¿ªÔÆÌåÓý

How to train PEC


 

Hi!

I'm trying to get my head around how I can train PEC in a good way. I understand that I can do it manually, that is with an eyepiece and manually issuing guide commands.

But: If I can do that manually, could I not use PHD2 to do the same? Use the RA guide commands to train PEC, instead of me sitting there trying to guide? That is: manually engaging the "Train PEC", then start the gudiing only in RA, and let it run. Take a number of runs, then average and smooth them. Would that work?

Thirdly, Pecprep seems to be a utility that can produce a "PEC file". Can I use that and someway upload to the Gemini 1? I understand Pecprep to be part of th EQMOD project, and that does not include Gemini, but is it doable anyway?

Finally, I can use Pempro. However, somewhat costly.... is that worth the money?

Magnus



 

I think the big question is what is PEC for? It is my view that it is there to get you better unguided exposures.



I believe that if you taking guided exposures, then you don¡¯t need PEC (at least not traditional PEC) but just rely on PHD2 (with possible use of PPEC). I¡¯ve never heard of anyone actually doing a proper double blind test to see which is the better approach.



And in answer to your question about using guiding s/w to train PEC, yes that¡¯s what I do: Set up PHD2 with the guide camera in the imaging scope, calibrate it and start guiding. The start the PEC training.



Dave

From: Losmandy_users@... [mailto:Losmandy_users@...]
Sent: 26 November 2017 14:09
To: Losmandy_users@...
Subject: [Losmandy_users] How to train PEC





Hi!

I'm trying to get my head around how I can train PEC in a good way. I understand that I can do it manually, that is with an eyepiece and manually issuing guide commands.

But: If I can do that manually, could I not use PHD2 to do the same? Use the RA guide commands to train PEC, instead of me sitting there trying to guide? That is: manually engaging the "Train PEC", then start the gudiing only in RA, and let it run. Take a number of runs, then average and smooth them. Would that work?

Thirdly, Pecprep seems to be a utility that can produce a "PEC file". Can I use that and someway upload to the Gemini 1? I understand Pecprep to be part of th EQMOD project, and that does not include Gemini, but is it doable anyway?

Finally, I can use Pempro. However, somewhat costly.... is that worth the money?

Magnus


 

Magnus,

You are correct : you can use your autoguider with the Gemini's built in PEC recorder /corrector.? You skip using the old hand controller method.? The correction signals from the autoguider work just fine.

That is not clearly explained in the Gemini 1 manual.??

You want to start by a precise polar drift align,? using a star near meridian and the celestial equator.

That is important because the PEC you want to correct is the RA's error, and even though the software says it can subtract out DEC drift, you want that as near zero as possible.

After drift alignment is clean and DEC drift is near zero over 8+ minutes, test that your autoguider is working correctly.? Remember to put the Gemini.net into G for Guide mode, else the Gemini will not accept any autoguide signals. Take a 10 minute? autoguide image.? ?Are the stars correctly round...no glitches?? If the stars are proper and no glitches, you can start a PEC recording.

Again select a new star near meridian and equator, perhaps point the scope to the west, then slew to a star just east of the meridian that will pass to through meridian over time.? You do that because you want to run about 3 PEC recordings and "add" and "average" them in the Gemini system.

The first PEC recoding will last exactly 8 minutes. Before that begins, the Gemini will transfer it's existing PEC data to its SRAM backup register, but that is all 0s.

? ?The system knows from its optical encoder where the starting position is (relatively, not exactly).? When the first run ends, the Gemini puts that data into its primary PEC register (in SRAM).??

When you start the next run, the Gemini will first transfer it's PEC data to the SRAM, then wait for the worm to get back to its same starting rotation.? That could take up to 7.99 more minutes.? Then the run itself will take 8 minutes.? At the end of that ND run, the 2nd PEC data is in the "top" PEC register.? You can tell the Gemini to "average" the 2nd run with the backup data.? That averaged data is now in the topmost PEC register in SRAM.??

You should use Gemini.net to save that data into your PC in a file.??

You can redo the steps of the 2nd run.? It is not clear to me how averaging the 3rd run works.? It might actually have (in another register) a counter so it knows this is the 3rd run, so all 3 runs are equally weighted.? ?Or it may be just takes 1/2 of run 3 and adds it to 1/2 of (the PEC backup data).? Only the person with the source code can say for sure.??

There are commands to smooth the PEC data, and other things.? I used those commands also.??

Save the resulting PEC to your PC in a new filename.? When you are done, test the PEC by taking a 10 minute image.? If you had a glitch you PEC data will have recorded the glitch and your PEC image could be worse than one plain autoguided.? If the PEC has a glitch, you can upload one of the earlier PEC files to the Gemini and see if it is correct.??

And:

You must park your mount before shutting off power.? That puts the worm in a known (in SRAM) rotation.? (It is said that even if the power goes out, the Gemini knows the last worm position, but on other mounts like the Atlas, you must always park.)

On power up, the Gemini always defaults to "PEC OFF".? Be sure to turn that on in the hand controller.? There is a checkbox in Gemini.net to automatically turn on PEC when it connects.? You could ensure your PEC is on using that.??

Epilogue:

PEC did work well for me, but I no longer use the PEC...I was lucky to get my PE without PEC down to ~1arcsec, so regular autoguiding gives me round stars.? Thats,all written up in the PDF "Improving the PE of the G11" in our group files section.??

(I will send you some Belleville worm washers,? Magnus.)

All the best, and hope these comments help you.

Michael?



On Nov 26, 2017 6:08 AM, "jfev5mnsvqiyzpq756wvokt55jve7sbg2esr5mip@... [Losmandy_users]" <Losmandy_users@...> wrote:

?

Hi!

I'm trying to get my head around how I can train PEC in a good way. I understand that I can do it manually, that is with an eyepiece and manually issuing guide commands.

But: If I can do that manually, could I not use PHD2 to do the same? Use the RA guide commands to train PEC, instead of me sitting there trying to guide? That is: manually engaging the "Train PEC", then start the gudiing only in RA, and let it run. Take a number of runs, then average and smooth them. Would that work?

Thirdly, Pecprep seems to be a utility that can produce a "PEC file". Can I use that and someway upload to the Gemini 1? I understand Pecprep to be part of th EQMOD project, and that does not include Gemini, but is it doable anyway?

Finally, I can use Pempro. However, somewhat costly.... is that worth the money?

Magnus




 

Do you have to use an Ethernet connection to do all the pec runs, or can you use just USB connection?? ?I only use USB and can't get Gemini.net, right.? I have a new G811G with gemini-2.? No pempro yet.


 

PE correction helps for both guided and unguided imaging. Some errors cannot be easily guided out, and PEC helps smooth things out so the guiding is better

My G11GT has a nasty 42 second recurring error that i see on my PHD guiding, which none of the algorithms including PPEC can effectively address.?

I'm waiting on pempro to have the two worm cycle error correction to address this


B


 

>?I think the big question is what is PEC for? It is my view that it is there to get you better unguided exposures.

I know this is a frequent subject for debates, but proper PEC programming does help prevent tracking errors before they happen. This results in better guided or unguided performance. An autoguider can only correct errors after they have occurred, PEC can correct them so they never happen.

If you want to compare PEC to PPEC, there may still be benefits, although these may be fewer
  • For one thing, PEC operates internally to Gemini. The corrections are not only precise, they are instantaneous, not subject to communications delays and PC timing like the corrections from PPEC

  • Gemini keeps track of the worm index position through encoders, allowing existing PEC to be used immediately after slews and goto's -- PPEC has to be relearned after a slew

  • PEC software allows many cycles to be used and programmed, filtering out a lot of the random errors that PPEC might not. Since PEC doesn't have to be relearned after each slew, dedicating an hour or two to training PEC can pay off big dividends over many months of imaging

  • Some PEC software (f.i., PemPro) has built-in knowledge about various gears and error cycles particular to each mount type. This allows PEC curve to be fine-tuned very easily to fix issues specific to a mount. PPEC doesn't do that. I believe an upcoming version of PemPro will even correct errors that occur over multiple worm cycles.?
PPEC is a great tool, but generally it's not better than PEC for both, guided or unguided exposures. In fact, PHD2 team recommends to train PEC before using PPEC if your mount allows it. Perhaps that's the best of both worlds: PEC correcting most of the errors, and PPEC picking up any small uncorrected errors that remain.

Regards,

? ? -Paul


 

Comments...

1. You don't need any PC connection, per se, to run the Gemini PEC recorder.? You could use an autoguider camera and the ST4 port to do it I suppose.

2. If you are looking at PHD2 on your PC,? and you can use its Manual guide to move your mount (remember Gemini must be set to Guide mode or easiest to select "G" on the Gemini.net applet) then your mount is getting pulse commands from the Ascom system. (You had to install ASCOM first and the Gemini.net driver).??

In that Gemini.net driver, I think is how you either select the Serial port, or ethernet connection method.

Anyway... play around and you will find a way. ...!

Have fun,
Michael?





On Nov 26, 2017 8:26 AM, "yh@... [Losmandy_users]" <Losmandy_users@...> wrote:
?

>?I think the big question is what is PEC for? It is my view that it is there to get you better unguided exposures.


I know this is a frequent subject for debates, but proper PEC programming does help prevent tracking errors before they happen. This results in better guided or unguided performance. An autoguider can only correct errors after they have occurred, PEC can correct them so they never happen.

If you want to compare PEC to PPEC, there may still be benefits, although these may be fewer
  • For one thing, PEC operates internally to Gemini. The corrections are not only precise, they are instantaneous, not subject to communications delays and PC timing like the corrections from PPEC

  • Gemini keeps track of the worm index position through encoders, allowing existing PEC to be used immediately after slews and goto's -- PPEC has to be relearned after a slew

  • PEC software allows many cycles to be used and programmed, filtering out a lot of the random errors that PPEC might not. Since PEC doesn't have to be relearned after each slew, dedicating an hour or two to training PEC can pay off big dividends over many months of imaging

  • Some PEC software (f.i., PemPro) has built-in knowledge about various gears and error cycles particular to each mount type. This allows PEC curve to be fine-tuned very easily to fix issues specific to a mount. PPEC doesn't do that. I believe an upcoming version of PemPro will even correct errors that occur over multiple worm cycles.?
PPEC is a great tool, but generally it's not better than PEC for both, guided or unguided exposures. In fact, PHD2 team recommends to train PEC before using PPEC if your mount allows it. Perhaps that's the best of both worlds: PEC correcting most of the errors, and PPEC picking up any small uncorrected errors that remain.

Regards,

? ? -Paul


 

Hi!

Great, thanks!! Will play with this next clear night.... next year or so, I guess... :)

Best,

Magnus


 

Brian,
? ? ?I noticed your comment about having a 42 second recurring error on your G11GT. ??
A couple of weeks ago I bought the Losmandy RA-AXIS Titan module as an update to my G11. ? I have been having problems being able to record a PEC file using the Gemini2 built in PEC training. ?Today I looked at a PHD2 guide log and it appears from that one guiding session my mount has a Spike on the 28% ?point of my chart (using?PEM Pro viewer). ? It doesn't match up with any of the harmonics of the single worm revolution.

? ? Tonight I ran a 3 cycle run and a 5 cycle run which I will be able to confirm if that is the case.

? ?The spike I saw on that guide log today was by far the highest peak of the log. ?Since it does not coincide with the 319 second worm period or the harmonics to the 5th or 6th order. ?I am wondering were to look for the problem. ?

? ?If you would like to have more of a discussion of this issue and how to resolve it, we could set up a separate thread on this or the Titan Mount group. ? Since all the new Titans have the tucked motors they may all have the same issue.

? ?aruckle



 

Hi aruckle,



There is some discussion over at PemPro forums on this.



My G11GT PE looks like this







Thanks



Brian





Brian Valente

Brianvalentephotography.com



From: Losmandy_users@... [mailto:Losmandy_users@...]
Sent: Tuesday, November 28, 2017 12:19 AM
To: Losmandy_users@...
Subject: [Losmandy_users] Re: How to train PEC





Brian,

I noticed your comment about having a 42 second recurring error on your G11GT.

A couple of weeks ago I bought the Losmandy RA-AXIS Titan module as an update to my G11. I have been having problems being able to record a PEC file using the Gemini2 built in PEC training. Today I looked at a PHD2 guide log and it appears from that one guiding session my mount has a Spike on the 28% point of my chart (using PEM Pro viewer). It doesn't match up with any of the harmonics of the single worm revolution.





Tonight I ran a 3 cycle run and a 5 cycle run which I will be able to confirm if that is the case.



The spike I saw on that guide log today was by far the highest peak of the log. Since it does not coincide with the 319 second worm period or the harmonics to the 5th or 6th order. I am wondering were to look for the problem.



If you would like to have more of a discussion of this issue and how to resolve it, we could set up a separate thread on this or the Titan Mount group. Since all the new Titans have the tucked motors they may all have the same issue.



aruckle









[Non-text portions of this message have been removed]


 

brian,
? ? Did you attach a photo that did not come through?

aruckle


 

Apparently so!

here's the pic again