• ASI Mount
  • Getting the best performance from my AM5

Jan75 My own observation is that OAG makes guiding performance worse, but at this point I don't know why that is.

With a strain wave mount, I suspect that it is because

  • you need guide exposures of 0.5 sec or shorter
  • this means that you will need multi-star guiding
  • in turn, this means the need for a good star field with all the stars in the field being as sharp as possible (this is where an OAG falls apart). If one star has coma and the other star does not, or another star has "wings", as the weights of their centroid changes over time due to scintillation, the weighted sum will also change.

Even before strain wave gear mounts, I have tried various OAG over the past 15 or so years, and have always gone back to using a separate guide scope.

Chen

    w7ay This might be the key to the entire problem of bad guiding performance. Can't wait to re-run all tests with a guidescope. I used to use it before I bought the OAG but I also had the ASI ZWO 120MM camera that is not as sensitive as the 220 so pairing the 220 with the guidescope will be a new experiment. Now that I use a counterweight I am no longer that worried about the weight (or torque) of the main telescope + guidescope group. Just wondering if a 246mm guidescope is a good match for a 1627mm main scope... Will see!

    • w7ay replied to this.

      Jan75 Can't wait to re-run all tests with a guidescope.

      With an RC8 type focal length, you will need to make sure there is no differential flexture, though :-).

      That being said, I have been able to get consistent 0.35" type total RMS guiding with ASIAIR, albeit with a different manufacturer's strain wave gear mount, with just a 250mm focal length guide scope with flattener and ASI178MM guide camera (large, clean FOV).

      That type of error should work well enough for the RC8. I have started to tune the autoguiding to prepare for when a Mewlon 180 arrives later in the year, although most of the time, it will be used unquided (planetary).

      Your 246mm should do fine.

      Good luck.

      Chen

      I have an RC6 on the AM5 at native 1370mm fl. Guiding is PHD2, not ASIAir.

      When it was being used on a CEM, OAG setup of Askar OAG (10mm prism) plus 174mm mini worked very well with guide exposures of 1.5-2 sec.

      On the AM5, I have tried 0.5 to 1.5 sec guide exposures and settled at 0.8 sec. Location is in a metro, so high LP (skies around 17 mpsas).

      The OAG works fine in star rich areas at 0.8 sec. I get 8-9 stars in multistars. Where star fields are poor (Virgo, etc), or near the light dome, I get only 1-2 stars, maybe 3. You might get more if you have low LP.

      I then tried a 60mm F4 guidescope with an old ASI385 I had lying around. With this, there are always 9 guidestars and I feel the guiding in star poor areas is better - lower by about 0.1"-0.2" rms total. Not significant when imaging at 1.6"/px, but desirable at 0.55"/px.

      In most areas, the difference in guiding doesn't show up in the images. But in the star-poor areas, it does. So I just use the guidescope now.

      I could probably go down to 0.5sec as well. I fiddled around with the minmove - had to reduce it compared to OAG.

      The guidescope is in rings which are bolted to the top vixen dovetail. The guide camera and imaging camera are well aligned - both put the same star in the centre.

      The 220mm actually has larger pixels than the 385. That would push the guiding pixel scale away from the imaging scale. I am planning to try a smaller pixel camera (a 290mm I have) and see if that improves it further as the guiding pixel scale will come closer to the imaging scale.

      Roughly speaking, the following is what I find with ASIAIR multi-star guiding with my strain wave geared mounts. ASIAIR uses signal-to-noise ratio (SNR) weighted centroid averaging (https://en.wikipedia.org/wiki/Weighted_arithmetic_mean). Autoguide programs that use uniform weighting works better, but if you are using ASIAIR, you have no choice.

      • more than 1 second exposure -- limited by the slope of the mount's periodic error -- basically limted to no better than 0.75" total RMS error, even if "seeing" were perfect (hah! only happens on Mauna Kea),

      For 0.5 second exposures:

      • 1 star: guiding is limited by atmospheric turbulence ("seeing"). Typically 2" RMS per axis, even for an average night, if you are at sea level.

      • 2 stars: RMS error is a factor of about 1.4 (== sqrt(2)) times smaller than the 1 star case iff the two stars have the same ADU. If one of the stars has an ADU of about double that of the second star, the SNR weighted scheme in ASIAIR will only reduce the RMS guiding over the one star case by about 12%.

      • 11 to 12 stars with the brightest star also picked: again, depends on the distribution of the guide stars. If one star is much brighter than the other stars (say, Vega is picked as a guide star), ASIAIR's SNR weighting when computing the centroid again makes the RMS error no better than a single star case.

      • 11 to 12 stars without an abnormally bright star picked: equivalent to about 3 or 4 stars. RMS error due to "seeing" is about half of using one star. A 0.5 second exposure equivalent to using 1.5 to 2 seconds of exposure using 1 star -- i.e., it starts being usable.

      • 12 stars plus an additional 10 dB camera gain after achieving 12 stars -- the brightest stars should now be rejected by ASIAIR because it has become saturated. If the brightest two or three stars are rejected, this might give the equivalent of using 6 to 9 stars when the weighting is uniform instead of SNR weighted.

      • 12 stars plus an additional 20 dB camera gain after achieving 12 stars finally, you are getting the benefit of using the centroids from 10 to 12 stars. A few of the brightest stars should now be rejected, making the rest of the stars have more equal weights. The limiting factor (under decent "seeing" and no wind) of about 0.15" to 0.25" RMS, so the error is most likely limited not by ASIAIR centroid computation, but limied by the mount itself. Notice that an OAG (or a heavy filter on the guide camera) may not be able to achieve this.

      • 12 stars plus an additional 30 dB camera gain after achieving 12 stars -- you are likely to hit high noise because of the excessive guide camera gain at this point.

      Notice that ZWO cameras uses 0.1 dB per gain unit. So, 10 dB of extra gain means adding (adding and not multiplying since the gain units are already in the logarithm scale) a gain of 100. 20 dB of excess gain means adding 200 to the gain.

      Given a night with good transparency, my 55mm/250mm focal length guide scope and an ASI178MM guide camera typically gives me 11 to 12 stars with a gain of about 60. I do have a flattener (but not reducer, to keep the focal length at 250mm) on the guide scope to get clean stars all the way to the edge of the guide sensor -- distorted stars will only hurt as you add more (since the SNR weighting will jitter the centroid), so make sure you have a decent guide scope with a flat field. I typically set the gain of the ASI178MM camera to between 250 to 300. When transparency is not so good but "seeing" is still respectable), I add even more gain; but at some point, the guide camera ends up being too noisy.

      To beginners and YouTube watchers: do not blindly use the gain that I use. Your gain will be different from mine. What you need to do is to slowly increase the gain of the guide camera until you get between 10 and 12 stars. Then add an additional gain of 200 to work around the ASIAIR SNR averaging problem. You may need a gain of 200 just to get 12 stars (in which case adding an extra 200 would make the guide images too noisy -- you would need a better guide scope). Ditto OAGs, if you can't get 12 stars at a gain of 200, you will only be able to add 10 dB of additional gain before the camera noise becomes intolerable. And in some cases, you won't even be able to get 12 stars with an OAG to start with.

      Chen

        w7ay what about when using a low end camera like the asi120mm mini where unity gain is at 28. I use one of those plus a 290mm mini where unity gain is at 110. I have no problem cranking up the 290mm mini gain to 210 but the 120mm mini maybe useless and a stretch at gain 128 and currently I find gain 70 is about max. Both cameras are on 240mm scopes. I may try a bit higher as I do notice that asiair does use a few overly saturated stars still.

        • w7ay replied to this.

          Kevin_A when using a low end camera like the asi120mm mini

          Nah, I stay away from bargain basement stuff when it becomes the limiting factor of a multi-thousand dollar (or Loonie) imaging set up.

          Cost wise, my current primary mount costs the most, then comes the OTA, then comes the imaging camera, then comes a set of filters. My guide system is already pocket change, comparatively.

          Favorite guide system right now: Borg 55FL objective, Teleskop TSMPT60 flattener, Astronomik L1 filter, ASI178MM camera, Borg helical focuser belt driven by ZWO EAF motor. When experimenting on Indigo, the QHY678M is swapped in, in place of the ASI178MM.

          Even the above (with the ASI178MM) is often star starved away from the Milky Way when using SNR-weighted centroid averaging.

          Chen

            w7ay i was looking at the 220 mini but the pixels are just too big. My scopes produce fairly good round stars but the choice of zwo guide cameras all have pixels 4uM and up now and the 178 wont fit my scopes. I am not sure why the ditched the 290mm mini with its small pixels.
            I still think that its not a camera issue but a asiair guiding issue that is too finicky with its devices.

              Kevin_A I am not sure why the[y] ditched the 290mm mini with its small pixels.

              Probably because the SC2210 (Chinese) sensor in the ASI220 is cheaper than the IMX290 (Japanese) sensor. It always comes down to nickel and dime with ZWO.

              This is why I chose the QHY678M (mono) for my Indigo guide experiments. I have learned long ago never to use a Bayer camera for guiding, since a star at the boundary of a sub-color pixel causes sudden centroid changes, even if you were to mono-bin it. Worst for 1-star guiding of course, but the SNR-weighted stuff in ASIAIR does not help.

              220 mini but the pixels are just too big

              Can you imagine that camera with a short guide scope and a MinMo of 0.2 pixels? Yeehaa!

              Chen

              w7ay what guide rate did you use to get these guide results for all the exposure lengths and different star saturation/number results? 0.5x or less?

              • w7ay replied to this.

                Kevin_A what guide rate did you use to get these guide results for all the exposure lengths and different star saturation/number results?

                Stock ASIAIR v2.1.1 guiding. 55mm Japanese objective (Borg uses fluorite glass from Canon), 250mm focal length. ASI178MM.

                0.5 second exposure (2 FPS), 0.25x sidereal guide rate, Astronomic L1 filter (which cuts a little more UV than a typical UV-IR cut filter, but probably no more than 5% to 10%; but helps with any UV star bloat).

                Max declination pulse duration of 30 ms (not typo -- thirty milliseconds -- since polar alignment of the order of 15", so there is really negligible declination drift), and max RA pulse duration of 20 ms (this is with the RST-135 'e', not the barebones RST-135, which would require more than 80 ms). Declination and RA feedback aggressiveness of 10%.

                Focused to achieve HFD of about 2.3 pixels (measured with ASIAIR detect star utility, at 3 second exposure).

                In addition to the Astronomik L1, the flattener glass also cuts some light, but not enough to make me go without a flattener. I aim for wide FOV with good star shape to the corners -- with multi-star centroids, you do yourself no favor with a wide FOV if some of the stars has coma or look like seagulls. You want the average centroid to remain constant as the stars scintillates and the centroids of different stars carry more weight. This "moving centroid" problem is also present if the polar alignment is not perfect and the guide stars field rotate after an hour or two -- PHD2 does not handle field rotation -- if only handles ∆x and ∆y image shifts -- i.e., Galilean Transformation.

                Chen

                  w7ay I am still waiting for asiair to offer 0.25X guide rate for the AM5 to tame the saw patterns but I am going to try tonight with a much higher than normal gain at 0.5s and see if that stops it picking 1-2 overly saturated stars with 10 normal stars. I spent a night refocussing all my guide scopes the other day and going to try a few setting changes.

                  w7ay I used a much higher gain last night on my poor old 120mm mini from a unity of 28 to 85 and it still tried to pick 2 huge stars. So I bumped it up to 100 and still it grabbed 2 huge monster stars.
                  So I ended up going back down to 85 and ran 0.5s exposures at 200 max durations with 35 aggression. Guiding was choppy in the 0.5 to 0.6 range all night long and the main star HFD was moving around like a mexican jumping bean as there must have been upper atmosphere turbulence even when apps said there was none. My star was going from 2.8 to 3.8 back n forth like a wild caged bird. Image was ok for 3 hours in a full moon but I was using my small FRA300.

                    Kevin_A 0.5 to 0.6 range

                    well, at least - I just gave up after three hours, my AM5 does not guide under 0.7 and that is "good" already. I'm mostly around 0.85. What I don't understand is that guiding never seems to settle. It jumps around, both axis, for no apparent reason. I really don't get it. I've tried pretty much everything one can do in PHD2. I was able to improve things for sure, but never better than mentioned above.

                    The NEQ-6 I used for 12 years always was in the 0.6 range, sometimes down to 0.5 and that over hours. No wild jumps.
                    I really regret buying the AM5 :-(

                    Talked to the dealer I bought it from. Yes, we can "repair" it under warranty, we just have to send it to China for that.
                    Yeah, great. Surely it won't take longer than two or three months...and the outcome is totally open.

                      CHriss I know what you mean about jumping around. Mine is not very consistant and even when it is guiding well my guide star jumps around like crazy all the time to the point I start to believe my skies are constantly turbulent. My guiding graph is never smooth either and I have never seen it flat for long periods. It always has zigzags. My previous iOptron mount ran 0.9 at best so I think 0.7 is ok for my image scale. I would like to see guide rate adjustability in asiair to use a lower 0.25X guide rate as that should smooth out the zigzags. Using the beta version my guiding went from a usual 0.7 to 1.2rms so i reverted back.

                      CHriss It jumps around, both axis, for no apparent reason.

                      Do you see any correction pulses being issued just before your jumps?

                      If you don't see any correction pulses before the "jumps," and if no one is walking close to the tripod, then it comes down to the mechanical construction of the mount. It could be the Chinese strain wave gears, or it could be the pieces that ZWO had added to it (shell, pulleys, belts, bearings, etc).

                      Any centroid estimation error will for example cause a jump, by issuing a bogus correction pusle. But if you don't see a correction pulse before the jump, it is not coming from software, but coming from the mount construction.

                      Unfortunately, if it is the mount construction, I doubt it can really be easily "fixed" without replacing it with a complely new mount (and you enter the next lottery).

                      (Notice that they are trying to shot gun the problem with software "fixes" that don't even make engneering sense. Ergo, I don't think they have a clue what the problem is -- and you are not alone either.)

                      Chen

                        w7ay

                        best to show a typical example, taken last night.
                        Here is a 10min log:

                        As you can see, guiding generally is rough. Total RMS 0.91, something the NEQ-6 would have produced maybe in the 10% of worst conditions one would even consider to fire up the telescope.

                        The part marked with the yellow X:

                        After the dither, guiding was running fine for a minute or so, something I would expect to see most of the time.
                        But then again, the violent jumps, for no reason.
                        I have tried many combinations now of aggression, hysteresis, min move, both axis, spending hours on that.
                        It really makes not much of a difference. The general behaviour is the same in the end.
                        What seems to work best (with sidereal set to 0.5x) is:

                        X guide algorithm = Hysteresis, Hysteresis = 0.150, Aggression = 0.350, Minimum move = 0.300
                        Y guide algorithm = Resist Switch, Minimum move = 0.300 Aggression = 25% FastSwitch = enabled

                        But "best" is most relative here, a RMS of 0.8 to 0.9 certainly is nothing to be happy about, especially taking the price of the AM5 into consideration.

                        Note the mount and scope are in an observatory, there is no wind or external influence of whatever kind. And all sits on a concrete pier, decoupled from the floor. No way to take any influence, even if you jump up and down right next to the mount.

                        I don't get it, really. Either my mount is a total lemon or the AM5 generally is flawed.

                        • w7ay replied to this.

                          Just one note about the pixel size of the guide cam. The ASI ZWO 220 mini has a pixel size of 4 µm. The 120 mini is at 3.75 µm and the ASI ZWO 1600 MM Pro (for example) is at 3.8 µm. I struggle to understand how a .25 µm/pixel could really make any difference at all. The 220 is more sensitive in the IR spectrum. I'd argue that it's better to have more stars rather than tiny pixels anyways. Am I wrong?

                          CHriss But "best" is most relative here, a RMS of 0.8 to 0.9 certainly is nothing to be happy about, especially taking the price of the AM5 into consideration.

                          As I have mentioned many times, a typical worm geared mount will outperform (w.r.t. autoguiding) a strain wave geared mount that is at least 3 times more expensive. The price of the strain wave gear mount has nothing to do with guide performance (and ZWO may be worse than folks who know how to machine precision mechanical parts); you are paying for other conveniences (such as portability) and not precision.

                          No one should ever buy a strain wave gear mount if they don't understand that price performance aspect of stain wave gears. Don't listen to the YouTube shills, it is not their money; and you never see a YouTube shill tout a 10Micron mount, have you? That should tell you something.

                          That being said, if you are getting 0.9" typically and ZWO advertises on paper (or web page) that you can get 0.7", that should be enough to return it and get your money back. Even if 0.9" is good enough for you -- you did not get the mount that you had paid for. If they use weasel words such as "typically 0.7", then keep that in mind the next time you need to buy something ("fool me once, shame on you, fool me twice, shame on me").

                          Just one note about the pixel size of the guide cam.

                          From the first guide graph that you showed, the large errors are not caused by bad correction pulses. It looks like a mechanical problem. So, not need to look at pixel scales etc. Look for lose bearings, bad belts, etc.

                          A telling point (which you did not mention earlier) is that the large error spikes are periodic. They occur every 110 seconds or so (you probably can do more work to get a better average using more cycles than is shown in the above graph).

                          Now, the problem with the above is that it happens on both the RA and the declination axes at the same time. They are independent and orthogonal motors. So, that appears to preclude the strain wave gears themselves as the source of the problem, but some mechanical defect in the rest of the ZWO mount.

                          Did you send the graph to your dealer? They should immediately refund the cost of the drive since there is something so obviously wrong from the large periodic spikes. If not, just dispute it through your bank/PayPal and do a stop payment.

                          The problem with the ZWO mounts so far is that everyone seems to have a different problem with it. The design does not appear to have been vetted by an engineer, with proper limits in place for QC to take defective products off the assembly line before shipping to customers.

                          Chen

                          Regardless of the issues (and Chen, you're right, there seems to be a plethora of them from what users report - thus not one problem to fix, rather a more systemic QC process improvement, as you mentioned), what is the general rule of thumb to choose a guide scope for a given imaging scope? If I use a 80 mm and 328 mm guidescope from TS, for example, and a ASI ZWO 220 mini camera (4 µm pixels) this yields a resolution of 2.52"/pixel. That looks like heavy undersampling to me when the goal is sub-1" guiding, so what am I missing? If the main imaging train is around 0.4-0.5" how can it work without OAG? And yet it does work and in fact, from the above, the number of guiding stars is more of a factor (bad point to OAG) than resolution, so it seems.