In moving-mirror SCT designs the primary is affixed to a tube, this tube is a close fit on the outside of the primary baffle tube upon which it slides, the primary mirror being maintained in axial alignment by it. The slide-tube and primary baffle tube are critical components of the focussing system, and are definitely the weakest point in the moving-mirror design. In order to focuss smoothly this design requires a minimum clearance between the two tubes to prevent binding, and also to provide simple mechanical clearance for the moving parts (for those with no engineering experience, a 1" round bar will not slide into a 1" round hole!). This 'working clearance' gap is normally filled with a special viscous lubricant which (when new) does a reasonable job of preventing the mirror moving from one side to the other of it's own accord (within the limits of said clearance). Unfortunately, only a very slight movement of the mirror is enough to cause a noticeable shift in the image position, and this becomes worse over a period of use as the normal focussing action causes the lubricant to be scraped off and deposited at the extreme ends of the primary baffle tube. A quick 'n dirty remedy for significant image shift can sometimes be achieved by racking the focussing mechanism from one end to the other, thus re-distributing the lubricant (in theory anyway, I had no success whatsoever with this). Altogether, this is a quite unsatisfactory mechanical arrangement for astrophotography, although it is not a problem for purely visual use. The image shift produced by mirror movement is difficult to erradiate completely, and is the major contributor to 'differential flexure' between the optical system of the main scope and a piggy-backed guidescope. A way around this is to use an off-axis guider which compensates for the image shift, but at the expense of increased difficulty in locating and tracking a guidestar. Note that the LX200 is no different in suffering from this problem than any other make of moving-mirror SCT, even expensive telescopes like Takahashi suffer from it to some degree. It can be reduced by using close-tollerance bearings on the slide-tube, but never eliminated without extensive design changes.
|
Diagram (not to scale) of how the modification works. |
This is my second try at reaching a mechanical solution to the problem, the first being an attempt to stabilise the mirror using a ball-bearing support attached to the front of the mirror slider tube (an ill-conceived idea which was a notable failure). Recently, I had a re-think about the whole strategy and came up with an alternative solution. It seemed to me that trying to tackle the mirror movement by modifying the slider tube was hopeless - trying to control flexure just an inch or so from the center of rotation was doomed to failure because of the substantial leverage exerted by the weight of the mirror. A far better strategy would be to support and guide the mirror at it's periphery. However, looking at the OTA's construction, providing support at the edge of the mirror would be difficult, the mirror does not have a full-diameter machined cell to work with. The old 'shipping-bolt' mirror-locking trick (described elsewhere) seemed to be reasonably effective for locking the mirror in place (though it is not much good for guiding the mirror along it's optical axis), and this supports the mirror at a radial distance of just 2" from the primary's center. The major drawback with this method is that the support is only at a single point, and there's little one can do other than thread in a 1/4" x 20 bolt to lock it up with a nut after focus has been achieved, or to lock it in place first and use a secondary focussing mechanism (the JMI Crayford focusser for example). Disabling the focusser in this way was not much of a solution to my mind, OK - it did solve the mirror 'flop' during long exposures, but it did nothing for image-shift, it restricted the focus travel, and is basically a kludge.
I felt a more radical approach wan needed and thus decided to use three seperate support points some 4" radially from the center of the primary (that's as close to the edge of the mirror as I could work with). The supports would be attached directly to the rear surface of the mirror using silicone sealant (the same stuff used to mount secondary and primary mirrors in large Newt Dobs), and would consist of three 1/2" diameter rods sliding in close-fitting reamed bores. Each of the three rods would be spring-loaded (to remove axial slop which results in backlash and defocus mirror 'settling'), and all three could be locked with thumbscrews in any position the length of the primary's travel to provide absolute and definite rigidity.
The three housings containing the reamed bores could conveniently by mounted on the outer face of the rear cell, this being machined quite flat and thus providing a good datum surface. Indeed, without this machined face the job would have been very difficult. It was critical to success that the bores are parallel and true to the main mirror's axis of travel, if they were not then it was inevitable that binding would occur. The rear surface of the mirror at the point of contact is unfortunately not flat (the edges are thinner to reduce weight), and this was a complication - but not an insurmountable one. An angled face could be ground on the end of each piston to approximately match the angle of the glass surface, and I then relied on the silicone sealant to fill any gaps. Using silicone had one other advantage, it's slight flexibility reduces the possibility of any stress being applied to the rear of the mirror (which might cause distortion of the mirror), yet the joint is firm enough to provide the necessary support. Care was taken that the steel end of the pistons do not actually contact the glass itself, a thin fibre washer embedded in the silicone separates the two.
The main mechanical issues were the following:
There was the minor detail of deciding how to mount the three housings onto the rear of the mirror cell, and also drilling the 5/8" holes for the pistons to pass through. This is a messy job and not one for the faint of heart - this is not the time to make a mistake, and a 5/8" hole through the primary mirror would probably not improve it. First, the positions for the three holes were marked out at 120 degree intervals, and the LX200 attacked with drills of increasing size up to the big 5/8" machine drill. Care was taken not to let the chips fall into the electronics (or the gearing of my electric focusser), but there was nothing to be done about chips falling down inside the OTA. This is not really a problem, just finish all the drilling then take the corrector plate off and vacuum it out. All three support housings had been made (everything from stainless steel, including the springs) before the drilling commenced. In the first of the pictures below you can see a mounted housing on the left, and the 5/8" hole together with the three 4mm threaded mounting holes for the the right-hand housing. The bottom hole has yet to be drilled.
 |
At the left the first of the three piston housings is seen mounted in place, on the right the 5/8" diameter hole and three threaded 4mm mounting holes for the second. |
 |
Here all three housings are in place, the one on the left has it's endcap removed so you can see the end of the piston and the compression spring. Note the clamp screw on the left side of this housing. |
In the picture below can be seen the components of each support, the piston at the top and the housing below. The piston has a 5/8" diameter cup-shaped washer on the end, this holds a blob of sealant when finally fixed in place. The compression spring runs inside and all the way to the end of the piston to provide even tension over the mirror's approximate 1-1/2" travel.
 |
Component parts of the mirror support/guides. |
All three pistons were cemented in place at the same time. The mirror-end of the pistons were first prepared by having a fibre washer cemented on, the silicone being used to form a 1/8" thick 'ring' of material which would isolate the piston from the mirror. This was allowed to cure for 24 hours. The mirror was moved all the way back then forward again just half a turn, then a blob of sealant was placed in the depression in the plate at the end of each piston and the three assemblies were all bolted into place on the rear cell. The top of each piston had previously been marked to indicate the orientation of the angled faces of each piston. The springs provided pressure to keep the piston in close contact with the mirror whilst the layer of sealant cured (the corrector plate being removed for this job to ventillate the acetic acid released by the sealant). When set (24 hrs later) the mirror was checked for free movement by winding it all the way forwards, all the time looking (i.e., feeling) for binding in the focusser. With the pistons assured of being in the correct position in the primary's most rearward position, it was then most unlikely that binding would occur in it's most forward position unless the mirror moved significantly off-axis over the 1-1/2" or so travel (a highly unlikely condition). That's also why the pistons were cemented in place with the mirror at the rear position. The pistons were lubricated with molybdenum-based grease (necessary to inhibit wear because none of the parts are hardened) before the endcaps were finally bolted up. It was found necessary to drill a tiny vent hole in each cap to allow compressed air to escape as the piston moves.
 |
Finished mod - rear view. |
 |
Finished mod - side view. |
Addendum:
The springs shown had to be replaced with lighter gauge ones as the pressure exerted was unnecessarily high, and in addition, spring guides (or shrouds) were needed as the 1-3/4" free length at maximum extention would cause them to buckle as they were compressed. The spring guides eliminated this problem. The mirror now moves very sweetly, you wouldn't know any modification had been done other than the 3 housings sticking out the back. There's also no mirror 'settling' (focus-shift that can occur due to the weight of the mirror causing it to 'settle' backwards) and no image-shift (the pistons take care of that). I can't test for mirror-flop until the weather improves so I can take some long-exposure photos but I simply cannot see how the mirror can move with the thumbscrews locked up. We shall see.
I had my first chance to check the performance of the mirror modification on the night 27/28th March, it was clear all night but the seeing was poor (4/10). The cooling fan I installed at the same time really helped, it took only 20 mins to reach equillibrium - certainly, the image didn't improve any during the following 6 hours or so.
The first thing I did was to check the effect of using the 3 locking screws in the 'partially-locked' mode. All three screws have a Teflon insert which bear on the three pistons, my thoughts were that these could be used to remove any play at all in the system and the mirror would still be able to focus without binding. In practice, this merely serves to make mirror-shift worse. I attribute this to the fact that there is still clearance at the front of the slider tube, and the whole assembly (mirror/slider-tube/primary baffle) will twist under pressure from the focusser. If the lockscrews are released then no image-shift is visible.
Next, I checked the effect of tightening the locking screws on a pre-focussed image. A star was very carefully focussed (as accuratly as I could using my 2-hole mask) and then the image compared after the lockscrews were tightened securely. There was no change of focus or change in image in response to tightening the screws, the lockscrews did not cause mirror distortion nor did they cause the focus to move. I should note that the poor seeing meant this assessment was not as critical as it needs to be, star images were not pin-sharp for this reason (confirmed in my 12" Newt).
The next test was to try slewing to various parts of the sky with the mirror locked to assess possible changes in focus. Doing this normally causes the focus to change due to weight-shift of the primary mirror. I first focussed at about 45 degrees DEC and then checked star images for critical focus at various points on the horizon and as near to the zenith as I could get (the particular CCD setup I used would not reach the zenith because of clearance problems with the forks). There was no discernable change in focus that could not be attributed to diffraction and/or poor seeing near the horizon. Again, a more critical test will have to wait for a night of better atmospheric conditions.
Further tests then followed with CCD exposures (autoguided) lasting up to 30 mins of objects approaching, crossing, and falling away from the meridian. Such exposures previously exhibited oval stars (and was the main reason for making the modifications). There was absolutely no evidence of trailing in the resulting images (my refractor guidescope mount are solid enough not to contribute to differential flexure). However, some 3 hours into the imaging session I noticed a perculiar effect on star images as illustrated in the 30 minute MX5-C image of M66 below:
 |
Tri-lobed stars surround this 30 min MX5-C colour image of M66. |
The stars images had become tri-lobed, clearly an indication that the primary was under mechanical stress and the three pistons were distorting the mirror slightly. Upon releasing the lockscrews the image returned to normal. I suspect the reason for this was that the ambient temperature was still falling sharply at the begining of the session (in fact the temperature dropped below freezing a couple of hours later) and thermal contraction had caused the pistons to differentially shrink in length compared to the rest of the focussing mechanism. Further, I had not taken the trouble to back off the focusser tension, nor did I finish focussing in any particular direction. After releasing the tension on the lockscrews, re-focssing and then re-tightening them, the problem did not recur. This problem certainly requires further investigation, but it illustrates that the the mirror was securely locked! In the worst case I may have to replace the pistons with new ones made from low thermal-expansion material - or rather, material of similar properties to those in the OTA (I have Teflon, Nylon and various other materials I could try), but I also suspect that it's merely a matter of technique, and relieving the tension on the focusser after locking the mirror up will likely resolve the issue.
The scope has now been used three more times since the modification to the primary mirror, one was an all-night session, the other two 4 and 5 hours. For some reason there has been no recurrence of the 'tri-lobular' star images, perhaps because on none of the subsequent tests did the temperature fall quite so quickly and by so large an amount after the mirror had been locked in place. Also, the scope had been thoroughly cooled to ambient using the fan I installed recently. I can confirm that in order to minimise 'mirror-shift' when focussing the three locking screws must be loose, but under those circumstances it is not significant - i.e., it does not interfere with the primary function of the telescope. An example: with the 416XT with f/6.3 focal-reducer, selecting a focus window some 1/6th of the full resolution image, focussing back and forth causes the image to shift by about 5% of the window width (OK, not exactly a specific measure, but it gives something to compare to!). It is clearly still detectable though, and I suspect the spring loading in the pistons is responsible in some way (the 'shift' is not in a straight line but curves a bit). I do know the three springs are not identical so the loading will not be even. I will have to try removing the springs to see if there is any improvement, but even if there isn't I can live with it.
After further testing I cannot detect any trailing due to mirror-shift once the locking screws have been tightened. From that point of view the modification is a success. Nevertheless, that doesn't mean the focus cannot change! There are other factors at work that appear to affect focus, and these are significant for high resolution photography. I found that after focussing critically on a star at about the meridian (actually, it makes no difference where the scope is pointed in practice) and locking the mirror down, it is then possible to slew to any point in the heavens and the focus remains correct. I tested this with the Hartmann mask and CCD camera, no change in focus was detectable using this method (note that my mask has two 3/4" holes at the maximum extent of the 10" aperture and is thus very sensitive). When I checked focus some 2 hours later I found that the focus had changed measurably (sorry, I can't recall whether it was slightly inside or outside of focus). The extent of this focus change could easily be detected using the hole mask, the image of a star appearing elongated instead of circular, but not so bad that there were two separate images. I think this change in focus is also due to the temperature falling during the observing period (again, sorry but I did not record the magnitude of temperature drop, but it was at least 5 degrees and might well have been much more - it did not drop below zero but was enough to cause me to go back inside and put extra pairs of socks and another jumper on!). This happened during the period 9:00 p.m. to 11 p.m., and after that did not seem to occur again, I checked by popping the Hartmann mask on at 30 mintue intervals. The focus change possibly stopped because the ambient temperature had stabilised by then. This is an important finding though, because it is a significant change in focus and it will increase the size of star images on film. The (partial) remedy is to check focus immediately prior to each extended exposure, and if 'short' exposures (say, 15 minutes) are the norm then it will be necessary to re-focus at about 1 hour intervals whilst ambient temperature is still changing. It's also important to ensure scope cool-down is complete before starting imaging, this would only make matters worse.
Michael Hart's article on the effects of temperature on focus in SCTs
Back to Telescope Accessories page