Guiding Issue In TSX-Is This a Valid Experiment?

[PRejto (Peter)]

I'm not having any major issue in guiding with my MEII, but I've always noticed that the X and Y plots in the guiding graph look different. Since I normally previously aligned my guide camera to North up (Position angle =0) the X plot is aligned with RA and Y with Dec. My thinking has always been that this would make it easier to diagnose an issue with a specific axis. I still think this is generally true. Differences that I saw I would assign to a particular axis. That seems reasonable and is common practice.

Then, due to the fact that CCDAP doesn't seem to like the guide camera positioned this way, I moved it to a PA =45. Now neither axis is aligned with RA or DEC. At first I thought this was a pretty bad arrangement but then I had the thought that actually this might prove quite useful in thinking about issues that I see and many others are trying to talk about on the Software Bisque Forum. The issue seen is that of random spikes in guiding seen pretty much mostly on the guide graph in one or the other of the axis. Now, with the camera aligned such that X=RA, one might assume that a spike in RA tracking is caused by some crud on the worm or something like that. However, with the camera at PA=45 PE, crud etc would cause a spike in both X and Y, not just one or the other.

Just the other night, with my guide camera at PA 45 I had very tight guiding along X, but major clumping and numerous spikes along Y. Sometimes a spike would show on both X and Y but that might happen only 1 in ten or more times. I cannot see how to reconcile these observations. Any mechanical issue from either axis show show a movement in both X and Y. Atmospheric movement should on average move a guiding star around randomly. If all is well I would expect to see on average pretty much the same thing on both X and Y. I don't see this so I can only conclude that my thinking about this is incorrect, or there may be an issue about the way guiding corrections are implemented in TSX. Others are speculating that there is some problem of the interaction between Protrack and guiding. I have attempted to improve what I'm observing by changing just about every parameter I can think of including exposure duration, delay, different guide star, aggression, min/max moves, guide correction rates (in Bisque TCS). The effect I see does vary with where I am guiding. It's best on the meridian and gets worse as I track to lower and lower altitude. Of course, seeing and guiding get worse with decreasing altitude but again I would expect to see it get worse in both X and Y. It gets a lot worse in just Y. In any case this is not a critical problem (for me) but seems to be for others on the forum at Software Bisque. What I'm mainly interested in hearing about from others here is if my analysis and expectation of what I should be seeing vs what I am seeing (with the guide camera at PA=45) is reasonable. I might be missing something obvious. Thanks for any feedback!

Peter

[Shiraz (Ray)]

Hi Peter.

The following has the caveat that I don't have anything like your system, so it is all FWIW.

Could it be that your guide software is transforming the guide offset data and reporting in RA/DEC, regardless of the guidescope alignment? - the CCDAP manual suggests that may be the case and FWIW, that is what phd does.

It is probably reasonable that turbulence will have more variability in the vertical direction, since it is due to air circulation. Thus, for the average target that goes over close to the zenith, the vertical turbulence motion may show up predominantly in RA at low altitudes - maybe??

also FWIW, my EQ8 has "crud in grease" spikes in RA and it is slowly getting more noticeable as time goes by. I guess that should not be surprising, since anything bigger than about 1/100 the thickness of a human hair will be noticeable.

[PRejto (Peter)]

Hi Ray,

I really appreciate your reply!

It has been firmly established that in TSX the guide graph purely represents the guide star relative to the chip orientation in X and Y. It does not translate that motion to RA and DEC. Of course it does do so behind the scenes. In any case this is why when using TSX to measure PE it is a requirement to arrange the guide camera so X=RA by putting the camera at PA=0 (or 180).

I'll think ore about the atmospheric turbulence idea. If star jumping around really isn't random then my idea would fail. How would one verify such a thing?

Peter

[iborg (Philip)]

Hi Peter

I have even less knowledge than Ray about these things.

But, I would be curious about the results of two things.

1: Change the camera angle through 90 degrees - does this change the axis with the main spikes?

2: Change the camera angle through another 90 degrees - does that reverse the direction of the spikes? (Can the spikes be reversed in direction?)

Good luck

Philip

[Shiraz (Ray)]

Originally Posted by PRejto

Hi Ray,

I'll think ore about the atmospheric turbulence idea. If star jumping around really isn't random then my idea would fail. How would one verify such a thing?

Peter

Your mount is good enough that tracking a target low down in the west and then comparing with low down in the north should show if there actually is a turbulence pattern difference - I think that vertical turbulent flow should be aligned predominantly RA when west and DEC when north.

I have certainly seen seeing blur that distorts the image more strongly in the turbulent flow direction when doing planetary imaging, but couldn't find any references to suggest that the same applies to DSO imaging - come to think of it though, this might just possibly explain some odd star elongation that I sometimes get - might do some testing .

[gregbradley]

When you callibrate using the Sky X or CCDsoft the guide camera angle is calculated and allowed for when it makes corrections so having the guide camera exactly square to the scope is not necessary. I prefer to have it close but I can't say I have noticed that it matters much so long as you do fresh callibrations if you change the guide cam in any way.

As far as sudden spikes are concerned that may also be an issue with the cam pin or belts shedding material onto the motor pulley. Best to use the latest grey belts if you are seeing that (not sure about PME 2). The black ones can shed at times. I just inspected the black ones on the PMX and they seemed ok but I got some grey ones just in case. I recall some early PMX owners having spike issues that came from the black drive belts shedding material and it stopped when changed over to the grey ones that don't shed. A sudden spike implies crud on the gears or worm or on the motor pulley. Or some unwanted pulse from the guiding. I have not really noticed too much of a balance between X and Y errors. They often are quite different, not massively so but different.

Last Protrack model I did I also noticed it gave worse guiding so I turned it off for guiding. On my PME though it improves the guiding. But that is an earlier build version of Sky X. Perhaps something got changed in later Sky X build versions. It may also explain why the spikes are when the mount is lower in the sky as you would expect more flex there and Protrack may be overcompensating.

Greg.

[PRejto (Peter)]

Originally Posted by iborg

Hi Peter

I have even less knowledge than Ray about these things.

But, I would be curious about the results of two things.

1: Change the camera angle through 90 degrees - does this change the axis with the main spikes?

2: Change the camera angle through another 90 degrees - does that reverse the direction of the spikes? (Can the spikes be reversed in direction?)

Good luck

Philip

Thanks Philip. That sounds like quite a good suggestion. I will try it at first opprtunity!

Originally Posted by Shiraz

Your mount is good enough that tracking a target low down in the west and then comparing with low down in the north should show if there actually is a turbulence pattern difference - I think that vertical turbulent flow should be aligned predominantly RA when west and DEC when north.

:.

Ray, unfortunately I can't get low in the north or south from my location. Otherwise it would be quite a good test. I will try it getting as low as possible towards the south, and then as low as possible to the north west so as to get 90 degrees between the observations. Thanks for the suggestion!

Originally Posted by gregbradley

When you callibrate using the Sky X or CCDsoft the guide camera angle is calculated and allowed for when it makes corrections so having the guide camera exactly square to the scope is not necessary. I prefer to have it close but I can't say I have noticed that it matters much so long as you do fresh callibrations if you change the guide cam in any way.

As far as sudden spikes are concerned that may also be an issue with the cam pin or belts shedding material onto the motor pulley.

Greg.

Greg, I totally agree with what you are saying. My point, however is that with the guide camera pointed at pa=45 any crud or PE, or whatever on either axis would show up on the graph equally distributed between X and Y. It would only show up "more" on one axis than the other as the camera continued to be rotated towards the 4 cardinals (0,90, 180,270). On a cardinal all of the error of a particular axis would show on either X or Y, but not both. Calibration doesn't change at all what is seen on the graph (which is precisely why calibration is NOT required to perform PE measurement in TSX. With this in mind it is tricky to figure out why one should see any major differences between X and Y plots on average. How can one explain a strong spike in only one axis with the camera oriented at PA=45?

I really appreciate any and all input on this as my question at SB is being totally ignored.

Peter

[gregbradley]

Yes I see what you're saying. Well if you can rule out crud on the gear system or belts then that leaves a software spike caused by some not wanted pulse. Does it only happen with Protrack enabled or without?

Greg.

Greg, I totally agree with what you are saying. My point, however is that with the guide camera pointed at pa=45 any crud or PE, or whatever on either axis would show up on the graph equally distributed between X and Y. It would only show up "more" on one axis than the other as the camera continued to be rotated towards the 4 cardinals (0,90, 180,270). On a cardinal all of the error of a particular axis would show on either X or Y, but not both. Calibration doesn't change at all what is seen on the graph (which is precisely why calibration is NOT required to perform PE measurement in TSX. With this in mind it is tricky to figure out why one should see any major differences between X and Y plots on average. How can one explain a strong spike in only one axis with the camera oriented at PA=45?

I really appreciate any and all input on this as my question at SB is being totally ignored.

Peter[/QUOTE]

[PRejto (Peter)]

Of course, as soon as I've started this thread I'm getting perfect guiding tonight. Good focus too. No spikes at all. I guess there is no substitute for good seeing. I still don't understand what I'm seeing on worse nights but would now think it wouldn't be software. How could it be? The software doesn't know what the atmosphere is doing.

Peter

[cfranks (Charles)]

Peter,

Not that I can explain it but I was getting massive , random, spikes on both X & Y axes on my MX. On good days (nights) I guide with usually <0.5 error but the spikes were ~ 26.0 with an odd excursion >45.0. I did notice that the guide star centroid definitely did not move that far and eventually came to the conclusion that the spikes seemed to develop with variations in the brightness of the guide star. Targets where I was lucky to get a bright guide star had few or no spikes. I can't prove this and I haven't had a clear sky in about 3 months so I can't test anymore. I just bought the SBIG Guiding Kit for my ST-I as that looks like it will give a brighter and larger FOV for the guider so I will see what happens when I get a clear sky.

Charles

[gregbradley]

If you are getting crazy large errors that sounds like the software is confusing a hot pixel or a white line edge defect for the guide star.

What guide camera are you using? I used to get that with a Lodestar which had a bad bright white line edge which I couldn't correct with CCDsoft but you can with Maxim (dark subtract).

Greg.

[PRejto (Peter)]

The spikes I see are nowhere near that large...usually less than as pix and sometimes a bit more. But, enough to ruin a good image. Curious to see if you come to a resolution Charles!

Peter

[gregbradley]

Originally Posted by PRejto

The spikes I see are nowhere near that large...usually less than as pix and sometimes a bit more. But, enough to ruin a good image. Curious to see if you come to a resolution Charles!

Peter

Are these spikes occurring with Protrack turned off? I did see worse guiding with Protrack turned on. It stopped when I turned it off.

Protrack seems a bit hit and miss. Martin Pugh said to turn it off as it worsens yet with my CDK and PME it improves things. But on the PMX it seemed to worsen at my home observatory on a portable pier (perhaps not as rigid as my other 2 permanent piers).

Of course wind and seeing do come into it as does guide star selection. But a spike isn't guide star selection. It can happen in cloudy conditions when a cloud interrupts the guider for a few cycles but I assume you realise this. Same with wind causing spikes.

Greg.

[cfranks (Charles)]

Originally Posted by gregbradley

If you are getting crazy large errors that sounds like the software is confusing a hot pixel or a white line edge defect for the guide star.

What guide camera are you using? I used to get that with a Lodestar which had a bad bright white line edge which I couldn't correct with CCDsoft but you can with Maxim (dark subtract).

Greg.

Hot pixels are always possible but I got the spikes with both ST-I and Lodestar cameras. I use TSX and guider full image calibration which, hopefully, reduces the chance of HP. I do have ProTrack on so that might be a factor. When I get a clear night again (Ha!), I'll do some testing.

Charles

[Paul Haese]

Hi Peter,

as you know I have had my fair share of guiding issues with my MX but have perfect guiding on my PME. Some of the things I did to sort my issues were as follows.

1. do PEC via Pempro. Do this a couple of times and make sure you get good results. I think you have done this before but just check. Try guiding without PEC enabled first. Then try it with it enabled. Monitor the effects. Make sure you recalibrate between each guiding run.

2. Do a 300 point model, get your pointing very close to the refracted pole. I have my PA so close I get no reported error. Now that means my Dec rarely moves off the line. It will also mean you can then use Protrack effectively. Protrack does not work well with poor data. Try the guiding with and without again. Make sure to recalibrate the guiding between each guiding run again.

3. Check all your cabling and in particular cables going through the mount. I found putting too many cables in the mount causes issues with guiding. Make sure you cables have some slack between the scope and the saddle plate. Check for balance. Having a cable too long might well catch and cause these sort of issues.

4. Check with different software. It you want to eliminate software as a problem then go and use another guide software. Check the results against each other. Do this on a night when the guiding is acting up on one. It will tell you if the software is the cause of the problem.

5. I doubt this is a mechanical problem with crud. Your mount is a huge mount and would not be affected by tiny amounts of crud. However for the sake of completeness just check your drives for dirt accumulation.

6. Eliminate wind gust and seeing conditions. If you typically have bad seeing this might well be contributing to the problem more than you think. Wind also can be problematic with heavier loads, even small buffeting wind can cause some issues but not likely. A wind shade will prevent this from occurring.

I tend to think this is more of a mechanical issue going on here, but the main thing is to cover all the bases. Do things with and without options enabled. That will get to the root cause of the problem.

[PRejto (Peter)]

Paul,

Many thanks for the long post. All very good advice! In particular I don't think my polar alignment or balance is quite good enough though Patrick Wallace says that Protrack will compensate for poor alignment if there is a good and accurate model. But, for sure with good PA both guiding and Protrack will not work as hard.

For sure in better seeing I'm seeing far fewer spikes. What I found (findQ!) impossible to think through is how I could only see a spike in one axis given the camera at 45 degrees. Next time I see this I intend to guide in Maxim and see if I get anything similar.

Thanks,

Peter

[SpaceNoob (Chris)]

I found the best thing to do with pro-track was to turn it off. I just found it was inconsistent. I also have no intention on maintaining 300 point models anymore as I'd rather extend the life of the mechanical moving parts in my dome, especially when it spends half of the year covered in ice.... With the PME my pointing is already extremely good, any error is quickly resolved via automated platesolve and correction to within a few arc-seconds.