EQ6 Pro: PEC vs Autoguiding

[MarkS]

Thanks for your comments about my Horsehead Nebula - it's clear that I need longer subs.  Without investing a lot of money to get myself autoguiding (remember my focal length is 2800mm), I want to see just how far I can get using the Periodic Error Correction on my EQ6 Pro. 

Mon & Wed evenings this week (in the freeeeeezing cold!!) I played around with the PEC but with only partial success.  Without error correction, I was seeing a total tracking error range of around 30arcsecs over the 8 minute worm period.  Though "lumpy", this was nicely periodic for 3 or 4 cycles.  After recording an error correction with the handset, the PEC tracking worked extremely well for the following 20-30 minutes, keeping the total error within just a few arcsecs.  After that, tracking errors slowly built up again.  I produced some interesting PE charts demonstrating this, which I'll post later.

My guess is that those errors are due to other (non-worm) inaccuracies elsewhere in the gear train.  If so, it may well be the case that these errors are perfectly repeatable if the gear train is wound back to its original position and then re-run.

So, if my hunch is correct, I could do the following to image the Horsehead with long subs:

1) Point scope towards the Horsehead, then use a nearby star to record a PEC.
2) Expose 20 minutes of  Horsehead subs with PEC switched on.
3) Slew the scope back to the start point - this returns the whole gear train back to its original configuration.  But then line the OTA up with the Horsehead again by releasing the R.A. clutch and spinning it forward.
4)  Expose another 20 minutes of  Horsehead subs with PEC still on.

Repeat steps 3 & 4 until bored (or frozen!).

I'll give this a go at the weekend (assuming a clear night) and let you know what happens.

[JohnP]

Hi Mark - sounds possible but also souns like hardwork... it'll be interesting to see the results... John

[Mike]

Why not just guide instead? You've got a small scope and a webcam I presume?

[Ian]

I would think that getting a focal reducer would help too. Imaging LX at 2800mm is something to work up to...

[MarkS]

Mike, no I haven't yet got a webcam or a laptop - but you never know what Santa might bring!  I could use my old scope as a guider though.

Ian,  you're right - I am using an F6.3 focal reducer so that brings down the effective focal length from 2800mm to around 1800mm.

Here are the plots before and after Periodic Error Correction (I wasn't very careful about precise focus):

Before PEC:


After PEC:


The traces are 60 arcsec vertically and approx 25min wide - each vertical pixel is 1 arcsec. For the "after" shot, the PEC had been running for about 5-10 minutes before I started the trace.  But you can see how the oscillations begin to build up after 10minutes of the trace.

[Mike]

Seems like a very high amount of PE to me. The line on the second graph seems to match the first after a while. Are you sure that the full PEC data was recorded?

[MarkS]

Yes, I definitely recorded the full 8 minutes of error correction.  In fact, you can see in the second graph that the first 8-10 minutes is pretty flat.  You're absolutely right that the second graph begins to match the original after a while - that's an interesting thing I hadn't noticed.

Is there a gear train between the stepper motor and the worm (I'm guessing there must be).  Then is it possible to find out what those gear ratios are?  If so, it might be possible to match their frequencies of rotation against the graph.  It wouldn't help much but it would be an interesting thing to do.

[Mike]

Mark there is an exploded view of it here - http://avex.org.free.fr/dossiers-pratiques/eq6/index-eng.html

There is the stepper motor, then a worm, then a gear wheel.

[MarkS]

Thanks for that link Mike - the graphs and other info there are very instructive.  Is the drive chain of the EQ6 Pro identical to the EQ6 featured there?

In any case, what the PE graphs and FFTs on that site demonstrate is that the gears make a large contribution to the total PE.  That is not a huge problem in itself - it's the periodicity of the gearbox error that is the problem.  Take a look at the FFT: there is a substantial contribution with periodicity around 140sec.  The big problem with 140sec is that the worm period of 480sec is not exactly divisible by 140.    So when a PEC (of 480 sec) is recorded it can cancel out the worm period error but can never cancel out the 140sec period error.

In brief, it is physically impossible for PEC to cancel out PE arising form the gearbox if that gearbox error has the wrong frequency.  And that is probably what I'm seeing in my own graphs.

[Mike]

Hmmm.. Interesting.

I have only ever done the PEC once on my EQ6 and what a pain in the ass job that is !! I can't say it made a huge difference to be honest. Guiding is the way to go. Best start saving your pennies for a laptop and webcam !

BTW - You are more than welcome to pop around one night and I can give you a guiding demonstration on your own equipment using my laptop if you like.

[Rocket Pooch]

Hi MarkS,

The EOS does have a big chip so you will get away with some creeping in the images, however you are going to be in the .81 arc seconds per pixel.  This is ok for guiding, but the EQ guys on the EQ Mod group at best PEC they have got is about 2 arc seconds PP.  Guiding I normally get about .4-.7 arc seconds with my 80mm guide scope, I think this is going to be the best route to go, but the C11 is on the upper limit of the EQ6 so I'm not sure you will get as low a PE?

Chris

[MarkS]

Autoguiding is definitely the way to go but until I've saved my pennies I want to see just what I can achieve without.  Chris, your accuray of 0.4-0.7 arcsec is really impressive.

Looking at the exploded view on http://avex.org.free.fr/dossiers-pratiques/eq6/index-eng.html I can see that the final gear ratio before the worm is 10/34.  This means the shaft before the worm has a rotation period of 480*10/34 = 141.2 sec.  On the FFT (before he installed replacement gears) there is a very large spike at exactly that point.  So one of the gears on that particular shaft is causing a large proportion of the PE.  The next ratio is possibly 11/28, I can't see exactly, but it would account for the  spike aroung 55 sec.  But I can't see enough teeth of the other gears to continue the analysis.

I've taken 5 charts taken with my own mount. I'm not seeing a period of 141.2 sec but I am seeing a nice sinusoid around 120-123 sec.  I'll try to get an accurate measurement, maybe tonight.  Maybe the EQ6 Pro has redesigned gears.

[MarkS]

I've now had a chance to measure the period of that sinusoid.

It is 122.4sec (+/- 0.5sec ).  Since this does not divide into 480 (the worm period) it can never be fully cancelled out by PEC - in fact PEC can make it worse: sometimes the PEC waveform will destructively interfere with it, cancelling it out, 25 cycles later it will constructively interfere causing its peak to peak amplitude to double to around 14-20 arcsecs. 

It entirely explains the increasing amplitude oscillation seen in the trace I recorded with PEC switched on:


So guiding here I come ...

[Ian]

got to wonder where it's coming from if it doesn't appear to be at a frequency that matches any of the drive train...

[MarkS]

It has to originate in the drivetrain.  I guess it's in the gears between the stepper motor and the worm.  If the final gear ratio before the worm is, for instance, 13/51 than that would give a period of 480*13/51 sec = 122.35 which is very close to the 122.4 that I am seeing.  But I don't intend to take it apart just yet to see if that's the case.

The original EQ6 had a final gear ratio of 10/34 but maybe the EQ6 Pro has been re-engineered.