PEC--Periodic Error Correction
"Smart Drive"
Subject: How PEC Works 
From: Bruce Johnson Date: Sep 2001
<dateach va.prestige.net> writes:
<< My question is this: How does the scope know where in the 8 minute cycle to
start the script? I mean suppose it starts the script when turned on, with
the first correction, say, at the 2 minute point. But suppose the first
error doesn't occur until 3 minutes? How do it know? >>
Tom, It's quite easy to understand..... once you know the secret! There is a small magnet mounted on the RA worm shaft, and
a sensor right next to it, mounted permanently. When you turn the scope on, one of the things it does.. in Polar mode.. is
to go forward to detect the magnet, then back up a bit, then approach it again slowly. When it detects it that second time,
it is in sync with the stored buttons and when they were pushed.
Likewise, when you train the scope, the actual training session begins just as the magnet is sensed, and continues for the
eight minutes you already know about.
The idea is pretty good, but at this moment, I am not completely sure just how well the magnet sensing is, timewise. I mean,
it seems possible that the sensor might detect the magnet, say, 100 ms later, the next time you turn the scope on. If that
happened to be the case, then your PEC training would be out by 100 ms for the night. Overall, that wouldn't make much difference
for the grand sine wave of the periodic error of the worm, but it could definitely throw your 'quickie tweats" out of
step, and you could end up with a significantly different PEC on one neight to the next with nothing else having changed.
This is something I haven't confirmed that happens as yet, but I have reason to be suspicious along these lines. Anyway, that's
how PEC timing is detected.
Subject: PEC (Smart Drive) Training --part 1 of 7
From: Fred Park
Observations:
1. Daytime work. With the clutch off, turn on your LX200 and go to the PEC (Smart Drive). Do an erase. Go to the RA and DEC
readout. Check the time on your watch and record time and the RA. Let the system run for about three hours and each hour record
the RA. Note that the is little or no change in RA. I got four second drift in three hours.
2. Nighttime work. Do a Periodic Error Correction (in Smart Drive) using Learn only. Now repeat #1. I got a drift East of
one minute in three hours.
3. Conclusion: When doing a PEC, if you don't correct the same for West errors as for East errors you will introduce a long
term drift which is undesirable for astrophotography.
Method of observation: Setting up as in #1, I logged the time every time the RA would change. The cycle of correction could
be observed and the repeatability of the gear could be observed. I was bothered when I saw a gradual drift in the RA, and I'm
still not sure what to do about it. Question: Am I nit-picking or is this a problem? More work: See how update effects the
RA.
----------------------------------------------------------
Subject: PEC (Smart Drive) Training (a method) --part 2
From: Doc G
Richard G. Davis posted on MAPUG a note on PEC training giving his error curve. On 7 July Fred Park discussed the fact that
PEC training can introduce long term errors in the RA rate.
In this note I discuss a training method which should allow for training of the worm which avoids the introduction of long
term RA error. Note that this technique does not account for the possibility or even probability that the RA gear has periodic
irregularities in addition to those introduced by the worm defects.
The technique is as follows and is based on the need to establish the correct baseline against which to make worm error corrections.
First take data similar to that shown by Davis over one whole period of the worm (8 minutes). It might be wise to take several
cycles to get an nice average curve. Then plot the curve and integrate over the entire curve for one period.
That is add up all of the positive errors and add up all the negative errors and subtract the two numbers. Then divide the
difference by the total number of data points. This number is the current bias setting of your crosshair. Now correct the position
of the crosshair by the amount of this bias (positive or negative as the case may be) and call that the correct (zero) position.
Now train the worm using the PEC technique by making the telescope point as accurately as possible to the zero reference.
This training method should reduce the wobble without introducing a long term bias in the RA rate.
PS: Note that this technique is similar to that used to balance an operational amplifier using integrating feedback. This
last comment is for all those Electrical Engineers out there.
----------------------------------------------------------------------
Subject: PEC (Smart Drive) Training: Some Conjectures --part 3
From: Richard Davis <richdavema.ultranet.com>
Some comments on this subject:
WHAT YOU GET WITH PEC (Smart Drive) TRAINING
Fred Parks has confirmed my findings reported earlier. The RA tracking curve data which I reported in a graphical plot of
telescope position against a constant object position, a star, did not show any appreciable long term RA drift. Those data
were obtained with UNTRAINED RA tracking, that is, the PEC (Smart Drive) was NOT operational.
In that report, after training the telescope carefully, a constant RA drift of about .6 arcsec/minute was observed with the
PEC (Smart Drive) TURNED ON. That residual drift is surely the result of "DC" bias introduced by the RA training
work that I did.
Conclusion: If you work real hard on PEC (Smart Drive) training, you can greatly reduce the VARIANCE in RA tracking due to
the WORM gear to less than 3 arcsec. However, a constant small RA drift rate is likely.
You can "guide" out the small, CONSTANT RA drift error, easily enough, if you have some way of guiding during exposure.
However, the persistent variance of RA tracking of some 2 to 3 arcsec has a cycle time of around 20 seconds. Autoguiders must
have a correction cycle time of at MOST 5 seconds to have any chance of reducing that inherent, TRAINED PEC (Smart Drive) variability.
You need bright stars, short CCD guider exposures, fast image analysis times and short correction times to make that work for
CCD guiders. ...or, you must use visual guiding that is continuous! Or, use a scope configuration which gives 3 to 6 arcsec
per CCD pixel image scale so that the RA variability is smaller than the pixels in the camera.
===========================================
"AVERAGING" OVER MULTIPLE TRAINING SESSIONS OF THE PEC (Smart Drive): "NOT"
My suspicion is that there is NOT any arithmetic averaging of successive sessions of the PEC (Smart Drive) training! Of course,
MEADE appears to be silent on this issue. (And, why should they speak up? We'd just use the information to beat them up for
not giving us a professional grade machine for junkyard prices! Of course, has QUALITY at the MEADE LX200 price point been
maximized? Hmmmm... Read on.)
My speculation of how this works is as follows:
The data retained in the PEC (Smart Drive) memory is probably as simple and compact as it can possibly be. If so, then all
that is recorded during a PEC (Smart Drive) training session are the RELATIVE TIMES, measured from the beginning of the training
session, when one of just three possible events has occurred. These events are: (1) go EAST at GUIDE speed (EAST button press),
(2) go WEST at GUIDE speed (WEST button press), (3) STOP moving at GUIDE speed (release of either EAST or WEST button).
After one training session, there is just such a list of commands in memory which is an exact record of your EAST and WEST
button presses.
Now, when you return to UPDATE the PEC (Smart Drive), the button presses you now make on the second pass are merely inserted
into the list of previously recorded presses. The only "averaging" that occurs is what you do visually by noting
the performance of the PEC (Smart Drive) on the second training session, and superimposing on that performance additional EAST/WEST
corrections.
Have you ever noticed that while training the PEC (Smart Drive) in an UPDATE session that there are times when you need an
EAST correction, you press the EAST button and there isn't any response? ...for a while. ...and as you continue to hold down
the EAST button, at some point, a correction eventually begins to occur?
What happened, I suspect, is that the PEC (Smart Drive) was already in an EAST correction command from the previous session
when you started pressing the EAST button. Therefore, no response is possible. The EAST moving rate is already at its maximum.
Eventually, the current EAST command terminated, the computer checked the EAST and WEST buttons, found that the EAST button
was pressed, and initiated an new EAST move. Then, you would see the newly commanded EAST move. You can determine when this
occurs because, if the PEC (Smart Drive) is in an EAST command state, the RA drive has been STOPPED. It will be silent! And,
as you command the EAST move, the scope can't go east any faster than it already is--that rate is determined by the rotation
of the EARTH!!
Likewise, any WEST commands issued by you while in an update session may appear within the intervals of previous WEST commands.
And, likewise, the effects of these new WEST commands can't be processed until the pending WEST command terminates. Even if
the command in progress were terminated, the rate of WEST correction wouldn't be increased above the GUIDE rate!
(I will not continue here to consider the cases where the current PEC (Smart Drive) command is in one direction, and you press
the opposite direction button. This message is already long enough. Try it for yourself.)
Now, for the inevitable steady RA drift AFTER lots of PEC (Smart Drive) training:
Notice that the mechanical problem for going EAST and for going WEST at guide speed ARE NOT SYMMETRICAL in a mechanical sense.
The EAST corrections involve stopping the RA drive motor and gear train. The WEST corrections involve speeding up the RA drive
motor and gear train. Due to the inherent "coast" of the drive mechanism, I suspect that the EAST correction cycles
are not going to have exactly the same amount of "zero-length" cycle effects as the "zero-length" corrections
in the WEST direction will have.
I believe that the well trained LX200 PEC (Smart Drive) will always have some small, relatively constant RA drift attributable
to the mechanical asymmetry of the RA drive resulting from stopping it in the EAST direction and speeding it up in the WEST
direction. The residual RA drift should be in the same direction for all LX200's, I suspect, and generally all scopes should
have about the same amount of residual RA drift. My best estimate of that is about 0.6 arcsec per minute.
Remember, the net RA drift over three hours is near zero. Fred Parks reported this observation, and my data also showed this
to be true. Then, after RA training, careful RA training, repeated RA training, the drift emerges. Some part of that is probably
due to the correction process itself.
Well, this little analysis is all speculation based on my experience and studies of my LX200. I may be wrong.
------------------------------------------------------------------------------
Subject: PEC (Smart Drive) and Training --part 4
From: Ralph Pass <rppassrppass.com>
I am not associated with Meade (other than being a satisfied user of their telescopes and a frustrated user of their CCDs).
The information I am providing is my best understanding of what is happening with PEC and training. It does reflect several
conversations with Meade over the years.
First, the RA motor is being driven by 45 pulses per second. Pulses are counted to provide the pointing location (this
is why it is more precise than typical amateur encoders).
Second, the period of the worm gear is eight minutes plus or minus a bit.
Third, Meade has divided this eight minute period into 100 subdivision. Each subdivision in 2.4 seconds long.
Fourth, PEC correction is accomplished by changing the number of pulses per second and is constant during each 2.4
time period (although it may be different for each subdivision).
Fifth, during training PEC is off.
Sixth, the one hundred adjustments to the pulse rate are store in the LX200.
Seventh, as training progresses the net adjustment to the average pulse rate in each 2.4 second period is stored (learn
mode) or averaged with the existing value (update mode).
Eight, in the x.34 version of the ROM (and perhaps in the x.30 version as well my aging memory is failing me), there
is a five digit number my the SMART entry on the keypad. This number is normally about 21600 ( = 8 minutes * 60 seconds * 45
pulses per second).
Ninth, deviations from 21600 counts may or may not reflect 'RA' creep in time. This reflects creep and the difference
between crystal frequency and true sidereal time.
Tenth, training at one temperature (e.g., winter) and using at another temperature (e.g., summer) will change the crystal
frequency and hence the apparent creep of the telescope.
Eleventh, any variation induced by the large gear may degrade the accuracy of the tracking (including PEC).
This summarized some of what I know about PEC and training.
Other comments:
What is clear is that if there are significant drive deviations over time periods less than 2.4 seconds, the current implementation
of PEC will not correct for this. I try to return my scope to a 0 HA at the end of the evening to use the same area of the
drive gear each evening. I have provided this information to help people understand the operation of their scopes. I think
that clarification of this information is appropriate follow-up. I do not think that criticism of Meade design choices is appropriate
follow-up. The magic number '21600' is the nominal value. The actual value shown is the number of pulses sent to the RA drive
motor with the current PEC correct applied. Deviations from 21600 reflect RA creep or variations from the nominal crystal frequency.
For the training I have, it is 21575. Mine ranges from 21550 to 21597. I have never seen either of my scopes exceed 21600.
I have not regularly gotten the number above 21590.
-------------------------------------------------------------------
Subject: More PEC (Smart Drive) Speculation --part 5
From: Doc G
Additional thoughts on RA motions.
I have read very carefully the many posts on the topic of RA motion and correction of the rate in Meade telescopes. First
let me say that my experience with guiding has been very good. I am pleased with the training of the worm drive and the good
short term corrections that can be obtained.
The many posts have been very informative and well thought out I think. I hope to make a few points about all these matters
that may have been lost in the large traffic in the last few days. I am starting with the assumption that the need for accurate
motion of the telescope is driven mainly by the needs of imaging. This implies that accurate motion over periods of an hour
would be generally adequate. To get accurate guiding, one must guide manually or with an appropriate CCD guider. A guider will
certainly make enough corrections to take care of motions over the longer periods of several minutes. The really important
question, it seems to me, is "can the CCD guider work fast enough to smooth out the short term wobbles." These short
term wobbles are in the range as short as a second or a few seconds and of amplitude in the range of a few arc seconds.
If the above situation is correct, then we need to be concerned with two sorts of imprecision in the guide rate. One is due
to the imperfections in the worm and the second is due to imperfections in the main RA gear. Since the worm turns once per
8 minutes, it rotates 180 times in 24 hours and thus the main RA gear has 180 teeth. We must face the fact that this is one-half
the number of teeth in a precision drive such as those from Byer or Losmandy. (at half the cost of course) Nevertheless, a
four minute worm with 360 teeth is inherently more precise. This is not a criticism of the LX200 design but a statement of
fact. The LX200 is amazingly accurate in its pointing capability and RA drive stability for a modestly priced mechanism.
A major question, which we are near answering, is the relative importance of the worm and the gear defects. I have observed
many cycles of the worm in my LX200 12" and feel there is a very strong repetition of the shape of the periodicity at
8 minute intervals. This means that the principle defects are in the worm and not the RA gear. (Though there are certainly
some defects in the RA gear as well.) This situation is very encouraging since it means that the 8 minute training of the worm
will eliminate most of the rate variations. I and others have found this to be the case.
Thus I feel that users should concentrate on doing a good job of training the worm using the recommended techniques and not
worry about long term drift that this method might introduce. Such long term drift is very small and has essentially no effect
during practical exposure lengths. (Anyway it is easily correct by the guider.)
The question still to be answered is how quickly a CCD guider can correct even a well trained worm.
I think that a bright star cycles the imager more quickly and results in more frequent corrections to the drive. Continued
study of the quickness of the corrections and the effect these corrections actually have on the drive motor rate and in what
length intervals is still needed.
At the present time, corrections seem to take place at intervals of 5 to 10 seconds. Possibly under computer control of the
imager and the telescope together, software could be devised to decrease the interval between these corrections.
-----------------------------------------------------------------------
Subject: Summary of PEC (Smart Drive) Operation --part 6
From: Doc G
Summary of some issues regarding the Meade drive systems.
>From the many posts on LX200 drives, I have gleaned some design information which seems to me to fit together and make
good engineering sense. I have designed similar system over the past many years and feel it is time to try to come to closure
on some of the issues that have run on over the past month. I believe the following descriptions are reasonably accurate.
First, the drives are closed loop control systems from the control electronics through to the motors. The motors are connected
to the actual main drive axes through a set of gears and a worm. The main shafts are not inside the control loop but are extensions,
through the gears, of the positions of the drive motors.
One source of slack or looseness in the drive systems is then any lash in the gear trains between the motors and the main
shafts. Any defects in the gear train or looseness cannot be fixed simply through control of the motors since it is out of
the loop. However backlash in the gear train can be compensated for to some extent with the backlash setting provided by the
design. Additionally some of the most dominant errors in the worm drive can be fixed by the PEC (Smart Drive) training technique.
To keep absolute track of the position of the telescope, the number of turns of the motor shaft must be known and kept track
of by a counting mechanism. I believe this is accomplished as follows: The motor has an optical transducer on the shaft which
sends to the control electronics a series of pulses which the electronics counts and compares to a computer generated number.
The number of counts tells the computer where the telescope is and the computer drives the motor to force the number of counts
sent by the motor to the number it knows is correct.
Counts are generated by a disk with 90 slots and two photoelectric pickups. The reason there are two pickups is so that there
is no ambiguity about the number or direction of the count generated by the motor. With only one pickup it is possible to have
false pulses generated and the direction of the encoder disk cannot be determined. With two transducers very slightly offset,
the transitions from the transducers will generate a two or four pulse code that gives both the number of counts and weather
the count is going up or down. This a simple type of Gray code. (I will not go into coding for unique output here. Too complicated
for this discussion.) Thus the 11.25 slots per second that the motor shaft moves at the normal RA drive rate yields 4 time
as many pulses to the control circuit or 45 pulses per second. (Both of these numbers have been reported in posts.)
Now the computer generates 45 pulses per second and the computer pulses and motor pulses are counted in an up/down counter.
When the motor gets behind the computer applies, through pulse width modulation, more current to the motor and speeds it up.
The computer demands that the motor follow. Sort of like the computer puts 'em in and the motor must take 'em out. (Or vice-versa
for the other direction) When the number is 0 the motor stops.
Then once the telescope is synchronized with a known position, it moves the telescope to a new position by simply demanding
a calculated number of counts from the RA And Dec. motors. This is a very simple, inexpensive and accurate method of positioning
a mechanism.
It has been reported that one slot movement corresponds to 1.3333 arc seconds of motion and that there are 11.25 slot moves
per second for the normal RA drive rate. Meade states that the position is accurate to 0.333 arc seconds. This is a factor
of four and corresponds to the supposition above that one slot motion generates 4 pulses just as expected from the transducer
scheme described above.
It is certainly not difficult for the computer to generate the required differential position distances in the form of RA
And Dec motions or for that matter Alt. and Az motions in the form of a pulse count.
Now consider the more subtle issue of correcting the backlash in the drives. The correction number is entered by the user
by selecting it manually. The computer simply remembers to add or subtract this number to the appropriate move command. Thus
the seemingly difficult backlash correction is taken care of easily by the computer through the same drive counting mechanism.
How about the correction for the irregularities in the worm gear. This correction is entered by the user in the form of E
and W pushes of the direction keys on the keypad. The computer knows exactly the position of the worm gear. For an 8 minute
period, one turn of the gear, a total of 200 corrections are entered. This is one correction for each 2.4 seconds. At the nominal
R. A. rate for example the computer demands 45 pulses per second or 108 pulses in a period of 2.4 seconds. Pushing the E key
stops the motor for the pushed period. This action must subtract pulses from the number demanded by the computer for that period.
Also, pushing the W key doubles the speed of the drive. So this action must add to the number of pulses demanded by the computer
in the period in question.
I cannot address the precise algorithm used to add or subtract pulses in the computer for the 200 periods. It could be on/off
or a proportional number depending upon the length of the key press or the key actions might be tallied and averaged in some
way. The details of the algorithm would be interesting to know, but it is not an issue of principle concern here. It is sufficient
to know that Meade has provided a very nice scheme for correcting the worm drive rate in 200 increments per revolution and
that the user can train the worm through setting up the PEC (Smart Drive) mechanism with great accuracy.
Many experiments can be done to ferret out the exact counting algorithm but I think it will be a difficult task. If someone
has figured it out with certainly, I would certainly like to know. I am satisfied at this point to feel I understand the basic
system and to simply train the worm as well as I can.
A way to determine the exact pulse train structure would be to use a logic analyzer and print out the entire pulse train.
This would be 45 pulses per second for a total of 8 minutes. (21,600 pulses) Clumsy but not impossible to do.
It has been reported that the total number of pulses deviates from the number 21,600 for one turn of the worm. I do not understand
how this can be the case since the number is determined by a gear train which can be calculated exactly. I suspect that the
worm rotation is not being measured accurately or some other systematic error is creeping in. This must be checked. Sixty pulses
represents 1 degree of worm notion. Alternately, the training routine might be introducing a systematic drift. This phenomenon
has also been reported.
Update:
Subject: PEC/Smart Drive --part 7 of 7
From: Doc G
But here is my impression of how it works. There is a sensor on the worm which is the start of the PEC process. I believe
this is the correct way to do the PEC thing. (Meade never tells how to do it, unfortunately)
Turn on the scope and Erase the PEC. Then go into learn mode and watch the display until it gets near 0. This may take up
to 8 minutes. When it gets near zero it will beep. Then start guiding as fast as possible, about every 2 seconds for as long
as the thing beeps. It will beep 200 times or 8 minutes which is one turn of the worm. That should do it.
Then if you want to refine the PEC you go into update mode and do the same thing all over again. After three times, you have
a well trained worm. (Or you fall in the floor with serious eyestrain) I got my wobble from about 50 arc seconds down to 3
arc seconds with three cycles.
Subject: PEC Questions from a Beginner 
From: Ralph Pass <rppassrppass.com> Date: May 2002
Dave Schanz wrote:
> I am just starting to work with PEC training as I get closer to doing some
> film work. I need help for the beginner.
>
> I started out with a very good polar and drift alignment. I was
> using a 12mm illuminated reticle on my 10" LX200 f/10 classic. I realize I
> might want to jack up the magnification, but for my first attempt at PEC
> training I wanted to keep things as simple as possible. The star I used was
> almost directly overhead and about 5 degrees west of the meridian.
>
> My questions are:
> What is the significance of the "pulse count" figure that shows on the
> display after training is complete? My first training ended with a pulse
> count of 21611 and I read in the archives that number is low (but no
> explanation is offered). After an update, the pulse count was 21597.
> However, I don't know what a good pulse count range should be and I don't
> know why mine were that low.
>
> During training my RA drift was very minimal and I was very careful to keep
> the star centered on one of the crosshair intersections. During the first 8
> minute cycle, there were several times when no corrections were required for
> periods of 10-15 seconds. Most corrections that had to be made were minor
> and to the west. DEC drift was also minimal, requiring perhaps 2-3 quick
> corrections to the north in total. Several articles in the archives mention
> making 2-3 corrections per second - which is way more than was necessary
> during my training. Is my initial experience of not having to make a lot of
> corrections indicative of anything in particular?
>
> It did occur to me after I was done that perhaps the reason I didn't need a
> lot of corrections was because of the relatively low magnification I used during training.
>
> Was my choice of a star near the zenith a contributing factor to the
> relatively infrequent RA corrections?
> My last question (for now) is perhaps a really stupid one -- once I have done
> PEC training, will the scope always use that training from now on -- until I erase it?
- Stars closer to the equator are better. The error effect is reduced by the cosine of the declination of the star (so at
40N it is reduced to about three quarters of what it is at the celestial equator). This is great for tracking but makes you
need to be that much more sensitive and careful in your training.
- NEVER adjust in DEC while training RA. There can be a wobble in RA due to Dec motion. This would cause a correction for
a phantom RA error.
- The motor is driven by 45 pulse per second. The cycle on the drive worm is 8 minutes. This leads to an expected total number
of pulses of 45 * 60 * 8 or 21600. Variations from this are due primarily to small frequency differences from the nominal
value. These vary with temperature. However, the PEC training is tied to the position of the drive worm and, once done, is
not affected by temperature (so you can train the summer and use in the winter.
- RA training is remembered from power off to power on, DEC training is not.
- More power helps
Subject: More PEC Questions & Answers
From: John Mahony. Date: July 2005
> From: Keith
> I'd like to ask further about.
> 1. John M wrote "In the end, the PEC count should be around
> 21600 when you finish." What does this mean, how is it
> useful, and where do you get or find this count?
The encoder wheel has 360 slots, and rotates once every 8 seconds, so there are 45 tics/sec. A PEC training session runs
for one cycle of the worm, =8 minutes=480 seconds. So the total number of tics in the training cycle is nominally 480X45=21600.
This number is displayed on the keypad somewhere in the PEC training process. When you train, if there is something causing
long-term drift (gear eccentricity, etc), you'll press one of the E/W bottons more than the other, so after being trained,
the scope will expect to see a total number of tics/8minute cycle that's more or less than 21600.
It's been awhile since I looked at this on our scope, but I think a good accuracy is if you're within a couple dozen tics
of 21600. If you're off by a few hundred, you'll want to do something about it. Long-term drift from fork flex or atmospheric
refraction can be minimized by training on a star near the meridian.
Main-gear eccentricity causes one part of the gear to be too close to the axis, and the opposite part to be too far, so the
scope will appear to run fast when using the first part, and slow at the opposite point, with good speed at the points half
way between. A way to find the good part of the gear is to aim at a star on the meridian with a crosshair EP, let the scope
run for 8 minutes, then check to see if the scope has run fast or slow. Then move the scope 2 hours west (30 degrees along
the celestial equator) with the keypad, then loosen the clutch, and manually move the scope back to the meridian. Now the worm
will be meshing with a different part of the main gear, so test again. Keep this up until you've tested half the gear, and
find the best part of the gear.
If you have a permanently mounted scope, once you've found the good part of the gear, avoid ever loosening the RA clutch
again. This way you'll be using the good part of the gear when you're aimed near the meridian, where you do the most imaging.
>2. and "The PEC records in 2.4 second time increments, ..."
> Does this mean the scope records the average of all
> corrections made during an increment and when running
> normally corrects once (the avg obtained during training)
> every 2.4 sec?
Right.
>3. I was wondering how much do the stars themselves tend to
> wobble and jiggle?
> The steadiness of the seeing seems to affect the performance
> I get. On what I'd call a bad night I'll get an average
> (some peaks higher, some less) of 4-5 arc sec peak to peak,
> a good night about 2-3.(all testing with the same scope
> setup across the same arc of sky)
Sounds about normal.
> 5. I have a 1997 10" classic that I use for photos, and
> have used the newer (last 2 years) gps versions but only for
> live observation. The newer ones seem to be a lot tighter.
The two obvious mechanical weak points in the classic were the sloppy gearbox, and the nylon dec bearings. Meade switched
to metal ball bearings for the dec in about the last year of the classics, and kept this in the GPS. The gearbox is also much
better in the GPS scopes (but if your's has plastic transfer gears, you might benefit from an upgrade to Buck's gears. Meade
eventually switched to metal gears).
> I was wondering if that "tighter" sense in fact translates
> into better astrophoto tracking.
It should.
>6. The new RCX400s seem to have had the astrophotographer
> in mind during it's design and developement. I'm wondering
> if anyone has an experienced opinion on it's tracking
> stability or accuracy compared to the recent and older LX200s
The RCX mount is basically the same as the 14" LX200GPS mount, which is the same as the smaller GPS mount, except for
a beefed up base casting.
Subject: PEC Training -- Stored? Update Often?
From: John Mahony <jmmahony@yahoo.com> Date: Dec 2004
A. Levine wrote:
>I have done the PEC for my 10" Classic about two or three times on
>consecutive nights. Are these corrections stored in the telescope or
>do I need to do it each time I power up?
It's stored.
> If it is stored, will using the erase function erase all
> of the stored PEC data or just the last entry?
Check the manual. I think there's a way to delete just the last entry if you decide it's no good. You can also completely
delete all PEC training.
> Also, even it is stored, should I from time to time continue to do the
> training. I understand that the more times you do the training the
> more accurate it becomes.
Not necessarily. The way Meade does this is not very good. If you re-do the training without erasing the old (use the "update" command),
it will average the new with the old. But it gives even weight to both. So after you've done it a few times, the old training
gets less weight. For example, if you do it twice, each gets 50% weight. But if you then do it a third time, then weighting
the new evenly with the old averaged training means you get 25%,25%,50% weighting for the three trainings. So the result is
that one bad training can screw up the previous training pretty bad.
If they had stored just one more number (the number of previous trainings), then they could give even weighting to all.
Subject: Keypad and PEC Training Trick
From: Doc G, Date: Nov 2002
----- Original Message -----
> In trying to do PEC training, the keypad seems to be very "jumpy" when
> pressed to make a small correction, and the star overshoots center on the
> reticle. When "learn" is complete, the corrections are worse than baseline.
> Is there a trick to using the keypad to get more precise star centering
> during PEC training? ------end of quote-------
The trick, if it is a trick, is to watch your guide star with the PEC zeroed out for several cycles. This way you can count
steps (beeps) and memorize what is coming. Then when you do the PEC training for "real," you anticipate the swings
and jumps. This gives you a bit of lead time to do the button pressing. It takes some practice.
I have found that with one initial training and one update you can get the RA guiding to 1/5 to 1/10 of the original.
Subject: RA Tracking by the LX200 (Analysis)
From: Richard G. Davis
Recent messages here have discussed the motion of the LX200 in Right Ascension (RA), and the performance of the Periodic Error
Correction capability in the LX200. What follows is a presentation of a quantitative method used to characterize the tracking
of my LX200. This is just one machine in the LX200 class of telescopes. I DO NOT SUGGEST THAT THIS IS TYPICAL OF THE LX200.
However, it is reasonable to assume that these results are typical until evidence to the contrary is presented.
The LX200 was mounted of the standard Meade tripod. A SuperWedge was used to mount the telescope onto the tripod. The scope
was polar aligned by the drift method. The declination drift was reduced to about 6 arc seconds per hour, which could be safely
ignored.
The scope was configured with a CCD camera. The scope was an 8-inch, f10 with a reducer operating at f6.3. The CCD camera
was a 9 micron pixel KAF400 chip device. Under these conditions, it was determined that one pixel on the CCD was equivalent
to 1.35 arc seconds of angular measure. The results are presented below in terms of arc seconds.
A series of CCD images were obtained. The camera was set up to take images as fast as they could be exposed and saved to the
disk of the computer used to control the camera. Thus, exposures were obtained at 2.67 second intervals. Exposure time was
0.2 seconds, which should make OTA motion insignificant.
Images were collected for about 20 minutes, which covers more that two revolutions of the worm gear. No telescope motion commands
were sent during the time images were being taken.
The data presented in the following graphic were obtained WHILE THE LX200 SMART DRIVE (PEC) WAS NOT RUNNING. Thus, the following
graph shows the worst performance of the LX200. The graph shows an idealized version of the results for a single 8 minute cycle
of the worm gear.
+----Seconds of Elapsed time
| ____________________________________
| | Relative Tracking Deviation
V | units= arc seconds
|-15 -10 -5 0 +5 +10 +15
sec. | | | | | | |
10 | x
20 | x
30 | x
40 | x
50 | x
60 | x
70 | x
80 | x
90 | x
100 | x
110 | x
120 | x
130 | x
140 | x
150 | x
160 | x
170 | x
180 | x
190 | x
200 | x
210 | x
220 | x
230 | x
240 | x
250 | x
260 | x
270 x |
280 x |
290 x |
300 x |
310 x |
320 x|
330 |x
340 x|
350 x |
360 x |
370 x |
380 x |
390 x |
400 x |
410 x |
420 x |
430 x |
440 x |
450 | x
460 | x
470 | x
480 | x
sec. +----+----+----+----+----+----+
|-15 -10 -5 0 +5 +10 +15
^ | Relative Tracking Deviation
| | units= arc seconds
| ____________________________________
|
+----Seconds of Elapsed time
These results show several significant features. First, there is a distinct higher frequency cycle motion superimposed on
the 8 minute worm gear cycle. This regular motion feature is about 55 seconds cycle time and has an amplitude of about 7 arc
seconds. Thus, in the relatively stable part of the 8 minute cycle, there is a regular deviation for this configuration of
about 5 pixels. Even this small periodic variation is large enough to be noticed.
At about 240 seconds into the cycle, rapid and extreme deviations occurs which go through two large alternating deviations
before resuming the more stable 7 arc second cycles. These large excursions cover a range of about 30 arc seconds. During these
large deviations, the rate of change in RA reaches a high value of about 0.45 arc seconds per second. These large, rapid deviations
are where the SMART DRIVE (PEC) in the LX200 is presented with its greatest challenge. Thus, the ability of the SMART DRIVE
(PEC) to correct this telescope is going to be limited by the maximum rate at which the drive can adjust RA tracking.
To correct a RA deviation to the EAST, the LX200 drive must be slowed down. If the maximum possible rate of EAST moving corrections
is going to be made WITHOUT REVERSING THE DIRECTION OF MOTION of the worm gear, one would just STOP THE DRIVE MOTOR. Thus,
the highest EAST correction rate would be the sidereal rate, which is 15 arc seconds per second. The worst RA deviation rate
observed in my telescope is about 0.45 arc seconds per second, which is very small in relation to the best correction rate
for the LX200.
The same data gathering procedures were repeated with the SMART DRIVE (PEC) well trained and turned on. Under these conditions,
the RA deviation rate was reduced to a steady rate of 0.57 arc seconds per MINUTE on the average. The deviations around this
average deviation rate had a maximum peak-to-peak amplitude of about 2 arc seconds. Thus, if the telescope configuration is
adjusted to give 2-3 arc seconds per pixel, this SMART DRIVE deviation would be less than the size of a pixel. Tracking adjustments
of 0.5 arc seconds per minute would keep the image on the same set of pixels. If the tracking correction rate is 15 arc seconds
per second, then a 0.5 arc second deviation could be corrected by stopping the worm drive for 1/30th of a second, a correction
that is well within the capabilities of the LX200 SMART DRIVE (PEC).
OK. SO MUCH FOR THE GOOD NEWS! The bad news is that there are other large RA deviations, which occur at frequencies below
one cycle per 8 minutes. They can be as much as 4 to 8 arc seconds deviations which occur in 1 to 2 seconds. These sudden,
large jumps are at rates from 2 to 8 arc seconds per second. These transients are so large that they approach the maximum possible
rate the SMART DRIVE (PEC) could demand, and possibly even exceed the basic capability of the SMART DRIVE (PEC). Likewise,
these rates are a challenge for autoguiders to correct at the GUIDE rate capability of the LX200.
These low frequency events appear to be related to irregularities in the large, main gear which the work gear drives, and
in spurious electronic variations in the LX200 drive electronics.
Subject: SMART DRIVE: Back to PEC
From: Jerry Gunn <jgunnmtco.com>
Thanks to all of you who spent so much time learning how the LX200 drive and electronics operate. You have all done a lot
of great detective work.
Doc. G. seemed to have it all the correct numbers which made sense. So it would seem from all this information that a sort
of averaging of PEC multiple training runs is what the LX 200 does.
1. The 8 minute worm period is time sliced into 200 pieces of 2.4 sec each. 200*2.4=480 sec = 8 minutes
2. The on board computer wants to see 45 pulses/sec to operate at the sidereal rate.
3. If we want to correct the RA drive rate, we need to store 200 pieces of information about how many encoder counts to add
or subtract to each 2.4 sec time slice to correct for worm inaccuracies.
4. During training, the numbers of pulses received during each 2.4 second time slice is recorded and difference value from
45 is stored. This is done 200 times, once for each time slice.
5. When PEC is in playback mode, these difference values are added or subtracted (east/west) to the normal 45 pulses per time
slice and the servo motor is either speeded up or slowed to get the proper drive rate.
6. Subsequent training runs would just add or subtract from the values stored in memory if and when the east or west buttons
were pushed during the time slice. Although this is not an averaging of values, (its really better than averaging) it accomplishes
increased precision with each training up to some mechanical limit.
Also, if all this is true, some constant long term RA drift will occur if the number pulses received by the computer from
the encoder is not equal to 9000 (200*45) over the 8 minute worm period. Theoretically, if the worm turns once and you have
accurately trained on a star the pulses must equal 9000, but this obviously cannot happen exactly. So long term drift is to
be expected.
Subject: PEC Training Experiences --part 1 of 2
From: Michael Clary Date: June, 2000
Based on inputs from LOTS of people, I did the following:
- Drift aligned for 15 minutes, east and south. No drift...status quo.
- Visual PEC training (has been done with CCD in past). Erased, Learn(ed) once, and update(d) once. My understanding is that
this results in an average for the LX200 mount rather than an improvement on the initial Learn corrections for PEC.
- Checked collimation with 9mm illuminated reticle (278X). Very slightly off, realigned.
- Bundled CCD power and SCSI cables and tied off to east fork and tripod leg to avoid cable drag. With cables on the east
fork, balance (with slight bias to the west) achieved by removal of 1/2lb ankle weight attached to east fork (1 pound left
attached to east fork). Good dec balance in all positions, mount has slight tendency to move west in all positions (very
slight).
- Did +/- X and Y tests with CCD. Guide speed (2X sidereal for LX200 mount) for 10 seconds each direction. MaxImDL/CCD used
for all camera control. Started with 00 dec backlash set by Meade paddle. On all tests, results showed orthogonality (i.e.,
move in +/- X direction did not show change in Y pixel location and vice versa). Tests were done sequentially (i.e., 7 tests
in +X and -X directions followed by 7 tests in +Y and -Y directions). Results: had to set paddle dec backlash to 75 to get
+Y and -Y distances almost equal (left a slight undercorrection)...this does cause some jerkiness in the movement!
- Located star field in vicinity of celestial equator at the southern meridian. Located and exposed guide star for 5 seconds.
Did guide star calibration resulting in settings of -10.34 in X axis, -4.87 in Y axis. Started tracking...relatively steady
X, oscillation in Y.
- Took sequence of 9 X 5min exposures. Results: football stars!
- During PEC training, I noticed that the object star did not move in either the + or - dec direction. Decided to try guiding
with only X corrections (assumption being that radical and jerky dec corrections might be effecting RA).
- Took sequence of 9 X 5min exposures. Results: still football stars, but not quite as pronounced.
- Took sequence of 9 X 2min exposures with both X and Y corrections turned off (i.e. unguided). Results: slight figure 8
stars (i.e., slight tracking error in both X and Y without correction). Conclusion: the mount DOES need guiding inputs, but
not as aggressive (BTW aggression was set at 1 for all above and following tests).
- Now the weirdness: Although calibration numbers were -10.34X and -4.87Y, I manually input -8X and -3Y. Took another sequence
of exposures. Results: almost ROUND stars! Further reduced numbers to -7X and -2Y, resulting in rounder stars, but with a
slight bulge in the y direction (i.e., dec still not perfect).
- Slewed to north meridian near zenith. Recalibrated and got similar numbers (-10.3X and -4.4Y). Manually input corrections
of -7X and -2Y. Sequence of 9 X 5min exposures still showed "almost" round stars (still slight bulge in +/- Y direction).
Long night but maybe some progress. The input of manual correction numbers works, but why?? As I understand it, calibration
said that a 1 second move in X equals a move of 10.3 pixels, so a move of 1/10 second moves just slightly more than 1 pixel.
Instead, I insisted that a 1 second move in X equals a move of only 7 pixels, so a move of 1/10 second moves 7/10 of a pixel.
I would think that my manually input numbers would cause MORE movement instead of decreasing the moves. What's up??
Anyway, muchas "THANKS" to all who offered advice. I'm obviously still not satisfied (won't be until they're all
perfectly round) and I'll keep klutzing around. Still need to try slightly "misaligning" from pole and various other
tweaks of parameters.
---------------------------------
Subject: PEC Training Experiences --part 2 of 2
From: Ralph Pass <rppassrppass.com>
Michael: It is suggest that you only touch the RA button during PEC training. As you noted, DEC corrections sometimes confuse
the issue.
Subject: PEC and Long-Term RA Drift - Solution
From: Craig Tupper <ctuppererols.com> Date: July, 2000
This problem is well described in the Topical Archive, but there is no real solution to it offered other than to live with
it, i.e., guide out the long term drift that occurs after PEC training.
I came up with a simple little solution, maybe so simple that it was obvious to everyone but me. Simply change to a higher
tracking rate.
After PEC training 3 times (2 runs each time) I found that I consistently got pulse counts of around 21535, pretty low, and
had very noticeable drift. Autoguiding with an ST7 at 63" focal length gave oblong stars on anything longer than about
15 sec guiding exposures, and sometimes you just can't go shorter than that, especially if you are using Track and Accumulate.
I did the math and found that my pulse count rate was about 0.3% low ( 1 - 21535/21600 ) . Multiplying 0.3% times the standard
60.1 Hz tracking rate, I decided to try a manual tracking rate (adjustable with the hand paddle) of 60.3 Hz. To make a long
story medium, I find that in practice a rate of 60.4 Hz gives me practically zero drift. I have taken several unguided 8 minute
test exposures at this rate and was shocked to see round stars. Needless to say, my autoguiding performance is now much improved.
Subject: Polar Alignment and PEC Accuracy?
From: John Mahony Date: June 2005
Alan Sickling wrote:
> I am intending to try some imaging soon, so I have attempted to
>
perform an accurate polar align on my pier-mounted 12" Classic.
> 2 nights ago I got some targets to set up on.
> I drift aligned for about 3-4 hours, and I am now wondering
> how good my finished alignment is for imaging work. I have
> a Meade Superwedge, to which I have fitted a long T-dog on the Az
> thread, and bushes/bearings in all the critical pivot positions. The
> adjustments worked smoothly with minimal backlash. I tried to measure
> the final tweaks I had to apply to get each axis spot on the final
> position, and as near as I could estimate they were in the region of
>
30-45". I presume from this that I can expect the RA axis to be at
>
the NCP within about 60" overall. After locking down all the
> fastenings, I checked both axes and could discern no drift off-line
> after 15 mins. The engineering in the pier and mount is such that I
> do not expect much (if any) creep from the set position.
>
> Any comments from you experienced imagers? Are my assumptions OK?
>
And how does this figure of 60" max error stack up with the
> practical requirement for good imaging? I have seen no actual
> quantified figures anywhere.
I use that as the "practical limit" for fine-tuning a good alignment. Due to atmospheric refraction, there's no
such thing as a _perfect_ alignment that will give no dec drift anywhere in the sky. Atmospheric refraction makes objects appear
higher above the horizon than they really are. For objects with alt higher than about 20°, the effect, in arcminutes, is about
1/tan(alt). From that it's not hard to calculate that a drift test done on an object 30° above the E or W horizon will leave
the mount's RA axis adjusted a few arcminutes above the pole. It is generally considered beneficial to have the axis slightly
high, since the tracking will more closely match the refracted path of an object this way for most objects, but the drift test
leaves the alt of the polar axis "corrected" for objects low on the E or W horizon (since that's where you tested),
so you might want to drop the alt of the polar axis a bit, since most imaging is done close to the meridian.
I use a direct method of polar alignment, at <http://geocities.com/jmmahony/LX10/polaralign.html>.
The idea is simple: set the tube to 90° dec (where here "90 dec" means relative to the mount, so that the tube
is parallel to the mount's polar axis), and then use the scope as a giant polar alignment scope. To do this, the tube must
be precisely parallel to the
mount's polar axis, but this is easy to check: with the tube at 90°, rotate the diagonal up between the forks so you can view
from the south side of the mount, then loosen the RA clutch, and view through the scope while rotating the scope slowly by
hand in RA. The image should rotate around the center of the FOV. If not, tweak the dec until it does.
To use the scope as a precision polar alignment scope requires a knowledge of the stars near the pole. Charts are available
on my webpage. Keep in mind that your view of these stars will be refracted by about 1', if viewed from mid-northern latitudes.
When rotating in RA, you'll notice that even with the dec tweaked precisely, the image never rotates in a perfect circle,
since the aim of the scope drops by a few arcminutes when the RA is near 6 or 18 hour angle compared to when the RA is at the
0 or 12 hour angle position. This is due to fork flex- the long axis of the fork cross section is parallel to the ground when
in this position, so they don't resist bending as well. This is another reason why I use 1' as a practical limit for polar
alignment.
Of course, orthogonality issues can also show up when rotating while at 90°.
> While I was watching the drift sequences, I also observed the
> lateral wanderings of the RA drive, which has had no training yet.
> (The PEC file has been cleared.) The maximum lateral excursions I
>
measured were in the region of 90-100" overall (i.e.+/-45-50"). I am
> quite surprised that it was this bad. Does anyone know what the
> normal range of PE error is for an LX200? Does mine seem
> particularly bad? I have set up the RA worm/pinion engagement to be
> as good as it can be got.
That does sound somewhat on the bad side, but not terribly. One problem with checking other sources is that many do not say
if their number is overall total, or +/-. IIRC, before I trained the PEC, ours was around 40"-60" total, and I've
heard similar numbers for other LX200s. Of course, a big question is whether the error is smooth, where PEC will help, or erratic,
where PEC may not be able to keep up.
Subject: First Pass PEC Graphs After Cleaning
From: Gene Chimahusky <lynol1000yahoo.com> Date: Oct 2005
Here is the final installment of the gear cleaning and PEC measurement saga: I did a second full learn cycle followed by
an update on the PEC after the gear cleaning. Seems to have taken out the RA creep I added the first time around.
Graphs and the image of M57 of the PE from a stack of 29 images each 60 sec long:
<http://www.geocities.com/lynol1000/lx200_internals/pec_100105.html> (about
38k total)
Subject: PEC Training with CCD Camera
From: Bruce Johnston Date: Sept., 2000
I managed to get my ST-7E to PEC train my LX200, and do it pretty
successfully!! When done, I watched a star for perhaps 15 minutes to see the results. The star stayed right on the crosshair
at 500x, with a couple of exceptions. Those exceptions are the same ones I've always fought, I believe. Those being, when there
is a rather radical 'hop' or speed-up/slow-down that would appear to me to be slight irregularities in the worm tooth. I've
always been chasing those buggers and really would like to find a good way to minimize them. (Perhaps the lapping of the gears??)
To train the scope, I set the speed at 63 Hz and trained, using corrections of .2 seconds. Then I set the speed for 57 Hz
and did an update. Presto!
I must admit that I've been spending a pretty fair amount of time with my CCDOPS for DOS program, trying my best to learn
how to minimize overcorrections, and although I still have a lot to learn about the finer points of the program, I think I
now know enough to give it a fair try.
The seeing wasn't very steady at all tonight, so I really didn't expect much in the way of results, but I was quite pleased
in the results. Hopefully, when I get a really steady night of seeing, I might be able to get the thing to get better. And
if I can ever get those pesky quick speed-ups and slow-downs mastered, I would have even higher hopes for the method. All I
can really say at this time, is that the results were about the same as what I've gotten when training manually. (I really
don't know the thickness of my 12 mm Meade illuminated eyepiece crosshairs nor their spacing, so I can't give a decent arc
second estimate.)
I just thought some of you might be interested, because, like me, you've probably thought it just couldn't be done!
Subject: CCD PEC Training URL
From: Frank Roddy Date: Nov., 2000
You might check the site for info on PEC techniques with a CCD.
<http://telescope.8k.com/Operating.htm>
Note: should open a new browser window over this one.
Subject: PEC Training with an Autoguider or Manually?
From: John Mahony <jmmahonyhotmail.com> Date: Feb 2003
>Doc G wrote:
>Tim Long is quite right about the autoguider technique being used by many over the
>years. The Pictors do send a signal directly to the CCD port. But the problem with
>using this technique is that there is a delay in correction depending on the
>camera and the brightness of the star and the update chosen as so forth.
>
>I still believe that the general conclusion was that a good manual training
>was still the best. The good trainer can watch the behavior of the mount
>for a few cycles and then anticipate its erratic motion. Thus the good manual
>trainer has an advantage over the CCD camera which always lags. The
>manual trainer has a big advantage in fixing sudden erratic motion of the drive.
>
>One good manual training and one additional training should reduce the
>errors in the drive by a factor of five to ten times.
Since you have a wide range of guide stars to choose from, you can choose a bright one and use .1 or .2 second exposures for
the pictor. The pictor takes less than a half second to check the image and do the correction calculation. Since the corrections
are usually small, that gives you corrections at nearly 2 per second.
If you practice doing it manually, you could eventually do better, especially with the human ability to anticipate, once you've
learned your drive's error. In fact it occurs to me that some software will give you the ability to record the error- in fact
you can just take an 8 minute exposure with the polar axis slightly off- for deliberate drift- to record the error directly
on an image, as a graph of time vs. RA error. This record could be useful for training yourself to anticipate upcoming error.
But the drawback to all this is that the PEC code on the classic divides the 8 minute cycle into rather coarse chunks- 2.4
seconds, I believe- (i.e., in each 2.4 sec increment, any corrections during that interval are added together to give a single
correction for that interval during playback), so in practice it would be difficult to do significantly better manually than
with the pictor. And the pictor will do it easily.
Anyone know what time increments are used for PEC on the GPS?
Subject: Refining PEC Finally Pays Off
From: Randy Marsden <jmarsdensan.rr.com> Date: Feb 2003
At various times I have mentioned the work I have been doing on my LX200 to reduce the periodic error and improve my polar
alignment and PEC training. It finally all came together last weekend. My first target was M82. The only convenient guide star
required 6 second exposures for guiding with a clear filter and 30 seconds to maintain guiding for the blue filter. The seeing
was very steady, which combined with very small guide corrections every six seconds, yielded 3.0 arc-second FWHM stars on a
ten minute exposure. The resulting color image is posted at:
<http://members.aol.com/rmarsden/Astrophotos/M82_color_2-8-03.jpg>
Then I moved on to M51 for which I wanted to do a longer series of luminosity exposures to get more of the detail. I found
a brighter guide star and was able to use 0.3 second exposures for guiding. This yielded 2.0 arc-second FWHM stars that are
very round. I was able to shoot ten 10 minute exposures before dawn began to break. The result of that series is at:
<http://members.aol.com/rmarsden/Astrophotos/M51_monochrome_2-8-03.jpg>
An example of the residual periodic error after one round of PEC training is posted at:
<http://members.aol.com/rmarsden/Astrophotos/PPEC_residual_error.jpg>
All of the results are from my 10 inch LX-200 that is about 5 years old. I am not positive of the exact age since I acquired
the scope from the estate of an astronomer who had died.
I was seriously considering buying an expensive equatorial mount for the optical tube. But persistence has finally paid off.
Now I will build an equatorial mount in my leisure to carry the OTA from my old LX-6 for which the mount is beyond hope for
astrophotography. I am using an SBIG ST-7E anti-blooming version.
After tweaking and tuning the mount for months, I concluded that most of the problematic error is actually inside the RA drive
gearbox. Only after going inside and carefully cleaning the gears did I get acceptable performance. The camera can correct
for slow periodic error but it cannot compensate for high frequency error which can come from irregularities, dirt and imperfections
in the gears inside the gearbox. Careful cleaning, tooth by tooth, of those gears yielded the biggest change in performance.
Careful polar alignment also greatly reduces the amount of guide corrections sent by the camera.
Of course, having an evening of exceptional seeing helped as well. The high declination objects like M51 and M82 also minimize
the effects of periodic error. The real test of the performance will come when I go back and try to image lower declination
objects like M16.
Subject: PEC Training on Different Teeth? --part 1 of 2
From: Doc G, Date: Aug 2001
Wayne Watson wrote:
> In an exchange about worm gears and PEC sometime ago by Bruce Johnston
> and Doc G, I found the following:
> =============
> Question (from Bruce Johnston?):
> > I'm disappointed in the inconsistency I get after a training session.
> > Next night, it's right back to the erratic movements.
> Answer (from Doc G?):
> That happens because the PEC trains on only one tooth of the wormier.
> The LX system does not allow for more sophisticated training. Though
> it should be realized that if you do successive training on the same
> star, you are training on several different teeth. The hope is that the PEC
> makes the scope smooth enough so that the CCD guider takes care of the rest.
> =============
Holy Moly what indeed is a "wormier."
Doc G: What I meant to say is this: The PEC trains on one section of the worm and on one tooth of the large gear. Note: I
do not call this the worm wheel since it is a simple spirally cut gear and not a true throated worm wheel. But that is a different
issue.
The PEC trains for one turn of the worm and thus it trains on one tooth of the large gear and on the surface of the worm,
over one turn, that happens to be in contact with the large gear. If you train further, the worm trains on another tooth of
the large gear, but one the same surface of the worm. It is clear that the worm surface used is the same, over and over, since
the worm remains in place tangential to the large gear and in one place. This goes on for each training session.
The theory PEC of training assumes that the main deviations are in the worm and not the large gear. Technically perfect training
of the worm/large gear would require training for one full rotation of the large gear and keeping the data for the entire 180
turns of the worm required to make the large gear turn one full rotation. This record would then have to be kept synchronized
with the data record.
Note that professional telescopes can and do use such precision training for the entire sky.
As a practical matter, the PEC trains for mechanical deviations in the worm and keeps a record for one turn of the worm only.
This record is used to correct the motion of the worm for each turn of the worm and for successive teeth of the large gear.
Thus it is not only possible, but likely that as the worm contacts different parts of the large gear, the mechanical deviations
will differ. Typically PEC training will reduce tracking errors by a factor of 5 to 10 times. This usually means from untrained
values of 30 to 50 arc seconds to 5 arc seconds. This is an excellent improvement, but not perfect and not good enough for
precision imaging. Thus the guider is required to give arc second or sub arc second guiding as required for long focal length
imaging. As a consequence of the PEC training method used, One might find good PEC on one part of the large gear and less satisfactory
results on another part of the large gear.
----------------------------------------------------------
Subject: PEC Training on Different Teeth? --part 2 of 2
From: Doc G
Gregg Ruppel wrote:
> Hi Doc-- Thanks for the detailed explanation (again) of PEC training. Let me restate
> a couple of practical points based on your description and get your feedback:
>
> 1. I have a permanently mounted LX200; always park it at 00 RA and 00 Dec.
> If I PEC train on a star near the celestial equator and local meridian, the
> PEC should be most accurate at that location in the sky because of the "one
> tooth" phenomenon.
Correct. But PEC will help everywhere.
> 2. If I slew to a region of the sky to the east or west, my PEC training
> may be less accurate because now I'm tracking on a different "tooth" (i.e.
> section) of the large gear. I may see periodic errors that didn't show up
> in the area of the sky where I originally trained PEC. Guiding is still
> better than with no PEC training, but may not be as good due to differences
> in the large gear.
Possibly and to be expected. But PEC will help everywhere since the major errors are in the worm.
> 3. As the scope tracks an object over several hours (as in long exposure
> photography), PEC performance may change considerably as the worm encounters
> different sections of the worm wheel.
Yes, that is why a good guider CCD is still required. A residual of 5 arc seconds will give oval stars. CCD guiding is essential
sine it can give you sub arc second lock on the star field. Atmospheric effects will then give you fatter stars, but you will
be guiding well anyway.
> 4. To maximize PEC (say for CCD imaging), training should be done with the
> scope pointed to the area of the sky to be imaged (I think this has been
> discussed before but I was never quite certain of the rationale).
It is not too practical to do the PEC at many locations because of the time factor. The same portions of the gear will be
used only if you never unlock the fork from the drive using the clutch.
I recommend this in any case for a permanently mounted scope. You can find a sweet spot on the gears and use that section
over and over.
Subject: PEC Training - Elongated Images
From: John Murphy Date: Jan 2002
------Original Message-----
From: Maarten Vanleenhove <maarten.vanleenhovepandora.be>
I have bouncing around getting elongated stars (at f/6.3). It is a real pain. It is a very fast error, which is not periodic.
I've tried just about everything to get it right, with no results. Some say it is caused by little chips of metal between
the gears, others say it is the RA drive assembly...
If anyone has a good solution for this problem please let me know.
PS: check my elongated ccd images at: <http://www.supernovae.be/>
I have a 12" LX200 and an ST8 and have experienced the same problems when attempting to image at f/6.3 to f/10. I have
talked to both Meade and SBIG to no avail. I have looked for metal chips and found none, and have tried almost every combination
of numbers possible in CCDsoft.
I finally backed all the way up to square one and did the following:
- - deleted the PEC
- - turned guider corrections off, acquired a star on the guide chip, dumped the errors to the log, imported into excel and
graphed the error ( about 30 arc mins).
- - trained PEC with camera
- - turned guider corrections off, acquired a star on the guide chip, dumped the errors to the log, imported into excel and
graphed the error ( about 10 arc mins).
- - erased PEC, trained PEC manually (1 plus update)
- - turned guider corrections off, acquired a star on the guide chip, dumped the errors to the log, imported into excel and
graphed the error ( about 5 arc mins).
(for clearing PEC, Meade recommends running learn mode of the PEC without pushing any buttons for one worm period, then train
over again." This is supposed to do a "better" job of cleaning out an erratic PEC. Sounded like bull pucky to
me, but I'll try anything once.)
I was also able to see that when I turned autoguiding back on, that the lower PEC led to better guiding. I was still seeing
random large errors, but the magnitudes of the errors were better. It is worthwhile to note that by reducing the focal length
to f/4 (f/3.3 reducer with no spacer) guiding works great, no jumps at all.
I now believe that my camera will guide at longer focal lengths without the random RA errors if I can get the PEC down low
enough. This probably entails adjustments to the worm, finding the sweet spot, and some other adjustments, that I haven't had
time to make yet. For me, having the camera train the drive resulted in worse correction than doing it manually. I have also
heard others report that training the PEC in the area of the sky in which you plan to image helps.
I guess, the moral of my story is don't guess about your PEC. Measure it, graph it, change it, compare it. Just altering the
values in CCDsoft had me chasing my tail for many, many nights. I figured something had to be broken. By taking a step back
and empirically looking at what was going on, I think I have proven that the equipment is doing what is expected. Such are
the travails of owning a mass produced (relatively) low cost instrument.
Subject: PEC - New Use!
From: Emery Hildebrand <emeryhearthlink.net>, Date: Jan., 1998
I do a lot of hour-long manually guided exposures and accidentally discovered something interesting in this respect.
If you erase PEC and retrain at the beginning of each exposure, you barely need to make any adjustments even for prime focus
shots. Because of the weight of all the accessories and differential flexure, it appears that best PEC values are different
in differing parts of the sky. This is true for both my 8" and 12" Classic LX200s, but it may not be true for all
LX200s. It would be interesting to see how many others can benefit.
Specifically, retraining does not seem necessary unless pointing more than 20 degrees from the last position it was trained
in. As soon as the scope is pointed to some other quadrant, required guiding corrections increase dramatically. Point back
to the original location and very few corrections are needed.
If you don't erase the existing PEC values this system will not work as well since it will merely average with the previous
training and, while improving things, will not make such a dramatic improvement. The improvement is so dramatic that a typical
hour-long exposure at prime focus of the 8" scope may need only 3 minor guiding corrections.
These are not artifacts of polar misalignment either since there is no evidence of field rotation on any of the resulting
photos.
If this result is repeatable, it can obviate the need for any CCD to be used as a guider.
Subject: PEC Problem Caused by RA Worm Damage
From: Doc G, May 2004
I am now advocating another fix for the RA drive. If the scope has ever been punished to the point where the main gear has
jumped the worm, you very likely have a damaged worm. The very parts of the worm that get damaged are the parts that contact
the main gear. If you have a PEC which has a sudden correction required, you likely have a damaged worm.
This can be fixed, or at least helped very much by polishing the worm. I suggest removing the RA drive and connecting the
motor to a single or possibly two battery cells so it rotates at about 2 or 3 rpm. Then take a very fine and tiny file, a triangular
or square file is good, and chase the worm from one end to the other about 50 times. Then take a piece of 600 emery and fold
it and chase the worm another 50 times. This will remove or reduce the damage from the worm bearing surface.
I have one LX200 12" that will track for 3 minutes without guiding.
Subject: PEC Programming as Related to Drift Alignment
From: Michael Hart, Date: May, 1998
BACKGROUND
Programming the Smart Drive (PEC) should be done on a night of good seeing, no wind, at high power, and on an excellent polar
aligned scope because improper polar alignment introduces RA drift that effects the PEC programming used in other parts
of
the sky. One possible variation might be to program the PEC near the object of interest to compensate a bit for polar alignment
errors which could including Dec drive training. I use 800-1200X on my 12" LX200 on a star quite near meridian close
to the celestial equator. I make 2-3 small corrections per second to avoid overcorrecting at each 2.4 second recording
cycle.
This is done in RA only. No Dec corrections are ever made. The measured results are consistently around 4-5 arc seconds periodic
error peak to peak over the 8 minute worm cycle. Once done, the PEC corrects for small drive frequency variations and virtually
stops RA drift for at least 2 minutes. With a well programmed PEC and accurately polar aligned scope, one does not need to
make frequent guiding corrections at all. In fact, a human is often better at guiding than an autoguider during poor seeing,
because the human can easily determine seeing errors from alignment/drive errors. Attempting to autoguide out seeing anomalies
is a daunting task, especially with a computer controlled autoguider approaching a limited maximum update rate of 1.45
seconds
or greater.
DRIFT ALIGNING BEFORE PEC PROGRAMMING
If your scope drifts consistently in RA, you may need to tweak polar alignment before PEC programming. Three things happen
when we are NOT polar aligned: The stars appear to drift in RA. The stars appear to drift in declination. The field rotates.
Using the drift method to polar align is recommended highly for high power visual work and precision imaging. As we approach
3000+ mm focal lengths, alignment that was satisfactory for a camera lens or small refractor is wholly inadequate. The term
"declination drift" itself implies a polar aligned scope will drift in declination only. In fact, a scope that is
not polar aligned drifts in RA as well, however, we ignore any RA drift during the process and watch "declination drift".
As we
approach polar alignment, RA drift slows as well. The reason for RA drift is easy to visualize. If our mount is not quite
aligned with the celestial poles, the amount of misalignment formed causes the RA to appear to drift. The greater the misalignment,
the faster the RA drift.
Note: Start by erasing the PEC to remove any programmed RA drift.
The idea is to line up our scopes to turn parallel with the earth's rotation. Those experiencing unusual RA drift should assure
polar alignment is quite accurate.
LEVELING THE POLAR MOUNT- IS THIS REALLY NECESSARY?
There is a good reason to level the mount along the altitude and azimuth axis. When drift aligning, a mount out of level will
cause adjustments in one axis to effect the other, increasing the number of iterations that are required arrive at a polar
alignment solution. For example, a correction that might require an azimuth move will also move a bit in altitude when the
mount is not level in azimuth, the amount depending on the degree of error from level. Once quite level, we are then ready
to start the declination drift alignment process. The method outlined in the Meade manual is based on using a diagonal. Straight
through users and south celestial pole users, should reverse the correction directions. It may be a good idea to label the
direction moved on the mount azimuth and altitude knobs to prevent mistakes that would unnecessarily prolong the process.
DECLINATION DRIFT ALIGNMENT OF A POLAR MOUNT
Now that we know which star to select and have done so, we're not quite ready to drift align. First, we must align our reticle
eyepiece with the RA axis. We can consider this the E-W axis. Everything above the E-W (RA) axis is NORTH, everything below
is SOUTH. If you are uncertain, merely moving the scope with the E-W keys will identify the E-W (RA) axis in the eyepiece.
It is vitally important that we understand that the use of the term north or south as described in drift alignment procedures
is not related to your position at the telescope, rather, the direction the star drifts with respect to the RA axis. Now, we
select a star within 5 degrees or so of the celestial equator and within 30 minutes of the meridian. This provides maximum
declination drift which readily speeds the alignment process. A moderately bright star often provides better results than a
very bright star. Using a 2-3X Barlow with extension or star diagonal will produce powers that are quite high, amplifying small
drift movements. If the seeing is so poor that your moderately bright star is too dim and/or moves about, postpone the PEC
programming for a day of better seeing. If the star drifts NORTH, use the azimuth control to move the scope EAST. Keep adjusting
and re-centering the star until the movement virtually stops over a 3-5 minute period. Now, locate a star at about 6 hours
RA not much less than 15 degrees of the horizon (this avoids refractive errors) If the star drifts NORTH, use the altitude
control to move the scope DOWN. Keep adjusting re-centering until the star does not drift 7-10 minutes. Now return to the paragraph
immediately above and repeat, but strive to improve star drifting from 3-5 minutes to 7-10 minutes.
RETURNING THE TELESCOPE TO A POLAR ALIGNED WEDGE
This leads to the subject of reproducing polar alignment when removing the scope from an adjustable wedge or other device which
maintains the scope parallel to the earth's rotation (aligned to either celestial pole). Once the
wedge is aligned, we can move the scope base up, down, right or left as long as the scope base is not TILTED by a non-flat
surface. I use indexing pins to position the scope base at the exact point it was aligned. However, are they really needed
if you have a flat mounting surface (no bumps or warpage)? An even better question is, (thanks to Paul Goetz), how non-flat
(irregular) must the surface be to cause a significant polar alignment error? If moving the scope base laterally across the
wedge surface (parallel to the earth's rotation) introduces a 12 minute error west (due to a .020" bump or warp in the
surface), a 60 minute exposure could drift around 3 minutes in declination. A .020" deviation in a short span of less
than .25" (maximum anticipated variance) is clearly visible on the wedge surface and quite easy to see in the eyepiece
during drift alignment. However, the maximum amount of field rotation seen in a typical 60 minute exposure near the guiding
point would be fairly small- around 30 seconds. If the amount of field rotation produced is so small, should most bother with
indexing pins? Probably not. Why then, bother at all?
Three reasons come to mind.- First, manual guided exposures are much easier with an alignment precision that requires almost
no Dec corrections or RA corrections in two minutes or longer. One can relax a bit. Thus, more precise polar alignment is desirable.
Second, two minute unguided CCD images will move less than 1 pixel, enabling electronic stacking with minimal loss of field
of view and object centering. Third, guiding errors while using an autoguider are narrowed to flexure, vibrations, wind, drive-train,
and seeing. We could use precision milling of the wedge surface to 0.002", but simple eyeballing the bolt hole centers
or simple indexing pins as a means to assure returning to precise polar alignment is quite adequate.
CONCLUSION
I believe the value of accurate polar alignment is often underestimated. For casual observing, casual polar alignment is quite
adequate. However, for improved PEC programming and good photography results at focal lengths of 3000+ mm, a well aligned mount
is essential to excellent results.
Subject: Difference Between Periodic & Pointing Errors
From: Bill Ezell Date: May, 1998
- > They are talking about periodic tracking errors, and they give an number (15
- > arcseconds in that case). My question is : is it 15 arcsec each second,
- > minute. What is the period of time?
There's a lot of confusion between periodic error and drift, and people seem to sometimes use the terms interchangeably.
'Periodic error' should be used to describe tracking errors that repeat in a cycle over some period of time, hence periodic.
'Drift' should be used to describe tracking errors that persist and accumulate in one direction.
Periodic error in the referenced article should refer to the total amount that the image moves back and forth in the field
over one period. This shouldn't be a time-referenced quantity. So, what this means in practical terms is that you should see
the target move back and forth over a period of time over a total distance of 15 arcsec, and this would repeat indefinitely.
Note that at one point in the cycle, the image should (theoretically) return exactly to the location it had at the same point
in the previous cycle.
Drift is caused by misalignment of the mount or an incorrect tracking error. The appearance of this is that the image slowly
moves in one direction, never reversing its movement. In practice, both drift and periodic error are usually present simultaneously,
although drift can be essentially eliminated by careful alignment, assuming an accurate drive clock.
For the LX200, the significant time interval for periodic error is 8 minutes. This is the time the worm gear in the RA drive
takes to complete one revolution. There are additional second-order error sources, such as differences between tooth spacing
on the main RA gear.
And then there's pointing error, caused by mount flexure, alignment errors, drive encoder resolution, etc. which causes the
scope to point to a different area of the sky than it should when slewing to a particular location.
Subject: PEC Training -- Constructing a Guiding Keypad
From: Michael Hart, Date: May, 1998
A CCD guider port keypad is quite easily built using Radio Shack parts. The CCD guider port keypad is much simpler to construct
than a LX200 hand control keypad I use that relies on digital signals.
Making your own keypad control box is really quite simple and easy to do and is excellent for manual guiding and PEC programming.
I prefer snap action type switches for this application as the button travel is shorter than standard push buttons. Unfortunately,
these are not available from Radio Shack. It is quite likely this will work on scopes other than the LX200, such as the Celestron
Ultima 8 & 11, hand controller which I have verified as roughly pin compatible for this application.
RADIO SHACK PARTS NEEDED:
- (1) 279-422 six conductor telephone modular to modular 25' extension $7.99
- (1) 275-1547 normally open push button switches (4 in package) $2.99
- (1) RSU 11907680 molded enclosure $ 1.99
- Solder
- Soldering iron or gun
- Hot-melt glue gun (optional)
CONSTRUCTION STEPS
- With a sharp Xacto knife, cut the six conductor modular cable at a length of around 3-1/2 feet. Cut another piece of around
4 inches and remove the wires. You will use these to wire up the switches.
- With the Xacto knife, cut a slot in the molded enclosure base for
- the six wire flat telephone cable to allow the lid to close. You may need to remove a bit of plastic in the lid to allow
the lid to seat
- Mark 4 positions in the lid for the bush buttons and drill with a 7/32" bit and install the 4 switches.
- Wire one contact on each of the 4 switches to each other. This is used as a common ground.
- Look at the telephone plus so that the wire is away from your eyes with the plug retaining clip on the bottom. The pin-out
is counted from the left to right to- 6, 5,4,3,2 & 1. Now tip up the plug so that you are looking at the retaining clip.
Write down the colors of the wires you see from left to right. They will be: blue, yellow, green, red, black, white --OR-white,
black, red, green, yellow, blue. Match up a color with the pin-out above. You will use this information to wire the switches
so that the telescope moves in a expected direction.
- Pin 6 goes to the right button (blue or white).
Pin 5 goes to the up button (yellow or black)
Pin 4 goes to the down button (green or red)
Pin 3 goes to the left button (red or green)
Pin 2 goes to the ground in step 4. (black or yellow)
Pin 1 goes to NOTHING (+ 5v) (white or blue)
- Final assembly involves optional hot gluing the telephone cable to the top to provide strain relief, if needed and attacking
the base.
Subject: Dec Blocking Autoguiding Device
From: Wesley Erickson <twesleytwesley.com> Date: Oct. 2002
I have posted a PDF document on my website that contains instructions for building the "Dec Blocking Autoguiding Device".
Credit goes to Adam Block, lead observer for the Advanced Observing Program at Kitt Peak, who suggested the concept to me in
a phone conversation. I built a prototype the next day, and, after testing it with my LX200 and Pictor 201XT and Pictor
216XT, I sent it to Adam, who has extolled its virtues on the SBIG group.
The device installs inline between the autoguider and the LX200 control panel, and prevents Dec reversals by allowing the
user to turn off guide corrections in either the north or south direction.
Both switches should be on during calibration, then the desired switch can be turned off before starting the autoguiding process.
The link appears as Block Autoguider Device on the Technical Reference page at:
<http://ourworld.compuserve.com/homepages/twesley/blockdev.htm>
Note: should open a new browser window over this one.
The Adobe Acrobat PDF file may be downloaded directly at:
<http://ourworld.compuserve.com/homepages/twesley/blockdev.pdf>
or here.
Subject: Effective Use of PEC with an Autoguider
From: Michael Hart, Date: Aug., 1998
BACKGROUND
Since the introduction of affordable and reliable autoguiders, autoguider use has increased considerably. And why not? Guiding
can be a tedious and a monotonous task. With an autoguider, one can set an alarm and wake up hours later refreshed when the
exposure is complete. Still, it is not unusual (though perhaps less so in recent years) to see an amateur using rather inexpensive
equipment, minimal accessories and no autoguider. That amateur may walk away for a minute or so, return, converse, and casually
make a small correction. Contrast this to the amateur with many accessories, a much better drive, more expensive scope, and
an autoguider. Of course, the former amateur will get worse results as compared to the later amateur. Perhaps not- here's why:
Conventional wisdom used for film imaging with high quality mounts under excellent seeing is not simply transferred to inexpensive
amateur mounts used under typical seeing by many amateurs. I have used inexpensive mounts simultaneously along side high-end
amateur mounts and produced similar results, but not using the same techniques. Clearly, it is more of a challenge to obtain
similar results with an inexpensive mount. However, if we modify our techniques a bit to compensate for known anomalies, good
results are possible on inexpensive mounts with a reasonable level of effort. This is important for those that have limited
resources to invest in astrophotography.
WHY USE PEC
PEC is, of course, needed less with high quality worm and worm-wheels at shorter focal lengths. It is easy to dismiss mass
produced mounts as unusable if we are fortunate enough to own or have access to a better mount. However, as we start moving
up in focal length to over 3000 mm coupled with long exposures, PEC again starts to become useful, even for the high-end amateur
mount. Fortunately, PEC is available on many of these mounts. If used properly, PEC and an autoguider can have a synergistic
effect. This is especially noticeable with long exposure tricolor CCD imaging at 3000+ mm using a CCD chip where the CCD chip
records small errors much more efficiently than film. Many PECs are capable of averaging corrections over periods of 0.50 Hz
(2 seconds) or less (LX200) while a typical autoguider is operating at 3-5 second correction rates or greater. If we increase
the autoguider correction rate too much (faster), we may start chasing the seeing resulting in oscillation of the mount, especially
under average to marginal seeing.
THE ROLE OF AUTOGUIDER CORRECTION RATES
We should not confuse the exposure time setting of the autoguider with the correction rate which is the sum of exposure time,
integration time, any software/hardware overhead, and drive hysteresis. Thus, the correction rate is always longer than the
autoguider exposure time. This is an important consideration is we try to use an autoguider to program the PEC. A short exposure
setting might not be adequate to assure correction rates are less than the desirable sampling rate of 1/2 the record period
or about 1Hz (1 second) for the LX200 Then, of course, the autoguider is fully capable of programming non-periodic seeing related
corrections (undesirable) easily ignored by the human when the seeing isn't superb.
CORRECTION RATE LIMITATIONS IN MOUNTS
We should not assume that because our mount appears to oscillate in response to guiding corrections as a result of seeing,
worm errors, vibrations, alignment errors, and/or PEC programming errors that a given mount is not capable of fast guiding
rates. For example, a slightly modified LX200 is capable of responding to simultaneous X-Y guiding corrections at 2 Hz (0.5
seconds). For the RA axis, this consists of typical slowing and speeding of the RA drive. In the Dec axis, full worm reversal
slows drift or changes the direction of star movement. Of course, obtaining smooth correction rates at 2 Hz requires a thorough
understanding of the relationships of various adjustments in the drive and mount. For most, such fast correction rates are
not needed, however, the reduction in hysteresis required to achieve these rates is desirable.
THE PEC ADVANTAGE
For a LX200, if the PEC was calibrated and updated under excellent seeing and polar alignment, not only are the worm errors
smoothed, but residual RA drift is virtually stopped as clock crystal drift and minute alignment errors are corrected. Then,
subsequent updates are used to add or subtract pulses (0.3333 arc second movements) to memory segments which tweak the drive
further allowing the drive correction to track the worm quite close to real time- much closer than is practical for an autoguider,
especially under marginal seeing.
It is important that permanent PEC programming is preceded with good polar alignment and initial PEC erasure. Why good polar
alignment BEFORE PEC programming? Because poor polar alignment produces changes in the RA drift as well. With PEC programming,
in addition to clock frequency errors and worm errors (which are repeatable), we introduce another error (RA drift) which is
corrected by the PEC as well. The next day, we set up and roughly align again and find our RA drive doesn't track as well.
This is likely because we programmed in the RA drift produced by the previous night's poor polar alignment. The previous night's
rough polar alignment is usually more difficult to replicate than accurate polar alignment.
Once the PEC is carefully programmed, we can use the autoguider independently of PEC at whatever correction rate that produces
the best results- perhaps several seconds to even minutes (if desired) for a well aligned mount. Now, autoguiding is truly
optional because manual guiding corrections may result in one correction every few minutes or longer as the PEC has corrected
residual RA drift and the excellent polar alignment checks the Dec drift. Those on a budget can save considerable expense without
sacrificing results using their PEC wisely on a well aligned mount while manually guiding.
CONCLUSION
This is not to say an inexpensive mount will outperform a high end mount, however, clever implementation of PEC such as by
the LX200 can minimize the differences in results- which is what is important. Moreover, high-end mount users can use well
implemented PEC to improve results at longer focal lengths as well as improve manual guiding results.
Subject: Smart Drive (PEC) Quantitative Info
From: Frank Roddy Date: March 2002
You might want to look at the operating section of the Oak Mountain Observatory web site:
<http://telescope.8k.com/Operating.htm#Tracking>
to get some quantitative information on the improvement that can be obtained with Smart Drive.
Subject: Guide Correction Speeds? ---part 1 of 3
From: Blair MacDonald <b.macdonaldns.sympatico.ca> Date: Jan 2003
----- Original Message -----
From: Don Tabbutt<dontabbutt.com>
For years, members have said that when East (West down under) is pushed during guiding in Polar mode, the RA motor simply
stops. It also says this in several articles in the Topical Archives. But the manual says this:
"The 2X sidereal speed in GUIDE has one difference in that it will not interrupt the Right Ascension tracking direction
to make Easterly (for Northern hemisphere) or Westerly (for Southern hemisphere) adjustments; it will merely slow down the
tracking drive to one half its normal speed."
The reason I'm bringing this up is that I just bought Ron Wodaski's book "The New CCD Astronomy", which has an
excellent section on calibrating an autoguider.
If the manual is true, the guide star will move at three different speeds on the autoguider's chip (reverse East and West
for down under):
- 2X sidereal in declination.
- Sidereal when correcting West.
- .5 sidereal when correcting East.
Has anyone specifically measured these speeds? They are critical when setting the calibrate time with any autoguider. I
could certainly measure it myself, but it will be at least a week before we get clear skies around here.
----------------End Original Message-------------------
Don: I have measured the speeds while I was writing APT. They are 2 times sidereal when the west button is pressed and zero
when the east button is pressed.
---------------------------------------------------
Subject: Guide Correction Speeds? ---part 2
From: Bruce Johnston
I've measured these speeds. Going West, you go at 2x Sidereal speed for corrections. That is, the motor does indeed run twice
as fast in guide-correction speed as does the scope when it's just cranking along, following a star on its own. Then, for a
correction to the East, the motor does indeed stop. This makes the sky itself do the 'moving' while the scope sits still, giving
it the 'effect' of running to the East at Sidereal speed.
One easy way to prove that the scope stops when going East is to just set up in the daylight. Aim the scope at something on
the ground. (Be quick! It'll be 'moving'! <G>) Then press the East key and you'll see that the scope just stops. I've
done this many times for various experiments.
---------------------------------------------------
Subject: Guide Correction Speeds? ---part 3 of 3
From: Eugene Lanning
I did am empirical test to verify that the motion is zero when the East button in Guide is pressed. I aligned the scope using
a imaginary two star alignment ( done during the day). I the slewed the telescope to a terrestrial object ( a lattice work
some distance away). Over a 15 minute period, while pressing the East button, the telescope location on the lattice did not
move. When the button was released, the telescope position on the lattice started moving. It convinced me that the motion was
stopped.
|
