Jump to content


Photo

PECPrep help/Aeroquest question

  • Please log in to reply
75 replies to this topic

#51 Ray Gralak

Ray Gralak

    Vendor (PEMPro)

  • -----
  • Vendors
  • Posts: 382
  • Joined: 19 Apr 2008

Posted 01 January 2013 - 01:28 PM

i believe Frank has said as much...

compare FWHM of short (5 to 10 second) exposures with FWHM of long guided exposures, this will tell if the FWHM is atmosphere limited or mount limited..

Frank's the major culprit to whom I was beating this point. I'm glad he finally accepted that. :)

-Ray

#52 korborh

korborh

    Viking 1

  • *****
  • Posts: 895
  • Joined: 29 Jan 2011
  • Loc: Arizona

Posted 01 January 2013 - 01:33 PM

Ray - thanks for your replies. I am going to do some more experiments to gain a better understanding from the observations on my equipment.

#53 orlyandico

orlyandico

    Fly Me to the Moon

  • *****
  • Posts: 5719
  • Joined: 10 Aug 2009
  • Loc: Singapore

Posted 01 January 2013 - 01:39 PM

don't quote me on that...

i vaguely remember Frank saying something of the sort but i'm not about to dig through a bazillion forum posts to find his particular statement.

it does make sense from my perspective which is why i mentioned it.

#54 korborh

korborh

    Viking 1

  • *****
  • Posts: 895
  • Joined: 29 Jan 2011
  • Loc: Arizona

Posted 01 January 2013 - 01:45 PM

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.


As you know every continuous periodic function can be expanded to Fourier series. FFT is a smart and efficient numerical algorithm that decomposes a sequence of values into components of different frequencies thus in essence it expands a periodic error curve (PEC) to Fourier series. This method is well established and reviewed by mathematicians and practitioners. DTF would be more accurate but more time consuming and there would be no practical difference. PEMPro, from what you said in the past, uses proprietary techniques that I am highly skeptical of. Publish your advanced techniques and let specialists review and approve them. In the meantime I will stick to PECPrep and FFT for PEC analysis.


I cannot understand how a well established and robust procedure (FFT) is not enough to model the relatively simple undulations of what we see in a mount periodic curve. FFT is used in complex DSP and spectrum analyzers for many hundreds of MHz of frequencies.

#55 Ray Gralak

Ray Gralak

    Vendor (PEMPro)

  • -----
  • Vendors
  • Posts: 382
  • Joined: 19 Apr 2008

Posted 01 January 2013 - 01:53 PM

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.


As you know every continuous periodic function can be expanded to Fourier series. FFT is a smart and efficient numerical algorithm that decomposes a sequence of values into components of different frequencies thus in essence it expands a periodic error curve (PEC) to Fourier series. This method is well established and reviewed by mathematicians and practitioners. DTF would be more accurate but more time consuming and there would be no practical difference. PEMPro, from what you said in the past, uses proprietary techniques that I am highly skeptical of. Publish your advanced techniques and let specialists review and approve them. In the meantime I will stick to PECPrep and FFT for PEC analysis.

BTW, how did you determine that PECPrep is even accurate? The critical FFT routines are NOT open source. For all you know there could be a bug. The author seems to have a standing policy of letting the users test his code rather than always doing it himself.

-Ray

#56 Alph

Alph

    Surveyor 1

  • -----
  • Posts: 1769
  • Joined: 23 Nov 2006
  • Loc: Melmac

Posted 01 January 2013 - 01:56 PM

I'm not about to give people more things to copy.


You can't stop people from using deobfuscators. :grin:

#57 Alph

Alph

    Surveyor 1

  • -----
  • Posts: 1769
  • Joined: 23 Nov 2006
  • Loc: Melmac

Posted 01 January 2013 - 02:27 PM

BTW, how did you determine that PECPrep is even accurate? The critical FFT routines are NOT open source. For all you know there could be a bug. The author seems to have a standing policy of letting the users test his code rather than always doing it himself.


The reason I touched on this subject is that PEMPro results in too many misfits. I have seen it on my own and other data. On the other hand I compared many results from PECPrep to Software Bisque's PEC software and they were in a pretty good agreement. Looks like we are digressing from the OP's question so this will be my last comment.

#58 Ray Gralak

Ray Gralak

    Vendor (PEMPro)

  • -----
  • Vendors
  • Posts: 382
  • Joined: 19 Apr 2008

Posted 01 January 2013 - 02:54 PM

I'm not about to give people more things to copy.


You can't stop people from using deobfuscators. :grin:

Good luck with that, Adam! :)

-Ray

#59 Ray Gralak

Ray Gralak

    Vendor (PEMPro)

  • -----
  • Vendors
  • Posts: 382
  • Joined: 19 Apr 2008

Posted 01 January 2013 - 02:58 PM

BTW, how did you determine that PECPrep is even accurate? The critical FFT routines are NOT open source. For all you know there could be a bug. The author seems to have a standing policy of letting the users test his code rather than always doing it himself.


The reason I touched on this subject is that PEMPro results in too many misfits. I have seen it on my own and other data. On the other hand I compared many results from PECPrep to Software Bisque's PEC software and they were in a pretty good agreement. Looks like we are digressing from the OP's question so this will be my last comment.

Show me!!

-Ray

#60 Ray Gralak

Ray Gralak

    Vendor (PEMPro)

  • -----
  • Vendors
  • Posts: 382
  • Joined: 19 Apr 2008

Posted 01 January 2013 - 03:06 PM

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.


As you know every continuous periodic function can be expanded to Fourier series. FFT is a smart and efficient numerical algorithm that decomposes a sequence of values into components of different frequencies thus in essence it expands a periodic error curve (PEC) to Fourier series. This method is well established and reviewed by mathematicians and practitioners. DTF would be more accurate but more time consuming and there would be no practical difference. PEMPro, from what you said in the past, uses proprietary techniques that I am highly skeptical of. Publish your advanced techniques and let specialists review and approve them. In the meantime I will stick to PECPrep and FFT for PEC analysis.


I cannot understand how a well established and robust procedure (FFT) is not enough to model the relatively simple undulations of what we see in a mount periodic curve. FFT is used in complex DSP and spectrum analyzers for many hundreds of MHz of frequencies.

There's nothing seriously wrong with FFTs but I think if you did some research you would find that there are definitely disadvantages and limitations to FFTs. There are definitely better (i.e more accurate) methods employed today. PEMPro does use FFT's to approximate and identify frequencies but it uses more advanced methods to actually calculate the magnitudes, phases, and removing noise from the signal.

-Ray

#61 Charlie B

Charlie B

    Viking 1

  • -----
  • Posts: 970
  • Joined: 22 Mar 2008
  • Loc: Sterling, Virginia

Posted 01 January 2013 - 03:43 PM

I am trying to decide if my mount (CGEM DX) would benefit from an aeroquest worm & ring gear replacement. My maine objective is to get a smoother PE curve that is more easily guided (a PE reduction would be nice, but I am mainly looking for smoother). I am using PECPrep to do a little analysis and could use some help understanding the data.



I'm not sure you got an answer to your question, although you got a lot of information on PEC analysis software. According to Ed Thomas at Deep Space Products in an email I sent asking the same question, a hypertune will generally remove about 40% of the PE from a bad mount and maybe less from a good mount. The aeroquest can potentially remove about 1/2 the remaining error, but that depends on other sources of error in the mount. His advice was that most people only need the hypertune and do not need the new worm and ring gear. However, if you want to get the best possible performance from the mount, get the better components.

From looking at your mount PE, which is close to mine, I would not get the new worm. If you are having problems balancing or other similar typical problems reported by CGEM owners, you may want to do a standard hypertune to the mount. However, I would train the mount first, using either Pempro or PECTool, and repeat the analysis that you show with PEC enabled. Your peak PE and rate of change may improve.

Regards,

Charlie B

#62 freestar8n

freestar8n

    Skylab

  • *****
  • Posts: 4051
  • Joined: 12 Oct 2007

Posted 01 January 2013 - 06:34 PM

Hi. I am traveling and limited internet access with a droid so cannot respond much to this thread. I will just say that the goal of guiding is small fwhm and techniques that show measured improvement in star size are what interest me. For
Mid range mounts i focus on prompt corrections with low latency and accurate centroid.

Frank

#63 Ray Gralak

Ray Gralak

    Vendor (PEMPro)

  • -----
  • Vendors
  • Posts: 382
  • Joined: 19 Apr 2008

Posted 02 January 2013 - 12:36 AM

BTW, how did you determine that PECPrep is even accurate? The critical FFT routines are NOT open source. For all you know there could be a bug. The author seems to have a standing policy of letting the users test his code rather than always doing it himself.


The reason I touched on this subject is that PEMPro results in too many misfits. I have seen it on my own and other data. On the other hand I compared many results from PECPrep to Software Bisque's PEC software and they were in a pretty good agreement. Looks like we are digressing from the OP's question so this will be my last comment.

Just a reminder... if you have an example of what you state above I would like to see it (please post the logs). Maybe there's a bug that I need to fix but... if you can't give me an example, maybe you are mistaken?

-Ray

#64 Alph

Alph

    Surveyor 1

  • -----
  • Posts: 1769
  • Joined: 23 Nov 2006
  • Loc: Melmac

Posted 02 January 2013 - 05:24 PM

Just a reminder... if you have an example of what you state above I would like to see it (please post the logs). Maybe there's a bug that I need to fix but... if you can't give me an example, maybe you are mistaken?


Ray,
There is no reason for a concern. Misfits that I was referring to are caused by (or related to) PE's that are not really periodic. I guess there isn't much one can do about it other than removing the 'bad' cycles from the analysis. However I am not aware of an easy way of doing it in PEMPro.

#65 Ray Gralak

Ray Gralak

    Vendor (PEMPro)

  • -----
  • Vendors
  • Posts: 382
  • Joined: 19 Apr 2008

Posted 03 January 2013 - 12:01 PM

Just a reminder... if you have an example of what you state above I would like to see it (please post the logs). Maybe there's a bug that I need to fix but... if you can't give me an example, maybe you are mistaken?


Ray,
There is no reason for a concern. Misfits that I was referring to are caused by (or related to) PE's that are not really periodic. I guess there isn't much one can do about it other than removing the 'bad' cycles from the analysis. However I am not aware of an easy way of doing it in PEMPro.

It depends on what you mean by bad cycles. You mean bad data points in the middle of an Acquire data? If that's the problem then I'm sure you would want to have a clean set of data (i.e. do a new "acquire data" in PEMPro) instead of trying to use compromised data. You should also also be able to delete the offending data (each line is a data point) from the file.

-Ray

#66 Alph

Alph

    Surveyor 1

  • -----
  • Posts: 1769
  • Joined: 23 Nov 2006
  • Loc: Melmac

Posted 03 January 2013 - 01:52 PM

It depends on what you mean by bad cycles.


Worm cycles that appear quite different from others.

#67 Chris Shillito

Chris Shillito

    Explorer 1

  • -----
  • Posts: 51
  • Joined: 03 Jul 2008
  • Loc: UK

Posted 03 January 2013 - 08:20 PM

FFT's are old-school. The techniques are easily found if you do searches for them but I consider the specific methods PEMPro uses as trade secrets (and there will be advances in PEMPro V3). I can tell you the techniques used are related to singular value decomposition. Many tool vendors, include the PECPrep author has copied PEMPro's ideas so I'm not about to give people more things to copy. :)

-Ray


BTW, how did you determine that PECPrep is even accurate? The critical FFT routines are NOT open source. For all you know there could be a bug. The author seems to have a standing policy of letting the users test his code rather than always doing it himself.

-Ray


For the record PECPrep is an original work that was inspired by PEAs and the requests of our users. No reference to PEMPro was made during its development and PEMPro remains software which I have never installed or used (nor I doubt ever will). That said I am aware that the layout of the frequency spectrum screen in PECPrep has a remarkably similar equivalent in PEMPro ?? perhaps its just a case that two applications doing the same job will inevitably take on a similar appearance (unless you know otherwise?)

I have no interest in copying your work or in using experimental/unattributed algorithms. FFTs are an industry standard, well proven technique for vibration analysis and thus far have served us well in the analysis of telescope mount periodic error which is hardly the most challenging environment it which FFT analysis has been successfully applied.

PECPrep is tested by me before issue to our EQMOD group for further testing prior to full public release via sourceforge.

For a commercial vendor such as yourself to cynically attempt to discredit the efforts of a community based project is a disgrace and I trust all can see your true colours. I'm frankly surprised that the moderators here allow such behaviour.

Chris.

#68 Ray Gralak

Ray Gralak

    Vendor (PEMPro)

  • -----
  • Vendors
  • Posts: 382
  • Joined: 19 Apr 2008

Posted 04 January 2013 - 12:21 AM


FFT's are old-school. The techniques are easily found if you do searches for them but I consider the specific methods PEMPro uses as trade secrets (and there will be advances in PEMPro V3). I can tell you the techniques used are related to singular value decomposition. Many tool vendors, include the PECPrep author has copied PEMPro's ideas so I'm not about to give people more things to copy. :)

-Ray


BTW, how did you determine that PECPrep is even accurate? The critical FFT routines are NOT open source. For all you know there could be a bug. The author seems to have a standing policy of letting the users test his code rather than always doing it himself.

-Ray


For the record PECPrep is an original work that was inspired by PEAs and the requests of our users. No reference to PEMPro was made during its development and PEMPro remains software which I have never installed or used (nor I doubt ever will). That said I am aware that the layout of the frequency spectrum screen in PECPrep has a remarkably similar equivalent in PEMPro ?? perhaps its just a case that two applications doing the same job will inevitably take on a similar appearance (unless you know otherwise?)

I have no interest in copying your work or in using experimental/unattributed algorithms. FFTs are an industry standard, well proven technique for vibration analysis and thus far have served us well in the analysis of telescope mount periodic error which is hardly the most challenging environment it which FFT analysis has been successfully applied.

PECPrep is tested by me before issue to our EQMOD group for further testing prior to full public release via sourceforge.

For a commercial vendor such as yourself to cynically attempt to discredit the efforts of a community based project is a disgrace and I trust all can see your true colours. I'm frankly surprised that the moderators here allow such behaviour.

Chris.

Sigh... here we go again... I didn't say anything that isn't true. If you think I did I would like you to tell me exactly what it is and I think I'll prove you wrong. You claim your source code is open source but the critical routines are not open source. How can you guarantee your routines say what you say they do? How is your guarantee any different than mine?

I'll tell you.. the difference is that the PEC curves created by PEMPro have been written back to thousands of mounts and the proof is in the results of happy customers. I have tried PECPrep and your PEC code in EQMOD and I have told you it does NOT work very well. You've refused to listen so what can I say. I've even asked if I could contribute fixes to the code base but you don't respond. How is it a community project if you are the only person that is allowed to contribute code?

And, Mon, who originally developed most of EQMOD, contacted me about making PEMPro work with EQMOD before you had even started work on your PEC code. I find it very hard to believe that you never looked at PEMPro's documentation nor even had any influence from Mon to try to copy some of the features of PEMPro. I think that PEMPro was, and still is, the leader in this area and for you to say you never heard of it or at least knew of what it does is quite astounding, especially since you heard of PEAS, which was created after PEMPro. Maybe you were you just really green in your knowledge of software for astronomy??

That said, PECPrep is not really a competitor to PEMPro. PEMPro has camera control and provides support and training capability for every type of mount with PEC. PECPrep is mostly an analysis tool for everything but some Synta mounts. PEMPro uses modern signal analysis techniques (not EXPERIMENTAL) as well as older techniques such as FFTs. If I wanted to use FFTs throughout PEMPro I could have used but I didn't. So you are welcome to remain ignorant of newer, better techniques, but I think that using new and better techniques often distinguishes a good software product.

-Ray

#69 orlyandico

orlyandico

    Fly Me to the Moon

  • *****
  • Posts: 5719
  • Joined: 10 Aug 2009
  • Loc: Singapore

Posted 04 January 2013 - 12:46 AM

I wasn't expecting such acrimony to come out of this discussion.

PECPrep is a good tool, and one I use extensively.

But so is PEMPro. Different tools for different needs. To take an example from my line of work.. nobody would tell Bank of America to run their core banking system on MySQL.

In the same way, I doubt there would be many Software Bisque or Astro-Physics users who would use PECPrep. But EQMOD users (for whom the tool was originally designed) - sure.

I won't get into a value judgement over which is "better" - but PEMPro is definitely more complete and mostly automates the entire training process, including capturing PEC data. Not really comparable to PECPrep.


.. and on a tangent.. back when I was doing this sort of stuff, the "gold standard" open-source FFT implementation was libfftw. Last I checked that was available with source..

#70 Ray Gralak

Ray Gralak

    Vendor (PEMPro)

  • -----
  • Vendors
  • Posts: 382
  • Joined: 19 Apr 2008

Posted 04 January 2013 - 12:56 AM

It depends on what you mean by bad cycles.


Worm cycles that appear quite different from others.

The frequencies that cause worm cycles to be different are non-integer fundamentals because they don't repeat an even number of times in a worm cycle. PEMPro purposely and by design does not automatically include non-integer fundamentals because they cannot be corrected by PEC so there is no point to include them except if you want to measure them.

That said, you can add non-integer frequencies if you want, just for the purpose of measuring them.

-Ray

#71 orlyandico

orlyandico

    Fly Me to the Moon

  • *****
  • Posts: 5719
  • Joined: 10 Aug 2009
  • Loc: Singapore

Posted 04 January 2013 - 01:03 AM

anecdotal evidence. I spent a good deal of last night PEC-training my Littlefoot controller with PEMPro.

according to PEMPro the fundamental was 11" to 13" p-p. PECPrep gave a similar figure, about 10" (the two programs also mostly agree on my CGEM mount).

however after PEC training PEMPro was now saying the PE was 3.5" but PECPrep said it was 15". Quite a head scratcher there - but it was glaringly obvious from an eyeball of the raw curve that the PE had been reduced.

I have been using PECPrep as a sanity-check on PEMPro. For example PEMPro reported a raw PE of 2.6" on my other mount, reduced to 0.42" after PEC; PECPrep reported 5" raw and 2.25" after PEC. While these two sets of figures don't agree that well, they do agree that the PE was reduced.

This was not the case with last night's exercise...

#72 Alph

Alph

    Surveyor 1

  • -----
  • Posts: 1769
  • Joined: 23 Nov 2006
  • Loc: Melmac

Posted 04 January 2013 - 01:23 AM

That said, PECPrep is not really a competitor to PEMPro.


No, it is not, when it comes to programming mounts. However it is a formidable competitor when it comes to analyzing PEC and you know that very well. It is actually the best tool out of the box for analyzing PEC.

#73 Ray Gralak

Ray Gralak

    Vendor (PEMPro)

  • -----
  • Vendors
  • Posts: 382
  • Joined: 19 Apr 2008

Posted 04 January 2013 - 02:17 AM

That said, PECPrep is not really a competitor to PEMPro.


No, it is not, when it comes to programming mounts. However it is a formidable competitor when it comes to analyzing PEC and you know that very well. It is actually the best tool out of the box for analyzing PEC.

Analyzing PEC? Analyzing "Periodic Error Correction"? Is that really what you meant to say? Maybe you meant to say "analyzing PE (Periodic Error)"?

PEMPro was purposely designed to hide many unnecessary details. Have been developing commercial software for over 35 years I have seen many a person confused by tools with too much information. But maybe you're the 1 in 100 that likes that (good for you!).

If so, you may be interested in the free PE analysis tool I've been slowly working on for the last 2 years. It has a modern GUI and (I think) more power than any other free PE analysis tool. I plan to launch it simultaneously with PEMPro V3. Maybe you'll change your mind when you see the capabilities of that tool. Again, it will be free to the community and process logs in many formats, including PEMPro's own log files.

-Ray

#74 orlyandico

orlyandico

    Fly Me to the Moon

  • *****
  • Posts: 5719
  • Joined: 10 Aug 2009
  • Loc: Singapore

Posted 04 January 2013 - 03:44 AM

I think PECPrep doesn't use the inputted # of worm teeth in the calculation. Because in my experience it never gets the worm periods correctly, even for the fundamental.

(and I can provide logs)

The calculated worm periods are always off by a bit (say 455 seconds vs 449 seconds). I am not sure if PEMPro "hard fits" the calculated periods into what it knows, or if PEMPro really manages to calculate them exactly.

#75 ccdmaker

ccdmaker

    Lift Off

  • -----
  • Posts: 20
  • Joined: 15 Nov 2011
  • Loc: Illinois, USA

Posted 05 May 2013 - 05:59 PM

Hi Orlyandico,
Do you mind posting the part number of the Vexta motor(/gearbox) that you picked up on ebay? I am thinking of retrofitting my Vixen GP. have already built a stepper driver that works well with the stock MT-1 motors, but the numerous gears in the motor housing (1:120)are introducing a lot of high frequency components. Beside, a smaller gear ratio and some microstepping should enable somewhat faster slew speeds too.
Thank you.






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics