PEC Curve smoothing

[IvanT]

Hi,

I made my first serious attempt at capturing my mounts PE tonight. I was wondering if any Guiding Gurus would be interested in analysing my PEC data and seeing if they can "smooth the curve" better than my feeble effort.

I've attached 3 files:

1) The raw capture data, it's about 50 mins worth (PHD_log_20Jan11)

2) My attempt at processing it and smoothing the curve (PHD_log_20Jan11_RA_EQMOD)

3) And lastly the PEC output file which I imported into EQMOD (pec)

After importing it, I tried guiding with Sidreal+PEC, the screen shot shows that it wasn't very good.

I'd appreciate any pointers or improvement files if possible.

Thanks

PHD_log_20Jan11.txt

PHD_log_20Jan11_RA_EQMOD.txt

pec.txt

[chrisshillito]

The screen shot show you have the PEC gain set to x0 which effectively disables PEC set it to x1. When PEC is active you will see the numbers next to the PEC display change indicating the level of rate correction (in arcsecs per sec) being applied.

Chris.

[IvanT]

Thanks Chris, I didn't notice that. That's one thing to try.

I followed the steps in the PHD section from the EQMOD VS-PEC document The document shows the output as a nice smooth curve on page 12 which translates to a waveform in PHD, as can be seen in my screen shot, mine is anything but smooth and wavey

Here's the steps I followed:

Stark Labs PHD

1. Find target star and start tracking

2. In PHD make sure the log info option is checked.

3. In PHD make sure the ”disable guide output” is checked or In EQMOD uncheck the

pulseguiding enable boxes. (see note below re: DEC guiding)

4. Delete/move/rename any existing PHD logfile for the current date.

5. In PHD set the maximum search region you can get.

6. Calibrate in PHD then start guiding.

7. Capture 5 or more PE cycles.

8. In EQMOD hit the [TimeStamp] button on the PEC display.

9. Stop guiding.

10. Run PECPrep.

11. Make sure PECPrep has the correct mount type selected.

12. Load the PHD file from the file menu.

13. When prompted the focal length you were using, the pixel size of the imaging device,

binning and star declination. The pixel width corresponds to the ‘X’ display axis.

14. Select the “apply EQMOD Timestamp” option and select your PHD log file. PECPrep

will automatically convert the PHD log file to perecorder format, load the file PECPrep

and save the new format file with “_EQMOD” appended to the filename. If you wish to

come back to this data at a later time you can load this _EQMOD file directly without

the need to go through the conversion process again.

[chrisshillito]

Well the high frequency wiggle is most likely the 10.2 second toot contact period.

Did you read the pecprep manual (http://eq-mod.sourceforge.net/docs/PECPrep.pdf)? It describes in detail how you can use the lowpass, high pass and magnitude filters to preprocess your PEC curve - look out for the worked example of PEC file generation (pg17). Also check out the

on creating PEC curves with

Ultimately the shape of your PEC curve will depend on the characteristics of your mount. I'm afraid with mass produced lower end mounts like the EQ6 you get quite a lot of variation in shape of PEC curve from one mount to another. Some like mine are quite sinusoidal, others have all sorts of low freqency bumps in them caused by the modulation of the worm with non harmonic elements like the transfer and stepper gears (like the one I used in my video).

Chris.

[chrisshillito]

.. oh and I should add that if/when you rework your data you just load the PHD_log_20Jan11_RA_EQMOD.txt file as a perecorder file (saves you having to do all the PHD specific stuff and apply the timestamp etc.)

[Neil I]

Hi Ivan

I've just attempted this as well, using a different app but with very similar results as you. Here's the link:

http://stargazerslounge.com/equipment-discussion/126352-neq6-pro-pec.html

The curve shown in your eqmod screen shot looks identical to mine as well.

Cheers

Neil

[IvanT]

Did you read the pecprep manual

Nope, the doc I used doesn't have the worked example. That & those excellent youtube vids will hopefully make things a lot clearer for me.

Thanks!

[IvanT]

The curve shown in your eqmod screen shot looks identical to mine as well.

Sorry, but you don't know how good that makes me feel I like you am doing this as I recently completed my pier. I used EQAlign (two iterations so far) to tighten up my PA and the 50 mins or so guiding showed only a little movement in the star, I'm pretty sure that most if not all movement is down to PE.

I was expecting to get a nice smooth curve in EQMOD but I think it's more down to my not understanding how to manipulate the data. It's why I posted the files to see if someone fancied a go & confirm that I'm a tube The link Chris provided will hopefully explains things better for me.

Am I correct in saying that when you've got PE setup that the mount should always be parked and clutches never disengaged for manual slew as this chance of putting the worm back precisely where it was is slim?

[chrisshillito]

I was expecting to get a nice smooth curve in EQMOD but I think it's more down to my not understanding how to manipulate the data.

If you look in the photo gallary of the EQMOD group you'll see a collection of PE traces - mostly of EQ6Pros.

http://groups.yahoo.com/group/EQMOD/photos/album/2026286228/pic/list

all are different so please don't be overly concerned if yours looks like no other!

Am I correct in saying that when you've got PE setup that the mount should always be parked and clutches never disengaged for manual slew as this chance of putting the worm back precisely where it was is slim?

You're half correct - for PEC sync to be maintained you should always park and unpark. Moving the mount with the RA clutch slackened will not move the worm or RA motor so doesn't affect PEC - it does of course screw up any alignment model you may have saved.

Even if the worm does move for some reason there is a way of recovering synchronisation. What you do is remove the RA axis worm end cap and note/mark the position of the flat on worm shaft when parked at "home" (or any other favorite park position). If for some reason your worm has moved between sessions you just start up, unpark and slew the mount until the flat aligns with your mark. You then use the "resynch encoders" button (having made sure the appropriate park position was selected in the park dropdown).

Chris.

[chrisshillito]

Ivan,

Can I just check that the parameters you entered into the PECPRep PHD import dialog (i.e. focal length, pixel size, declination and binning). Looking at your data it would appear that the raw PE was only logged as around 3 arc secs peak to peak which would be quite amazing for these mounts if it is true.

Chris.

[IvanT]

Ivan,

Can I just check that the parameters you entered into the PECPRep PHD import dialog (i.e. focal length, pixel size, declination and binning). Looking at your data it would appear that the raw PE was only logged as around 3 arc secs peak to peak which would be quite amazing for these mounts if it is true.

Chris.

Ah, ermm, hmmmm, you see, it's like this...... I'm a twit !!!! I've entered the focal length of my 200P 1000 instead of the FL of the 9x50 straight through finder scope. That may have had an impact. Does that count as a bona fide brain fart?

The values I entered were:

Data to load: RA Error

Focal Length (mm): 1000 (what would the 9x50 be do you know?)

Pixel Size (um): 5.6 (think that's correct for the QHY5?

Binning: 1x1

Compensate for DEC (un-checked): 0

Apply EQMOD Timestamp: Don't recall if I ticked it or not.

[IvanT]

Just running through the PECPrep YouTube videos now. One piece of info I'm not sure of is what to enter for the focal length of the SWE 9x50 straight through finder with QHY5 attached (no barlow etc). Is it simply the 118mm end-to-end body length of the finder (not counting the camera depth) that I need to enter? Also, the DEC compensation value, is the format of that along the lines of 45:19:32 as displayed in EQMOD?

[chrisshillito]

Ah, ermm, hmmmm, you see, it's like this...... I'm a twit !!!! I've entered the focal length of my 200P 1000 instead of the FL of the 9x50 straight through finder scope. That may have had an impact. Does that count as a bona fide brain fart?

The values I entered were:

Data to load: RA Error

Focal Length (mm): 1000 (what would the 9x50 be do you know?)

Pixel Size (um): 5.6 (think that's correct for the QHY5?

Binning: 1x1

Compensate for DEC (un-checked): 0

Apply EQMOD Timestamp: Don't recall if I ticked it or not.

Oh OK, you've gone wrong on several fronts then.

Yes, you need to enter the focal length of the instrument to which the QHY5 was attached. To work out the precise focal lenth of your finder objective you would need to point it at a light source and project the image onto a piece of card - measure the distance between lens and card for focuser - my bet is it will be around 200mm.

That said, For the purpose of PE capture why not use your main scope? You may as well get as much resolution in the data as possible.

PHD only records movement in terms of pixels. Stars at DEC=0 therefore will produce far more pixel movement for a given angular movement than those up at say 60 degrees (in fact twice as much).When using a PHD log you must therefore check the compensate for DEC box. If you are applying an EQMOD timestamp then provided this "compensate" box is checked the declination will be read out of the timestamp itself (EQASCOM stores it there automatically) and you don't have to enter a number. If not applying a timestamp (i.e. just doing PE analysis rather than generating PEC curves) the enter the DEC value to the nearest integer of degrees.

Please note for PHD imports the taking and applying "timestamp" part of the procedure is critical if you are intending to generate PEC curves. For PEC we need to relate a given worm position with a PE. PHD itself does not read motor/worm positions from EQASCOM and so the only way to establish this relationship is to use this "timestamp" mechanisim whereby the motor position at a fixed time is written to disk. PECPRep can then use this, in combination with the PHD log start time, to work out where the worm was for each log entry.

With other applications like perecorder and Metaguide things are much simple as these are able to read the motor/worm positions directly and write them into their logs. They also log data in terms of arcseconds so as result timestamps and declination compensation are not required at all. Of course your particular camera type may not be supported by these applications.

Anyhow, assuming your original timestamp is still valid (i.e. you haven't made another since ) then you can just reload your PHD log into PECPrep with the corrected value for focal length and with the compensate for dec and apply timestamp buttons checked.

Chris.

[IvanT]

Thanks Chris, appreciate you taking the time to help explain all this to me. I re-did a capture last-night and will process it with corrected values. As for why not use the main scope to capture the data, that didn't occur to me, would you recommend with or without a 2x Barlow?

Sent from my iPhone using Tapatalk

[chrisshillito]

As for why not use the main scope to capture the data, that didn't occur to me, would you recommend with or without a 2x Barlow?

You could use a barlow if you wish. It will bprobably all depend on how well polar aligned you are as you need to get 5 cycles of data and in that time you might experience a fair amount of drift (potentially both DEC and RAb depending where you're pointing) which can move the star off chip at higher focal lengths. I my observatory I capture my PE using a web cam at prime focus with my C9.25 which is fl 2350mm.

Also if you're after improved spectral analysis from PECPrep then increase the PHD sample rate by using a bright star and short exposures. The shorter the period between log samples the the greater scope PECPrep will have in diagnosing any high frequency issues present. Of course this isn't such an issue if you're simple interested in building a PEC curve as you will be filtering out all those high frequency bits anyway but I always take the approach that its better to capture data and be choose to discard it, than not see it at all.

Chris.

[bamus]

Hi All

I have recorded 6 PEC curves last night and today in the morning train to work I have processed it and tried to apply PEC file in EQMOD. Unfortunately EQMOD didn’t want to load my saved PEC.TXT file ( no error message, but no action as well). I think there are 2 options:

- EQMOD must be connected to mount in order to apply PEC file

- I have done something wrong with TIMESTAMP (I have followed instructions and HIT Time STAMP in EQMOD just after starting guiding and shortly before finishing guiding)

Any other Idea what could go wrong?

thanks

Adam

[chrisshillito]

Hi Adam,

I don't believe EQMOD needs to be connected to the mount to load the PEC file (can you load your pec file into the simulator?). How are you starting EQMOD if it isn't connected to the mount -most clients will shut it down is connection isn't established. Any chance you can attach the PEC.Txt so I can take a look at it?

Chris.

[bamus]

Thanks Chris for your replay. Have to wait till break time to turn my netbook on and to see how I have managed to turn on EQMOD without being connected to mount and check if simulator accept PEC file.

I will upload pec.txt file in the evening as my netbook hasn’t got wireless internet connection here in work place.

[chrisshillito]

Adam,

Just did a quick check and yes EQASCOM does need to be connected to the mount in order for the PEC file to be loaded (there is a check made during the load that the PEC file data matches the mount type). The file should load into the simulator ok.

Chris.

[bamus]

Hi Chris

Its working under simulator, thanks!

in the evening at home will try to connect it real mount, but now it should work. According to PECPerp by applying PEC in EQMOD I should go down with current PE (-20.73, + 23.49) to PE-PEC +11.95 to -12.84. I know its is not great but it would be huge improvement and I hope I could do bit longer unguided exposures with my WO72 on Eq5.

I will do one more PEC recording to check if I will get same results, as I have to admit RAW PE looked pretty rough with PE jumping from -30 to +35.