Return to the Cloudy Nights Telescope Reviews home pageAstronomics discounts for Cloudy Nights members
· Get a Cloudy Nights T-Shirt · Submit a Review / Article

Click here if you are having trouble logging into the forums

Privacy Policy | Please read our Terms of Service | Signup and Troubleshooting FAQ | Problems? PM a Red or a Green Gu… uh, User

Equipment Discussions >> Mounts

Pages: 1 | 2 | 3 | 4 | (show all)
orlyandico
Postmaster
*****

Reged: 08/10/09

Loc: Singapore
Re: PEPrep help new [Re: korborh]
      #5598950 - 01/01/13 12:17 AM

hi Chris,

Frank (freestar8n) claims "sub 2" FWHM" achievable with Metaguide and an OAG, on a mid-range mount (CGE, CGE Pro).

You might want to try using MG for guiding instead of PHD and see where that gets you... MG "supposedly" can track fast errors better as it uses video. Frank also mentions that it can pre-emptively correct periodic errors if you input the expected cycle of the PE (e.g. the 8/3).

I hope Frank can weigh in here with his input as well..


Post Extras: Print Post   Remind Me!   Notify Moderator  
Ray Gralak
Vendor (PEMPro)


Reged: 04/19/08

Re: PEPrep help new [Re: korborh]
      #5599300 - 01/01/13 10:09 AM

Quote:

Quote:

Not necessarily true because guiding performance cannot be separated from scintillation.




Scintillation would be random while guiding errors due to gears/worm would me mostly periodic. So it seems one should be able to separate them, no?





No you really can't. My example of the perfect tracking mount with different RMS errors under different seeing conditions I think supports this point.

Another example... in lower end mounts there can be a lot of "gear noise" that Orly mentioned earlier. This noise is random so it cannot always be distinguished from seeing anomalies. Some mounts, the the Atlas employ stepper motors which produce measurable vibrations visible in an FFT spectrum if you use a webcam so that you can sample the star fast enough. The stepper frequency is similar to the 1 second autoguider exposures commonly used so (because of the Nyquist sampling theorem) the vibrations cannot be isolated with typical 1-second exposures.

Quote:

Quote:


You cannot judge guiding entirely from the RMS number.

-Ray




I agree with this (for my mount) based on my own observations of the tracking graph vs. FWHM in resulting images.




I believe you have to do more than that even. FWHM measurements can be affected by many things including seeing conditions, telescope aperture and quality, focus, brightness of the star, which camera you use, which software you measure with, duration of the exposure, vibrations, etc. I think to measure guiding quality you need to take several very short exposures (or measure many stars in a single short exposure) to get an average "short-exposure FWHM". This short-exposure FWHM is the baseline against which you can measure autoguider performance by comparing it to the average FWHM you get in a longer duration autoguided exposure. How little the FWHM increases I think would be the most accurate determination of autoguiding quality. But even in this case seeing can be an unknown factor so it's important to average many stars.

-Ray Gralak


Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Postmaster
*****

Reged: 08/10/09

Loc: Singapore
Re: PEPrep help new [Re: Ray Gralak]
      #5599307 - 01/01/13 10:16 AM

i think Maxim can guide from multiple stars?

or is it some other software?

p.s. Ray i finally paid for PEMPro (based on the assumption that I now have a mount that deserves it...) to my dismay the uncorrected PE according to PEMPro is 2.33" peak-to-peak (across 3 cycles). I'm not so sure if that can be improved... uploading the curve takes a LONG time..


Post Extras: Print Post   Remind Me!   Notify Moderator  
Ray Gralak
Vendor (PEMPro)


Reged: 04/19/08

Re: PEPrep help new [Re: orlyandico]
      #5599327 - 01/01/13 10:32 AM

Quote:

hi Ray,

you are absolutely correct about the phase shift introduced by an AG.

however we're talking about a CGEM here.. with its 88 PEC cells. really, one could train the PEC with the AG, download the curve from the mount, smooth and otherwise massage it, even edit it with MS Excel (it can be saved as a CSV file) and then re-upload it.

I am not dissing PEMPro - far from it, I recognize its the best out there - but for a CGEM with its 88 PEC cells and large 8/3, I think PEMPro is overkill. Things like the centroid calculation shifting etc. and fast gear noises can't be corrected by the CGEM PEC anyway - 88 cells over 479 seconds means a guide correction is only being applied once every 5.44 seconds.

so the cheapest and I suspect fairly effective way to train a CGEM's PEC is to use a long slow scope, and use a 5-second guider exposure.

that said... how many PEC cells does the GTO CP3 have? I have not had success with PEMPro in the past.. the CGEM was.. well, the 8/3 on mine is huge; my AP600 both in its original QMD guise and after I replaced the controller with a Littlefoot, can only be programmed via the ST-4 interface, and I could never get a good PEC training. I finally am at the point where I have a mount that can really benefit from PEMPro... something I'll have to look into because I ruined the built-in PEM training.



I wasn't trying to bring PEMPro into the discussion. My comments were from research and actual measurements from many mounts over the years.

That said, I think there are a couple problems that I can see with your arguments. First, an 88-cell PEC table can be affected just as much as a 970-cell table (most AP Mounts) by phase shifting. You still want to make sure the phase is centered correctly. In fact it's probably more important in an 88-cell PEC table because of the potential of a longer "wrong move" that a cell might cause.

But here's the problem with your argument of training the CGEM with an autoguider. As you know the problem with the 8/3 fundamental is that in each cycle it does not repeat in phase. So, just recording PEC to the mount with an AG does not work (well, the PEC curve works in one out of every three worm cycles). Simple averaging of 3 cycles does not work either. You cannot remove the 8/3 fundamental with averaging. If you are doing that it's no wonder PEC isn't working for your CGEM.

But you can use PEMPro to create a PEC curve that excludes only the problematic 8/3 frequency and upload it to the CGEM. Of course the uncorrectable 8/3 frequency will still exist but any correctable periodic error should be removed.

-Ray Gralak


Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Postmaster
*****

Reged: 08/10/09

Loc: Singapore
Re: PEPrep help new [Re: Ray Gralak]
      #5599338 - 01/01/13 10:38 AM

hi Ray,
that's right, but the thing is PECTool allows to export the curve as CSV. So you can massage it, do curve-fitting, etc. in Excel. Strictly speaking this is outside the scope of PECPrep of course... but most of what you can do in PEMPro can also be done with MS Excel and PECTool...

That said, both PEMPro (trial) and the AG with PECTool method worked. As I mentioned earlier I got my 40" PE down to 25". PEMPro and PECTool both performed at about the same level. Since my CGEM's 8/3 is about 20" to 22" its no wonder that the PE couldn't go lower than 25".

The OP seems to have a good CGEM though. PEC training would work very well for him... I still think PECTool is "sufficient" for his needs. Of course PEMPro would be better.


Post Extras: Print Post   Remind Me!   Notify Moderator  
Ray Gralak
Vendor (PEMPro)


Reged: 04/19/08

Re: PEPrep help new [Re: orlyandico]
      #5599363 - 01/01/13 10:54 AM

Quote:

hi Ray,
that's right, but the thing is PECTool allows to export the curve as CSV. So you can massage it, do curve-fitting, etc. in Excel. Strictly speaking this is outside the scope of PECPrep of course... but most of what you can do in PEMPro can also be done with MS Excel and PECTool...

That said, both PEMPro (trial) and the AG with PECTool method worked. As I mentioned earlier I got my 40" PE down to 25". PEMPro and PECTool both performed at about the same level. Since my CGEM's 8/3 is about 20" to 22" its no wonder that the PE couldn't go lower than 25".

The OP seems to have a good CGEM though. PEC training would work very well for him... I still think PECTool is "sufficient" for his needs. Of course PEMPro would be better.



I think many CGEM owners would be interested in hearing what steps you took in Excel to eliminate the 8/3 fundamental from the data. I'm sure it's possible but I would like to hear the steps you took.

Thanks,

-Ray


Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Postmaster
*****

Reged: 08/10/09

Loc: Singapore
Re: PEPrep help new [Re: Ray Gralak]
      #5599381 - 01/01/13 11:08 AM

i just did a curve-fit with a quadratic function

Excel has some built-in curve fitting routines and there are also some shareware curve-fitting libraries (which do up to 5th order.. just like PEMPro)

that's based on the assumption that the fundamental is just an ordinary sine wave... the whole "cost vs benefit" is based on the assumption that you have Excel (or maybe the free OpenOffice).

if you have to PAY for Excel then PEMPro is suddenly a bargain...


Post Extras: Print Post   Remind Me!   Notify Moderator  
Ray Gralak
Vendor (PEMPro)


Reged: 04/19/08

Re: PEPrep help new [Re: orlyandico]
      #5599389 - 01/01/13 11:15 AM

Quote:

i think Maxim can guide from multiple stars?

or is it some other software?

p.s. Ray i finally paid for PEMPro (based on the assumption that I now have a mount that deserves it...) to my dismay the uncorrected PE according to PEMPro is 2.33" peak-to-peak (across 3 cycles). I'm not so sure if that can be improved... uploading the curve takes a LONG time..




Orly, I assume this is your Mach 1? If so, you should be able to get it under 1 arc-second with no problem. Yes, the upload to 970 cells can take a while. I'll have to see if there is a way to improve the performance in the AP V2 driver.

The good thing is that you'll get a free upgrade to PEMPro V3, which will have lots of improvements, including the ability to select a particular star if you want to do that (implied from the question you asked on the PEMPro forum).

-Ray


Post Extras: Print Post   Remind Me!   Notify Moderator  
Charlie B
scholastic sledgehammer


Reged: 03/22/08

Loc: Sterling, Virginia
Re: PEPrep help new [Re: Ray Gralak]
      #5599401 - 01/01/13 11:24 AM

Quote:

RMS is not necessarily a good value to use to judge autoguiding performance.




Concur Ray! I take an image and measure the PSF of single stars. The FWHMx and FWHMy should be nearly the same to produce round stars. I consider an r-value (the ratio of the two values) of 0.95 or better as really good for my equipment. Also, the guiding performance will change as a function of the portion of sky being observed. With my G11, I had good guiding until just past the meridian looking straight up. I then fell into a DEC null zone, where the guiding oscillated back and forth independent of any internal periods, but more a function of my PHD parameters and independent of RA guiding.

I've still not worked out the best guiding parameters for this condition.

I thought metaguide was interesting, but, since it only works on video cameras and could not be used by my guide cameras, I stuck with the more universal PHD guiding.

Before I bought Pempro, I used the PECTool and found it inadequate. Yes I could average and massage it in excel, but, for some reason, could not get the same good results that I got with Pempro. I probably don't know good massage techniques.

Regards,

Charlie B


Post Extras: Print Post   Remind Me!   Notify Moderator  
korborh
scholastic sledgehammer
*****

Reged: 01/29/11

Loc: Arizona
Re: PEPrep help new [Re: Ray Gralak]
      #5599425 - 01/01/13 11:42 AM

Quote:

I think to measure guiding quality you need to take several very short exposures (or measure many stars in a single short exposure) to get an average "short-exposure FWHM". This short-exposure FWHM is the baseline against which you can measure autoguider performance by comparing it to the average FWHM you get in a longer duration autoguided exposure. How little the FWHM increases I think would be the most accurate determination of autoguiding quality. But even in this case seeing can be an unknown factor so it's important to average many stars.

-Ray Gralak




The complication is that in short (focus) exposures, the FWHM estimates can be inaccurate as the stars can be quite non-Gaussian. Most astro-software rely on Gaussian fit to get FWHM and centroid. Also during a long-exposure, the local seeing could actually improve compared to the (earlier) short exposure. So I am not so sure if the short-exposure is a good baseline to quantify guiding quality.


Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Postmaster
*****

Reged: 08/10/09

Loc: Singapore
Re: PEPrep help new [Re: Charlie B]
      #5599426 - 01/01/13 11:42 AM

Charlie,

PEMPro almost universally gives better PE values (both pre- and post-PEC/PEM) than say PECPrep. This is why alph always wants to see raw logs...

This could account for the difference in values. Normally a person using PECTool for training wouldn't use PEMPro for measuring the PE... they'd use PECPrep.

I am (right now) capturing data with PHD so that I can verify - with PECPrep - if the reduction from 2.47" to 0.42" (reported by PEMPro) is believable.

In my experience though PECPrep and PEMPro values are largely in agreement.. (although PECPrep still gives higher values)


Post Extras: Print Post   Remind Me!   Notify Moderator  
cn register 5
scholastic sledgehammer


Reged: 12/26/12

Re: PEPrep help new [Re: Ray Gralak]
      #5599428 - 01/01/13 11:44 AM

Is anything more complex than averaging the PEC data in multiples of three consecutive cycles required?

It won't eliminate the 8/3 fundamental but it will prevent it affecting the averaged PEC data.

Chris


Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Postmaster
*****

Reged: 08/10/09

Loc: Singapore
Re: PEPrep help new [Re: cn register 5]
      #5599445 - 01/01/13 11:52 AM

the 8/3 will cause lumps in the data. so you'd need a lot of cycles (maybe 10+) to smooth out the 8/3 lumps.

again this depends on how large your 8/3 is. For the OP, the 8/3 is about 6" so a good PEC training would benefit him well.

what I did was do one round of PEC training with the AG, export the curve as CSV, then fit it to a sine wave in Excel. Based on the assumption that the worm PE is a pure sine wave. That way you don't need to waste so much time capturing.

PECTool can also do smoothing on its own, which in my experience is "good enough" - just press the "Smooth" button enough times until the lumpiness is gone and only the fundamental is there.........


Post Extras: Print Post   Remind Me!   Notify Moderator  
Ray Gralak
Vendor (PEMPro)


Reged: 04/19/08

Re: PEPrep help new [Re: orlyandico]
      #5599447 - 01/01/13 11:56 AM

Quote:

i just did a curve-fit with a quadratic function

Excel has some built-in curve fitting routines and there are also some shareware curve-fitting libraries (which do up to 5th order.. just like PEMPro)

that's based on the assumption that the fundamental is just an ordinary sine wave... the whole "cost vs benefit" is based on the assumption that you have Excel (or maybe the free OpenOffice).

if you have to PAY for Excel then PEMPro is suddenly a bargain...




But there are multiple sine waves (at least the 1x and 8/3x fundamentals) so the trick is that you have to isolate and subtract the 8/3x fundamental from the excel-fitted waveform. So, while curve fitting may get you a curve you still need to subtract the 8/3x fundamental to get the curve you *really* need. In fact this is probably not something you can do in Excel very easily.

-Ray


Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Postmaster
*****

Reged: 08/10/09

Loc: Singapore
Re: PEPrep help new [Re: Ray Gralak]
      #5599457 - 01/01/13 12:03 PM

hi Ray, the simple curve-fitting in Excel is extremely primitive, and it won't try to fit the higher-order e.g. the 8/3. Really, it will just fit the fundamental.

what i do is plot the raw data and the curve fit. the curve fit is always a sine wave.. (with no 8/3)


Post Extras: Print Post   Remind Me!   Notify Moderator  
Ray Gralak
Vendor (PEMPro)


Reged: 04/19/08

Re: PEPrep help new [Re: cn register 5]
      #5599462 - 01/01/13 12:08 PM

Quote:

Is anything more complex than averaging the PEC data in multiples of three consecutive cycles required?

It won't eliminate the 8/3 fundamental but it will prevent it affecting the averaged PEC data.

Chris



Averaging the results of will NOT give you the best solution. To get the best results the minimum you need to do is to remove the 8/3x frequency from the data. The simplest (although not the best) is to break apart the data using FFTs (Fast Fourier Transforms), remove the 8/3x from the FFT data, then do an inverse transform to get back a curve.

-Ray


Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Postmaster
*****

Reged: 08/10/09

Loc: Singapore
Re: PEPrep help new [Re: orlyandico]
      #5599466 - 01/01/13 12:10 PM

ok.. so PECPrep said my raw PE was about 5" - and PEMPro says 2.5"

after the PEM training.. PEMPro says 0.42" and PECPrep says 2"

of course 2" is much less impressive than 0.42" - which again is probably why alph is always asking for PHD log files (and correlates with my experience that PEMPro invariably gives lower PE values).

from my point of view though I am happy that (even according to PECPrep) there was a factor-of-2 reduction in PE. that reduction IMHO is worth the price of PEMPro.

now if I had a CGEM.... (wait, I do) knowing I can reduce the PE from 40" to 25" with the free PECTool.. (BTW PEMPro reported 30" for my CGEM, and reduced it to 14" with PEC training) - so the ratio of improvement is more or less the same, but the absolute values are off a bit. is the price of PEMPro worth it for my CGEM? ..... my opinion, no. If I had the OP's CGEM though.... I might think otherwise.

that said i gotta work on my polar alignment.. i have about 5" of drift every worm cycle....


Post Extras: Print Post   Remind Me!   Notify Moderator  
korborh
scholastic sledgehammer
*****

Reged: 01/29/11

Loc: Arizona
Re: PEPrep help new [Re: orlyandico]
      #5599476 - 01/01/13 12:15 PM

Quote:

hi Ray, the simple curve-fitting in Excel is extremely primitive, and it won't try to fit the higher-order e.g. the 8/3. Really, it will just fit the fundamental.

what i do is plot the raw data and the curve fit. the curve fit is always a sine wave.. (with no 8/3)




Have you tried gunplot? It is free and quite powerful. A bit of learning curve though, however you can find some examples in the doc to get going.
gnuplot


Post Extras: Print Post   Remind Me!   Notify Moderator  
Ray Gralak
Vendor (PEMPro)


Reged: 04/19/08

Re: PEPrep help new [Re: orlyandico]
      #5599486 - 01/01/13 12:18 PM

Quote:

Charlie,

PEMPro almost universally gives better PE values (both pre- and post-PEC/PEM) than say PECPrep. This is why alph always wants to see raw logs...

This could account for the difference in values. Normally a person using PECTool for training wouldn't use PEMPro for measuring the PE... they'd use PECPrep.

I am (right now) capturing data with PHD so that I can verify - with PECPrep - if the reduction from 2.47" to 0.42" (reported by PEMPro) is believable.

In my experience though PECPrep and PEMPro values are largely in agreement.. (although PECPrep still gives higher values)




Having looked at PECPrep's source code I saw that it only uses FFTs, which are only fair approximations. PEMPRo uses FFTs to ballpark the frequencies but then PEMPRo uses much more advanced techniques to get the correct frequencies, amplitudes, and phases. PECPrep's values are probably higher because it is including noise that is removed via PEMPro's more advanced techniques.

-Ray


Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Postmaster
*****

Reged: 08/10/09

Loc: Singapore
Re: PEPrep help new [Re: korborh]
      #5599490 - 01/01/13 12:18 PM

ah gnuplot...

i am an old unix-head. and have used gnuplot in the past. heck one could do the PEC calculations with octave.. but most folks won't be old unix-heads.

imho.. if you have a CGEM, try PECTool first. my experience with it has been fairly good. if you have a higher-dollar mount, just get PEMPro and forget about it.. the mount deserves it.

i would guess once you're at the G11 level just get PEMPro. for the CGEM i'm on the fence about it.


Post Extras: Print Post   Remind Me!   Notify Moderator  
Pages: 1 | 2 | 3 | 4 | (show all)


Extra information
28 registered and 43 anonymous users are browsing this forum.

Moderator:  Dave M, richard7, bilgebay 

Print Thread

Forum Permissions
      You cannot start new topics
      You cannot reply to topics
      HTML is disabled
      UBBCode is enabled


Thread views: 4038

Jump to

CN Forums Home


Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics