PEC am I missing something?

[Doug]

Recently I purchased an EQ6 goto mount and although it is intended to give me some portabitity, I have mounted it on a pier to see just what its strengths and seaknesses are before venturing out into a dark sky site.

Ok, so I have spent a fair amount of time drift aligning it; I can not see SCP because of huge cypres Pine:sadeyes: , I can't use a compass because of close proximity of colour bond fences and pool safety fences:sadeyes: . I used the goto function to locate SCP for me and then drift aligned it until it is drift aligned period.

A 1 star alignment puts Sirius onto a webcam chip using an ED80.

I have tried PEC but with limited results.
I'm wondering if I have been holding the manual upside down or something; here is what I do:
Center a star using webcam and K3CCD-tools, with crosshairs active and zoom at 'full screen', I then start pec training. The star image is kept to within its own dia. , (about 1/3 at most). When complete I select Tracking speed =Siderial +PEC. But the Star image still wanders East/West by about 1 dia. or more.
I'm wondering what other EQ6 owners are doing with PEC?

Below are a couple of test images of (more or less) NGC2422. The first is 120 sec unguided, the second is a median combine of 6 unguided 120 sec + 1 240 sec unguided image, manually aligned. Of the series, 3 images looked more like a basket of eggs than a star field and were discarded.
If I could get PEC to replay as it was recorded, I could image unguided, but as is I can't. All these tests are done before the mount is parked and switched off.

Any suggestions welcome. (the images are for guiding tests not images for images sake)

cheers,
Doug

[Doug]

I forgot to mention that although I pec trained using a webcam, the images were acquired using a Meade DSIc and Envisage S/W.


Doug

[[1ponders]]

No matter how good your pec is Doug, for long shots (120sec + ) I believe you will still need to guide. The pec will take care of your worms periodic error, but you will also have some non-periodic errors as well. And no matter how well you drift align to get your polar alignment as close as you can, there will always be some small amount of movement in your mount that will cause drift errors. It is just isn't possible to build these errors out of a mount or scope.

[Doug]

Thanks for your reply Phil, I agree, however until an easy guider glides my way from Lumicon, I can't auto guide with the equipment I have. My point is though that the PEC playback is nowhere near as good as I beleive the recording to be, and I am wondering if others are finding this too, or is it just me?

Doug

[[1ponders]]

Is the PEC updateable? By that I mean, with my old LX200 if I did the PEC on one night, I could do another training on another night and then average the two, and then another one and so on. It improved the final PEC, but it still wasn't perfect, I still had to guide for longer exposures.

btw, do you at any time release the clutches and move the mount by hand?



ps. I dont know the EQ6, so I'm only looking at this from a generic point of view.

[[1ponders]]

bbtw have you measured your periodic error?

[Doug]

No I haven't measured it yet, K3CCD is supposed to do it but as I only downloaded it yesterday, I have yet to figure out how to do it; but it is significant. Mind you 70% or so of the worm is very good.

Unfortunately unlike the Meades, Pec is not refinable, nor it seems can I use ccd guiding to pec train which can be done with the LX200. The other draw back is that unless you park the mount before removing power, the synch between the HC and the worm encoders will be lost, requiring a fresh PEC training session. It is in this area that I suspect I have problems, maybe I'm reading the manual upside down or something, I'm not sure. Maybe I have more drift (masked during refresh times) than I am aware of, I just don't see how I could be that far out and not know it.

Doug

[[1ponders]]

To use K3 to measure your pe, go into Option> Telescope and CCD settings; input your telescope aperture and FL and your CCD details (ToUcam: 5.6x5.6 micron pixels) and then save the settings. You will get your resolution per pixel at the bottom as well as the FOV for the chip.

Once saved close that box, slew to a star and center it. Make sure that you camera is "square" to the imaging train. By that I mean when you press the RA buttons on your controller the star moves left and right on the screen and not up and down. If it does go up and down rotate your camera 90 deg. To avoid going through the process of measuring your camera angle try to get the camera as close to square as you can. To help with this I'd suggest clicking on the reticle icon on the tool bar.

Once you are square, start the drift explorer, roughly center your star, click on the target icon on drift explore and then click on your star. Your dec error and RA motion will start to appear on the graph. You can log this to open later and graph in Excel if you want to. I'd suggest just watching it for a while to get some idea. The numbers down the side are pixels and the numbers across the bottom are sec.

If you want to recenter the graph, click on the second icon from the left in the drift explorer tool bar. Under the tool bar you will see various readings for RA and DEC. I tend to ignore the RA ones and concentrate on the DEC. The figure immediately under "DEC" will give you your drift in arcsec/min.

To manually work out your RA error watch for the max and min readings for a full gear rotation and then multiply them by your pixel resolution (under options> telescope and ccd ...).

Hope you can follow that. If not let me know ;)

[Doug]

Thanks phil, I'll try that when the sky clears later this week sometime.
Clear as a bell *pm tonight clouded out by 9:30:(

Cheers,
Doug

[garymck]

Doug,

unguided imaging with anything less than an ap or paramount is unrealistic. Everyone autoguides because pec is not perfect. There are a number of issues.

Firstly with the eq-6. This is a mount built to a budget, most have a pe of under +/- 10 arc secs. However while the general profile is repeatable there are variations between the individual teeth of the worm gear that cause slight variations from worm cycle to worm cycle. In addition there are small random errors of an arc sec or so caused by roughness in the actual worm. There are also variations in spacings of the worm wheel from the worm due to manufacturing tolerances that can lead to the exaggeration of these random errors. The actual worm gear is not hobbed with an exactly matching profile to the worm. This allows for greater freedom in putting the mount together, but exaggerates the pe when there is a tooth to tooth transition. These things are typical of mid to low end mounts.

Essentially what I'm getting at is that pec assumes that the gear errors are perfectly reproducible, and I've given reasons why in the eq6 they aren't.
Below is a link to a survey of improvements from pec from the PEMPRO site :

http://www.ccdware.com/products/pempro/examples.cfm


You will note that none of the mounts has zero pe despite having programmed pec. Though I gotta say several are pretty damn close and below my seeing levels in Melbourne!!! But most of these mounts would not be able to take lengthy unguided images unless it was at a short focal length....

Recording pe for correction is difficult. Everyone forgets that when recording pe for correction you have to see the error occur, then issue a correction. This takes time, the mount may have already moved further after you issued the correction, you are always chasing the error, and always behind it. So when you program it, it replays always a fraction late (unless you've a means of editing and shifting the timing of the pe corrections.) Hence you are unlikely to be able to image for any long period of time.

Generally the eq6 is a good mount - I have had one that was superb, and have another that is a total dud, - and awaiting parts from Tasco to fix it (They have been very pleasant to deal with and I have no gripe with them at all) - and I have another friend with one that is just as good as my first one was. It is through analyzing this mount that I learned much about gears and pe.

BTW, you do not need an expensive extra scope to autoguide, go to:

http://www.skyinsight.net/astrophoto/current.php

and download the latest issue (free) and read the article on building a compact robust guidescope with a 40mm objective!!!

hope this helps

fwiw

Gary

[Stephan]

Gary,



Where did you get this number from? The only internet-published PE-recordings I found showed PE's of around +/- 20 arc Secs.

Stephan

[garymck]

Stephan,

from my own measurements of several local mounts, correspondence with people o/s, plus comments from yahoo group.

The mounts with +/- 20 arc sec errors are faulty. The problem with them is in faulty machining of the worm.

The worm is machined and then the end journals that fit into the ball bearings are ground. The problem is that some mounts have the journal on the end that the spur gear fits ground off center. You can fairly accurately predict the periodic error of an individual mount by measuring with a dial guage and v blocks the wobble in the worm. The good ones have about 1/2 a thou or less of wobble which seems to fairly reliably predict a total pe of around 10-15 arc secs. Those that have a total of 40 arc secs will have a wobble of about 2.5-3 thou. The problem with the larger wobbles are that the speed of the pe becomes horrendous. My "dud" mount moves 20 arc secs in 30 seconds on 2 places in the worm cycle - autoguiding simply cannot keep up with this. The actual worm thread is very well machined, but has some mild surface roughness.


cheers
Gary

[Stephan]

Gary,

Could you post an image of the PE-curve for an average EQ-6 and maybe one of a really bad EQ-6?

Stephan

[garymck]

Stephan,
don't know how to take screen shots, can you tell me how to do it? If you can I'll take a shot of a pempro screen

Gary

[iceman]

Press alt-printscreen together.

Then paste (ctrl-v) it into a picture program, resize as required, save as web size.

[garymck]

Image of a "good" pec run with dud eq6,

This is after re grease, optimizing spacings on gears, and ensuring no backlash.

Was over 55 arc secs total when I started.

Tasco are obtaining a new worm for me, not sure how long before I have a usable mount though......

Gary

[Doug]

Phil it looks like there is about 13 arcsec error. I suppose I could say +/- 7, but since I started the run with the track at its lowest point, all variation was upward. That was plotted 'as is' The plot while PEC training seemed to be mostly about 4 arcsec with a few minor indiscretions. Then something went batty; the Dec trace looked like about 30 on the richter scale! I wonder if I pay the registration fee, will K3CCD behave? The star image stayed still, the white square box was jumping up/down all the time.
After things were restored to normal, I replayed the PEC training but the error was still about 13 arcsecs, So something I'm doing is wrong or else there is a problem with the Firmware in the Controller.
Guiding gave reasonable tracking, though there were a few sharp rises and falls, this can probably be refined with adjusting aggressiveness.

Cheers,
Doug

[Doug]

G'day Gary, You said Yes that is true, as I indicated earlier I have an easy guider on order; in the mean time If PEC was working for me I could manage a few good imaging sessions. I can always use an offset filter to fix egg shaped stars; PEC working would involve less post processing.

I think you mean lapped. Hobbing is only a special type of milling operation.

I saw that or the prototype about 18months ago. The trouble with that is flexure. Unless mounted on a decent dovetail plate system.~ $350 it'd be a waste of time and a Dovetail plate on an ED 80?

It might pay me to have a new worm and wheel precission ground and lapped if the existing ones are easy enough to duplicate. The PE errors are really quite abrupt; worse that your bad one, though not as extensive.

Since nobody is popping up to tell me I'm recording or playing back incorrectly, I guess the problem is with the 'Clayton's' PEC training sub routine.;)

cheers,
Doug

[garymck]

"I think you mean lapped. Hobbing is only a special type of milling operation."

What I mean is that the worm and worm gear are not matching profiles, this means that the angular contact area is small and exaggerates the small random errors. This is not uncommon My understanding is that gears are hobbed, worms are cut - I probably wasn't as precise as I should have been :-)

"It might pay me to have a new worm and wheel precision ground and lapped if the existing ones are easy enough to duplicate. The PE errors are really quite abrupt; worse that your bad one, though not as extensive."

I've already checked this out, having a worm precision ground costs about $150 on mates rates, and that is after having one made with a matching new worm gear !!!! You might want to check out a set of Gerlinger gears from Germany - these will cost about $500 to get here. Have not seen any real reports of the results of fitting, but they were popular in Europe. They have matching profiles and are supposedly much smoother than the originals, but require VERY precise set up.

You might want to check and adjust the mesh and alignment of the gears - I was able to reduce the pe from about 55 arc secs to that shown, and at the same time the small jigglies shrank as well.

I also am not sure that the pec routine works properly as I have had strange results with it as well....

cheers

Gary