De-Rotator & Guiding

       

Subject: De-rotator & Guiding Discussion

From: Doc G

The following is a long note on guiding and de-rotating. It is an accumulation of thoughts and posts I have made during the past several months. Also discussed is the use of two or more tubes for guiding and imaging including considerations of wedge mounting, de-rotating and piggy back camera mounting. The several possible cases are discussed in detail starting with simple cases and going to the more complex.

These are my own thoughts based on calculations and experience with pointing my two telescopes over the past two years. I hope they do not deviate too much from others experiences or reality.

The first thoughts on guiding are for cases where a camera (imager) is used piggyback and the main tube is used for guiding, the main tube is used for imaging and a separate tube (telescope) fastened to it is used for guiding or some such combination. This does not preclude using several imagers on several tubes with a guider on another tube all on the same mount and possibly all in use simultaneously.

The advantage of using a separate guider is apparent to anyone who has ever used an off-axis guider. The separate guider tube can be of fast focal ratio so as to get a bright guide image, its axis can be adjustable and so be moved a bit from that of the imaging tube so as to center the guide star on the guider chip and an independent, stand alone guider chip can be used in its independent operational mode. It is very convenient to have guider and imagers independent. Of course the ST 7/8 offers additional options. I have not yet used the ST 7 in the guider mode so this arrangement is not discussed here.

To start with I assume accurate POLAR alignment. Assume that the guider finds a star and is working correctly. This situation insures that the guide tube, is pointed at a star and locked onto it. This is the most straight forward setup and is relatively easy to establish when a reasonably fast guide tube is used. I use a C5 on my 12" LX200. If polar alignment is not excellent, some rotation of the field will take place. Also assume for this simple case that the atmosphere causes no distortion of the celestial sphere. This assumption is not quite correct, so this issue will be discussed later. For the ideal case described, the declination is fixed and the telescope needs only to track the R.A. perfectly. There will be no de-rotation required and a de-rotator is not needed. If the telescope moves precisely in R.A. no guider correction is required either. Now assume that the telescope does not move precisely in R.A. Naturally it doesn't because of the worm drive defects. The LX200 has a fine system for correcting periodic worm drive defects which can reduce them from 50 arc seconds to 3 or so through a training program. Even then the guider may need to make some corrections to the RA drive rate. This has been done manually in the past and now, with a guider, can be done automatically. If there is no distortion of the celestial sphere by the atmosphere, then the guider tube can be pointed at a declination or R.A. other than that of the imaging tube and guiding will still be perfect.

This is true since the undistorted celestial sphere in its entirety moves with the same angular rate in R.A. everywhere. Also note that the above assumes that the guide tube and the imaging tube are rigidly held together with respect to each other. Considerable care must be exercised to insure mechanical rigidity of the tubes. A counter example is given by suggesting that this scheme will not work with the guide tube pointed at the pole star. And sure enough it won't. So what is wrong? It is this. The effective length of the guider telescope is longer when it is pointed at a guide star that is moving the greatest linear amount for a given angular motion of the celestial sphere. This linear amount is proportional to the sine of the angle from the pole to the declination being guided upon. Thus when pointing the guide telescope at the pole star. The motion of the star is nearly zero and the guider fails. To get the best accuracy with the above conditions, the guider should be pointed to a declination at 90 degrees to the pole. Then, the guider will be locked to the celestial sphere where the linear motion of the guide star is great and the guiding will be accurate and all will be well. Even if the polar alignment is perfect and the guider is perfect there is a complication. That is the atmosphere. Except at the zenith, the atmosphere distorts the position of the stars in the celestial sphere as it is seen from the telescope position. Thus as a star moves from near the horizon to a position higher in the sky and again toward the horizon, its motion is not perfectly regular in angular rate of R.A. nor does it maintain exactly the same declination.

The amount of this deviation is quite small. The Meade telescope has a first order correction for this effect in its computer. This is why the correct latitude and longitude must be entered into the computer to insure refraction correction that relates exactly to the local horizon and thus insure accurate pointing. Thus it is clear that the guide tube should be pointed somewhere near the position of imaging tube. If they are close together they both see the same atmospheric distortion and the guider will make the required R.A. and declination corrections. But the two tubes do not have to be exactly axial aligned. Usually, imaging is done away from the horizon because of light pollution anyway so the atmospheric distortion is usually not a great consideration. This distortion is usually masked by "seeing defects" in the atmosphere in any case. In summary, polar mounting eliminates the principle tracking defects and is thus very attractive for imaging. Several tubes can be pointed from the same platform and all will function well with a single guider. A separate but important consideration is the focal length (more importantly, the effective focal length) of the guider tube compared to that of the imaging tube. The guider cannot point more accurately than the angle subtended by one pixel within the guider image and may not be that good. Thus, the focal length of the guider tube should be similar to that of the imaging tube. It is usually recommended that the guider be at least 1/2 the focal length of the imager. For example, I use a C5 (fl=1300) for my 10 inch f6.3 (fl=1200) and also on my 12 inch f10 (fl=3000).

I feel this is an adequately long guider focal length. Any shorter focal length imaging lens such as a typical piggy back camera would normally use is then easily guided. Now, what about the Alt/Azm mounting for which celestial field rotation is a problem when imaging. Normally a de-rotator is necessary. The rate of de-rotation is a complex function of the R.A. and declination to which the telescope is pointed. A table of de-rotation rates is attached to this note. The table is normalized and so the numbers must be multiplied by the local rotation rate which is 15.2 X cos(latitude). Also, values above 80 degrees are not given since they become very large. A MATLAB graphic of this data is available on request. (very large file about 300K) The formula for the rate of rotation of the star field is rate = (const.) X cos(latitude) X cos(azimuth) / sin(zenith distance) The rotation only goes to zero at azimuth 90 and 270 degrees (due East and due West) and on a line connecting these points. Even on this line the rotation becomes singular (infinite) at the zenith. At the pole the rate is 360 (approx.) degrees per 24 hours. At other points the constant must be calculated for the observers location. The rate of de-rotation required is seriously large so that only very short exposures are possible. Depending on you desire for perfection, only a few seconds up to a minute for some pointing directions. If the de-rotator is working correctly is will take care of all the calculations and turn the de-rotator at the correct rate as long as it knows where the telescope is pointing.

I am told that the de-rotator on the 16" Meade does this operation very well. (Note that I have not use one.) Note also that the telescope must be leveled and the correct location entered so that the telescope knows exactly where the pole star is. This is because the de-rotation is calculated on the basis of the known R.A. and declination and azimuth distance. So, the Alt/Azm setup cannot be more "sloppy" than that required for polar alignment. There are severe limits to the use of a de-rotator. The guide star must be in the field of the de-rotator and be de-rotated with the field being imaged. An off axis guider will work if it is rotated with the imaging camera. But off axis guiding is already difficult and with the guider rotating as well may require a triple jointed neck. The ST 7/8 cameras solve this problem by mounting the guider chip next to the imaging chip Thus it rotates with the field and will correctly guide when using a De rotator. This is a excellent idea but it does not solve the problem of field rotation in any other tubes or piggy back imagers that may be on the same platform. If a series of short exposures is satisfactory, a single chip could be used as a part time guider. Appropriate software would be required and short exposures would be necessarily acceptable. But single long exposures have better signal to noise ratios. So using a chip to share functions as a guider and imager does not seem like a good idea. Again, the shift and combine method used by SBIG is a good, if partial solution to this problem. There are also software image processing techniques that might be applied. In summary, if the guider points the telescope to the correct R.A. and declination and tracks a guide star accurately and the telescope computer knows the value of the pointing position, it will calculate the correct rate of de-rotation.

The guider tube and the imaging tube need only be aligned to the accuracy and with the rigidity required for the polar guiding case. But conversely, they need to be set at least with this accuracy and not less. Considerations for atmospheric distortions and cetera are the same for either setup. Neither Polar nor Alt/Azm setups are much simpler each than the other to achieve the same accuracy. In one case you need the wedge and in the other the de-rotator. One case is a well known solution the other, currently, somewhat unknown. It is my opinion that for a permanently mounted telescope the polar mount is the least problematic. That is because it is a simple, well understood and versatile solution to imaging guiding. For a moveable telescope that has to be reset for each imaging session I still think the polar wedge solution is the best. Alignment, leveling and the like are not that much more demanding for polar as for alt/azm setup I believe. I have chosen not to use a de-rotator and have opted for a wedge for both my permanent and trailer mounted telescopes.

Thus the following opinions by be somewhat biased. The de-rotation is limited to the main tube for which the de-rotator is designed. The piggyback image is not de-rotated. Use of a separate guider telescope is not possible. De-rotation does not work well near the azimuth because the rate of rotation becomes very large. At the zenith, the Alt/Azm mounted telescope is rotating rapidly on the fork axis in order to guide and the de-rotator is turning rapidly in the opposite direction to correct field rotation. The clearest part of the sky, at least 30 degrees wide, is lost to imaging. Of course the pole region is lost to many mounted telescopes. But it is not of much great interest in general. The de-rotator is just another complex mechanical mechanism with bearings motor, computer software and the like which can fail to work precisely. Some autoguiders seem to work best when the pixel axes are lined up with the axes of the telescope. With the de-rotator in the guiding path the situation is constantly changing. Which is I think not a good idea.

Again the ST imager software reduces some of these problems. My final conclusion is, do not use a de-rotator on a small telescope which can be easily mounted on a wedge. Since this was written, Meade has come out with a de-rotator for the LX scopes. I have chosen to pass on it for some of the above reasons and others mentioned below. There is considerable added extension to the rear of the telescope (about 75 mm). The attachment uses the standard Schmidt thread which is of small diameter (internal tube opening of about 34 mm) which vignettes 35 mm format. It would be fine for CCD chip sizes. For imaging that requires guiding, it must be done with an off axis guider that rotates with the imager. Or, of course, the ST 7/8 two chip method. In the case of the Meade LX series, the PEC does not work in Alt Asm mode. The angle of exclusion at the zenith is stated to be as large as 40 degrees from the zenith limited by the physical arrangement of the tube and or fork. There are in the instructions implications that one could get somewhat closer. I must point out that I have only looked this unit over, measured it and weighted it. I decided not to go the way of de-rotators so I do not own the unit and have not used it. Those interested in the field rotation problem when using a telescope in Alt/Azm mounting might find the following calculations interesting. I have calculated the rotation rates using the formulas given in Meeus The primary reference is Meeus, Astronomical Algorithms. Chapter 13, using the concept of the Parallactic Angle, explains rotation. The discussion is quite brief and thus not a clear as it might be.

A more convenient formula which is very easy to tabulate is: Angular rate of rotation = (a constant) X cos (azimuth angle) / cos (altitude angle) The constant is the angular rate of rotation of the earth times the cos (local latitude). For my location, Wisconsin, the constant is about 11.3 degrees per hour. One must be careful with this equation since there are several singularities. i.e. points where the cos goes to zero. The singularity causes a line of zero rotation going from 90 to 270 degrees azimuth, which is due East and West. This line intersects the zenith but at the same time the values of rotation at the zenith are infinite since the cos at zenith is zero as well. The tables, calculated with MATLAB, are attached at the end of this note. The first table is for altitudes up to 80 degrees and the second for angles from 80 to 89 degrees. Rotation at the zenith, 90 degrees altitude, cannot be calculated.

Conclusions from the tables that follow. The rotation rate is smallest pointing East or West and largest pointing North or South for a given altitude. When pointing at an altitude of 60 degrees, the rate of rotation can get to be 2 times normal (the constant in the equation above). At 80 degrees, the rate can go to 6 times normal drift rate. Above 80 degrees, as shown in the second table, the rate can get very large. The telescope is rotating on its vertical axis in order to track while the de-rotator must rotating the opposite way to de-rotate the star field. This is not to say that the problem of de-rotation is impossible but only to point out that movement of the telescope and de-rotator are fast and must be very accurate when pointing near the zenith. That is why, I believe, de-rotators are not generally used at pointing angles closer to the zenith than about 20 degrees or more. But, the best seeing is often at the zenith or at least high in the sky where de-rotation is the most difficult.

The following table is for de-rotation rates below 
80 degrees alt. The values are normalized and 
must be multiplied by the normal 
rate at a particular latitude.

The following table gives values above 80 degrees.
As can be seen,they get very large near the zenith.

    Field Rotation Plot

I hope this information and my ideas about pointing your telescope have been of some value. Sincerely -- Doc G

------------------------------------------------------------------

Subject: Derotator Advise (more)

From: Doc G, Date: March 2005

Believe me when I say that I wish the de-rotator had been more of a success. A decade ago I got my first 12" LX200 classic with a de-rotator and a Pictor 1616. I worked diligently for 12 to 18 months with this equipment and never got it to work well. The litany of problems was overwhelming. The equipment never worked all at the same time. I lost over a year of imaging with this equipment. Perhaps that is why I warn people to proceed with great caution when trying to get this equipment to all work together. There are significant disadvantages to the de-rotator when used with a fork mount.

After a year of struggle I got a wedge and an SBIG ST-7. Within one week I had dozens of fairly nice images even though in monochrome. This was the period, a decade ago, in which I worked very hard to tune up the LX classic to the point where it was suitable for imaging. I came to the conclusion that for small scopes, like a 12", the wedge was the way to go. I certainly hope that the new GPS and RCX scopes have improved mechanics and will be able to be more easily tuned up for imaging. I will not be trying this newer equipment since I have moved on to different equipment which better suits my needs. (Paramount, SBIG Canon, Takahashi and the like)

Tentative reports indicate that the newer scopes optics are excellent, mechanics have greatly improved and the software has also improved significantly. I really hope this is the case since Meade, with its pricing, opens the joys of astrophotography to a much larger group. The SBIG group is the premier astro imaging group in my opinion. But they also for the most part have significantly more costly setups. Paramounts, Takahashi, RC and big SBIG cameras. All big bucks. I one time did a survey of the SBIG group and found that only about 5 % of them use Meade scopes.

I think you will find an interesting and challenging experiment when you go to a de-rotator. I wish you the best of luck. I rather this you will find the need for mother luck to be on your side.

By the way, I do not really understand the problem with wedges at lower latitudes. I have seen Meade scopes used on wedges in Aruba which is at 12 degrees and I have a friend using one on a wedge in Singapore which is at 1 degree. I was just to the Winter Star Party and saw hundreds of scopes on wedges and that is at 24 degrees. Wedges work well at any latitude. They just look funny at very low latitudes.

Subject: De-Rotator Mod to Use with NGF-S  

From: Dave Dixon

The Meade 1220 de-rotator does a very nice job of solving the problem of field rotation for an ALT-AZ mounted LX200. At the same time it introduces problems in backfocus and flexure especially when heavily loaded with an OAG, CCD camera and autoguider and trying to use a JMI NGF-S zero shift focuser. However with one modification the de-rotator and NGF-S can be utilized together in a combination that reduces back focus by 30 mm and almost completely eliminates flexure. Figure 1 shows the de-rotator with the SCT mount replaced by a 2" barrel that slides into the NGF-S as if it were a 2" eyepiece. Figure 2 shows the assembly mounted on an LX200. The only negative is that the weight of the de-rotator becomes added to the other accessories that the NGF-S has to move when focusing. A combination of the de-rotator, a Van Slyke Slider II, Meade 416xt, and 208xt are right at the limit of my NGF-S to lift when at an altitude of about 55 degrees.

 Figure 1

Subject: Field Derotator Experiences

From: Radu Corlan <rcorlanprofis.ro> Date: Sept., 2000

In case any of the list readers wondered if they really can avoid carrying a 15kg wedge and still make photographs -- well, you can (that is, the de-rotator works). But "there may be dragons".

Made the first test with the de-rotator (#1120) on a 12'' LX200 yesterday. I used the following setup: LX200->de-rotator->off-axis guider->35mm camera. The scope was leveled carefully an daligned with a 12mm reticle eyepiece.

First, the good parts:

1. The up-swing limit is not as bad as it would seem. Although (for some unknown reason) the scope sets the limit to 40 degrees from zenith, in practice it is below 20 degrees. I can imagine having to have a tighter limit when using a camera that extends a lot towards the mount. But you get within 20 degrees from zenith even when a 35mm camera is positioned vertically.
2. The camera/guider doesn;t rotate that much. You can make 1 hour exposures without the guiding being much more difficult than without rotation.
3. This is the really pleasant surprise: guiding in dec is as smooth as the r.a. The dec drive doesn;t reverse any more (it just accelerates/stops), so the corrections are smooth and instantaneous. Although i didn;t use an autoguider (yet) i can tell the thing is "autoguider-friendly".

I tool some 20-min unguided shots, and the result is that star images make nice arcs, about 15'' in diameter. This is to be expected, and at the normal level of the drive gears. The circular pattern comes from combining the dec and ra drives' periodic tracking errors.

I also took some manually guided shots (10 and 20-mins). When guiding yu go for example w for a minute or two, then w and south, then south, then e and south etc - as expected to cancel the circles. I took the shots in the Pleiades region, so i have enough bright stars in the field to check the tracking. All came out nice and round, and without signs of firld rotation.

>From the above, you can see i'm sold to the de-rotation process. But now comes the unpleasant part: the de-rotator unit itself. When i got the telescope, i was a little disappointed. The construction and finish weren;t what i expected in a quality optical instrument (look at a real microscope or similar piece of gear, and you will know what i mean). But it really is "a lot of telescope for the money" and the overall design is optimised, (at least for cost ;-> )

But the de-rotator is another matter. First it's unnecessary bulky. It has a heavy steel cover, and eats way too much back focus unnecessarily. On the 12'' it could have been mounted directly on the 3'' thread, flush with it's baseplate, and still clear the scope. That alone would decrease it's backfocus by nearly 4cm.

Second, when i first unpacked it, it had about 0.5mm free play in the rotating bearing - imagine your camera flipping up and down 0.5mm! And it had play in the rotation direction.

I opened it up (at least it's easy to assemble/disassemble).

The unit is built on a backplate. On the telescope side, there is an SCT thread coupler (way too tall IMO). On the other side, there's a circa 4'' dia worm gear rotating on a bearing. The "output" coupler is bolted on the gear. The gear is turned by a worm, which is driven via a totthed belt by a gearbox reduced stepper. The stepper and gearbox are a single unit (and look mighty strong - way larger than the LX drive motors).

The rotation-wise play was easy enough to fix - the worm's position was maladjusted (or should i say unadjusted?) Pushing it a little against the worm gear did the trick. But when i took the worm off to check the worm gear play in the bearing - surprise! Not only was the play in the bearing _huge_, but the wheel felt rough when turning.

I suspected that the wheel had some sort of nylon bearing, and they forgot to put the bearing element in altogether. So, off come the many screws that hold the thing together. It turns out that thegear is rotating on a ball bearing. But not a preassembled bearing, but rather one built from the gear itself and the hub. The gear has a semi-circular race machined on the inside, and the hub is made of two beveled parts which screw together. When fitted, they form a "v" shaped track. A lot of loose bearing balls complete the bearing.

The rough rotation's cause was pretty evident: all the beaaring area was full of aluminium shavings. Fortunately, the races themselves were smooth and even - the shavings must have come from the many threads that are around. But the play was just grossly imprecise machining! Not having the required gages around, i couldnt determine if the play came from the semicircular race in the worm, or just wrong diameters. In the latter case, it can be cured by machining the hub halves, making the "v" narrower, and thus pushing the balls towards the gear. If the race itself is the problem, the thing can still be fixed by using larger balls.

Until i get to do one of the above, i mounted some Teflon pads underneath the gear, and the play is gone. Still, this is obviously not a final solution.

Does anybody have better experience with the de-rotators? I may "just" have had a bad unit.