I ran another test last night and it looks as though you were right it was just luck that the phase was correct!
So the problem definitely seems to be that the TCS software is not picking up the index correctly.
Just to confirm this can you open the text files of the two long PE runs you made and read what is the index for the maximum positive error. Is it the same for both runs, does it match either of your "Periodic error curves for Bisque TCS"?
One thing for sure the two curves you made are incredibly similar in shape so if applied correctly the residual PE should be amazingly small!
I'll second the recomendation for PemPro. With it, I set my MX system up in 1 night including a more accurate polar alignment than with the SB/TSX procedure. I haven't run an accurate test but have since managed unguided images of more than 5 min exposure with a WO FLT132.
Charles
I greatly appreciate your further investigation. It really confirms that something is happening that ougtht not to be happening.
I'm not able to tell from the guiding logs where the high or low points are (at least not the log generated by TSX. The CCDSoft log is easier to figure out) but the raw data certainly shows the phase difference between the two curves (which I have posted below).
I'm thinking about an experiment where I have the TCS window open and watch the index number. When it gets to the very end of the cycle I would start guiding, and then stop guiding on a specific number. An examination of the guiding log would show if the same numbers were being logged as those shown in the TCS. I'm pretty sure they would be.
That leads to the question of whether the numbers seen in the TCS really corespond to the worm position? Could a long cable run lead to some sort of latency where the worm is actually more advanced than what is logged?
Charles, I am ready to try Pempro and I'm glad to hear of your positive experience. Would you share what your uncorrected/corrected numbers are? Unfortunately my trial version has expired. I never got to test it months ago due to a camera incompatibility. It would be an expensive experiment particularly as I've had good luck with T-Point and most likely wouldn't use the alignment method either. Still, if it works I could move on from this endless frustration.
Clearly the PEC is bugged and is not working. I would try Pempro. I know Chris Venter did this when he hit a wall with SB PEC and it worked fine.
PEC did not work on my PEC on the first few goes. I updated with a later build and it worked. I did a 20 minute PEC run done in good seeing at my dark site.
Sometimes later builds cause things that were fixed to unfix. Not sure if this is one of those cases.
The PMX is a great mount but it really is a bit of pain in the a...
When it works its great, but its fussy, a little temperamental and a little fragile. I lightly bumped into the counterweight shaft the trip before last. It had been guiding beautifully on the 2 nights preceding. I started getting backlash and oscillating corrections (large positive to negative swings). I knew from having this happen before the tensioning screw must have been loose.
I found the cam pin had popped up and was loose again. Once reset it all worked again (3rd or 4th time I have had to adjust it). So PMX is weakly engineered in some regards - lovely when it works but too fragile. Should have gotten an AP1200.
I put your two PE curve quickly into Excel and both curves suggest a maximum positive error around a value of 300 for the RA index. So the data you collected is definitely consistent. So this is not an issue of acquisition just the interpretation of the data by Bisque TCS.
On the mount this max error should correspond to the maximum negative correction at an index of 300.
I loaded both your curves to my mount and as you can see they are completely out of phase and either side of the target of 300.
I think this proves beyond doubt that the problem is the index is not being picked up correctly by the Bisque PE calculation, probably when the "FIT" is made.
I put this on the SB forum so lets see what they come back with.
On the positive side the uncorrected PE is incredible for both our mounts and the consistency of shape between your two PEC curves suggests if the phase was correct the correction would be very good indeed.
Many thanks for your analysis, and for posting at SB. I hope this gets some attention!
Would you please explain how you are able to "know" when loading these curves into Excel where the worm is in the cycle? If the data of the worm position is incorrect in the guiding log how would you know what the correct index is?
Clearly the PEC is bugged and is not working. I would try Pempro. I know Chris Venter did this when he hit a wall with SB PEC and it worked fine.
PEC did not work on my PEC on the first few goes. I updated with a later build and it worked. I did a 20 minute PEC run done in good seeing at my dark site.
Sometimes later builds cause things that were fixed to unfix. Not sure if this is one of those cases.
The PMX is a great mount but it really is a bit of pain in the a...
When it works its great, but its fussy, a little temperamental and a little fragile. I lightly bumped into the counterweight shaft the trip before last. It had been guiding beautifully on the 2 nights preceding. I started getting backlash and oscillating corrections (large positive to negative swings). I knew from having this happen before the tensioning screw must have been loose.
I found the cam pin had popped up and was loose again. Once reset it all worked again (3rd or 4th time I have had to adjust it). So PMX is weakly engineered in some regards - lovely when it works but too fragile. Should have gotten an AP1200.
Greg.
Hi Greg,
Thanks for your comments. I certainly will go to Pempro if SB doesn't offer up a solution soon!
My experience with the MX has been difficult but I put that mostly down to my beginner status with an overwhelming amount to learn from the start. I suppose the journey might have been easier with brand AP, but I'm not yet at the point of wishing I'd made a different choice. Hopefully the bugs will be rectified. Hope springs eternal!
I am not sure I understand exactly your question but this may cover it.
The recorded text file includes the RA index for each point. So if you plot the RA index against the PE error you can see at which index values the error is a maximum.
I previously used the maximum positive error but the maximum negative error might be clearer to see on the excel sheet.
In both cases the maximum negative error is very close to 600 but when loaded to the mount the maximum positive correction is applied at about 800 for A and about 100 for B, not at 600. (The horizontal scale on the mount plot being 0 to 1000.)
Based on this my assumption is the error in the phase is made in the making of the correction curve not during the acquisition.
I am not sure I understand exactly your question but this may cover it.
The recorded text file includes the RA index for each point. So if you plot the RA index against the PE error you can see at which index values the error is a maximum.
I previously used the maximum positive error but the maximum negative error might be clearer to see on the excel sheet.
In both cases the maximum negative error is very close to 600 but when loaded to the mount the maximum positive correction is applied at about 800 for A and about 100 for B, not at 600. (The horizontal scale on the mount plot being 0 to 1000.)
Based on this my assumption is the error in the phase is made in the making of the correction curve not during the acquisition.
Regards
Kim
Thanks Kim, I get it now. I misread your earlier post and made up something you actually didn't say!
Nothing much from SB yet so I tried Pempro last night during a lucky break in the clouds. Remarkably I got my first positive result. My measured PE in Pempro (same camera, resolution, same everthing except using Pempro instead of Bisque TCS) measure my PE as 2.6 arcsec peak to peak (vs 2.2 in TSX) after 9 worm cycles in not great seeing. After applying the correction it is now .7 arcsec peak to peak.
Reasoning that there can only be one correction curve that "works" it seemed possibly interesting to compare results from TSX to Pempro and see what the differences might be. I used data collected a few nights ago, a run of 44 min and compared it to the Pempro results from last night. The 44 min data was collected with a position angle of close to zero with the telescope pointing west, as it was for the data collection. Therefore the "west" selection would normally not be necessary to check (which flips the data vertically).
#1 just places the two curves on top of each other
#2 Better fit with large phase shift.
#3 Bisque curve flipped vertically. Shows phase error, possible match but there are obvious differences in curves.
If this analysis is correct there are two possible problems with the TSX PE calculation. Something is causing the data not to align properly with the worm index, and there might be something not correct about the use of the "west" selection. (though I think the fit is better with the curve not flipped vertically which would mean the major problem is that the curve is just out of phase with the true worm position).
I will be adding this to the discussion at SB.
Thoughts?
Good to see you have something positive for a change.
I would post that data on the SB forum, it is more proof that there is something wrong the the way SkyX calculates and imputes the PEC table.
I was up till two last night and played around with the North and Southern config. But I still have not found any way to get Sky X to correct the PE.
I posted this info in your SB thread.
I did a test with ProTrack on and no PEC, this gave me a better result than when the PEC is activated. (image below)
It must be a Southern hemisphere thing, I don't see any Northerners posting anything about PEC problems.
Charles, I am ready to try Pempro and I'm glad to hear of your positive experience. Would you share what your uncorrected/corrected numbers are? Unfortunately my trial version has expired. I never got to test it months ago due to a camera incompatibility. It would be an expensive experiment particularly as I've had good luck with T-Point and most likely wouldn't use the alignment method either. Still, if it works I could move on from this endless frustration.
Thanks,
peter
Hi Peter,
I'll have to set up my system again to get those figures. Hopefully, the weather will clear soon. I do remember that the RMS error after a TPoint run got to less than 4.
Charles
I can't get my head around why North or South hemisphere should make a difference but it looks that way.
I can't get to my mount for quite a while but what happens if you fit and produce the Bisque TCS PEC curve from one of your uncorrected PE runs then clear it change to Northern hemisphere then reload the same file not changing anything else?
How do the PEC correction curves compare?
I may be wrong but should they not be identical? After all the PEC is just moving the RA faster or slower at a given index, as RA moves clockwise for us and for the North ant-clockwise, but we still want the same correction (with or against RA) at the same index.
Peter,
Did you load the PEC table from PEMPRO into Sky X ?
If so, did you do a tracking log with the PEMPRO's PEC table activated so you could compare the PEMPRO graph to the Bisque TCS graph.?
This would show the difference using SKy X, SB tends to respond better when the evidence is shown in their own software.
Cheers
Phil,
Unfortunately even though log guiding was selected in CCDSoft I think Pempro turns that function off when using CCDSoft's camera to gather data. I will just remeasure the corrected PE using TSX or CCDSoft when the weather improves.
Peter
Hi Peter,
I'll have to set up my system again to get those figures. Hopefully, the weather will clear soon. I do remember that the RMS error after a TPoint run got to less than 4.
Charles
No worries Charles. I've got a fairly good result now using Pempro so knowing your numbers isn't too important.
