EQ8/Guiding Sawtooth glitch oddity - any ideas?

[Sp@ce_d]

Well... my last attempt to post this seems to have been devoured in the ether!  Are new posts auto saved server side? My browsers certainly lost it!

Anyway. Been having an occasional glitch while guiding I cant put my finger on. There's a tell tale sawtooth in the graph I've seen 4 or 5 times since I put the new set up together. First I thought it might be a cable snag or something bumping the pier/mount. However, haven't found any evidence yet as I tweak & adjust the set up.

Noticed it again last night during a 3 or 4 hour session.

Any ideas anyone?

Snapshots with EqMod PEC & Guide pulses shown, Maxim Guide Graph. Also the resulting subs from the two scopes imaging at the time. Movement isn't in the direction of tracking errors I'm still tweaking. As any tracking errors run bottom left to top right.

[ollypenrice]

Extra questions;

- which axis is x, is it RA or Dec?

- how is x axis aligned on the images? You seem to be saying diagonally but I didn't fully follow your last paragraph.

Dual rigs can run into some tricky geometry as seems to be happening here since the errors in both scopes, though similar, are not identical. The top one has trails top and bottom of the star while the other one trails only on the bottoms of the stars.

- What's the guider, OAG or scope? If it's an OAG was it working through the top image's scope or the bottom? If it's a scope, is it closely aligned with the imaging scopes?

I'd check for differential flexure between the scopes. How are they mounted?

I'd also align the guide scope carefully with the imaging scopes on a dual rig. I've never bothered on single rigs.

Do you get any clean subs or are they all trailed?

Olly

[Sp@ce_d]

Extra questions;

- which axis is x, is it RA or Dec?

- how is x axis aligned on the images? You seem to be saying diagonally but I didn't fully follow your last paragraph.

Dual rigs can run into some tricky geometry as seems to be happening here since the errors in both scopes, though similar, are not identical. The top one has trails top and bottom of the star while the other one trails only on the bottoms of the stars.

- What's the guider, OAG or scope? If it's an OAG was it working through the top image's scope or the bottom? If it's a scope, is it closely aligned with the imaging scopes?

I'd check for differential flexure between the scopes. How are they mounted?

I'd also align the guide scope carefully with the imaging scopes on a dual rig. I've never bothered on single rigs.

Do you get any clean subs or are they all trailed?

Olly

Hi Olly...

X is RA

The camera's are all aligned the same as in pic below.. (As you know, Atik labels indicate sensor orientation)

The guider is currently a lodestar on a finder. I do notice the scope with it attached has slightly better stars. I've tried OAG's in the past but I only got perfect stars with the scope it was on. I got better results generally across all scopes with the finder/guider  (I'm currently testing out using the ZS71 as a guider as its in middle, so equally spaced from each Esprit.).

They're mounted on an ADM Losmandy triple bar. One is directly fixed the other uses an ADM Max guider Saddle (yes I know.. if I could get hold of a couple of Cassidys ). The ZS71 is on a SW guider saddle I modified with ADM parts. I cant detect any flex not to say there isn't any of course. Yes the cabling's a mess, it's in the process of being tidied into looms & I wondered if one was snagging but I haven't found any evidence of this.

Those 2 subs are the only ones spoilt during nearly 4 hours on a couple of targets last night (this morning). Normally, apart from slight eggy stars, more so on one scope than the other at certain parts of the sky, the subs are clean. It's only this very odd "saw tooth" glitch I've noticed randomly happening that spoils a sub run. The next sub after is fine. Very odd.

I guess as I've only noticed this since installing the EQ8 I'm keen to eliminate any possible problem with the new mount.

[orion69]

 I've tried OAG's in the past but I only got perfect stars with the scope it was on. 

Sorry couldn't resist...

[Sp@ce_d]

Sorry couldn't resist...

yes I know..

OAG's are great on a single scope setup (if there's a star in FOV). As Olly points out, guiding & the geometry of a multiscope layout has to be considered. I found that having the guider closer to one scope than the other(s) results in slight trailing on longer exposures on the further scope(s).

I had a play with guiding last night. Could just about see Deneb through the mist & cloud. Using the ZS71 as guider which is central to the rig, I got a round star on both Esprits on 900s subs over a couple of hours. No X Axis glitches either, although I don't think that proves anything in this case.

Sent from my iPhone using Tapatalk

[ollypenrice]

Well, I'm bound to be suspicious of the Maxguider HD but maybe you have two issues here.

The runaway sawtooth is weird. Quite suddenly you get a lurch in RA which is then nicely (if slowly) restored by the guider. So what causes the sudden movement in the first place?

Power surge? Cable snag? A rogue guide command sending it awry in the first place? (Think USB.)  Backlash? (Have you checked the backlash throughout the rotation of the wheel?) Sudden movement in the gudescope? (Don't just look for snags on the imaging scopes.)

One thing's just struck me though. The trailing effect on the image you posted seems out of proportion to a single 'saw tooth event.' Because the correction seems to have been made fairly effectively the scopes were not off target for long. I'm surprised the trailing is as obvious.

Olly

[Zakalwe]

Upgrade the mount firmware to the latest version. Some of the firmwares had a glitch that caused a "spike" when guiding with Lodestars.

[Sp@ce_d]

Upgrade the mount firmware to the latest version. Some of the firmwares had a glitch that caused a "spike" when guiding with Lodestars.

Ah.. ok...  hadn't realised there's new FW. I was using an Atik 314L+ on the guidescope but I've just switched to a Lodestar on a finder.

In fact since my last post things have gone way downhill. Guiding started to get worse generally & now just wanders off in one or both directions. I've taken everything off the mount & started again. just put an Esprit with finder/guider on it to try & work out whats going on. I can only "guide" directly above now if the mounts more or less horizontal. Once I start pointing any side the closer to the horizon the faster either RA or DEC or both drift off. Almost as if the clutches are slowly slipping or the pulses are being ignored by the mount. I'm going to try it with an ST4 cable & check using another EQMod cable from my heq5, as I'm using an FTDI one at the moment.

[ollypenrice]

Beware the Lodestar ST4 cable. It is a dreadful thing, far too flimsy and redesigned on the new SX guide camera. I have very often had bothers with both of mine. You'll see all sorts of solutions being asserted, like the plastic cable guide form TS or a home made equivalent, or tying the cables together, or taping them to the body. I've tried all of these and they don't really work. They do seem to work but what is really happening, in my view, is that making a change to the cables is improving a contact. So I've had just as much success from removing a modification as I've had from introducing it! I concluded that basically the cable needs an occasional wiggle.

This might explain the inconsistent behaviour of the mount when its position is altered. Good connection here, bad connection there.

Zakalwe's post does look promising on the glitch.

Olly

[Zakalwe]

I think that version 2.09 fixed the random spikes that happened when guiding with a lodestar. Best have a look on Skywatcher's site (or the Yahoo group) and download the latest. Some of the recent firmware made significant changes- for example improving the polar-alignment routing.

If you do upgrade, then hook the USB-to-Serial up to a powered USB hub. Updates kept failing on my (bricking the handset! Luckily it's easy to restore it) when the USB-to-Serial was plugged into the PC. Plugging it into a powered USB hub solved the problem immediately.

[Sp@ce_d]

Ok, looks like there's the handset update V4.36.02 and motor update v2.11

I take it I'm only interested in the motor FW as far as my issues are concerned, I use EQMod. I don't use the handset?  At this point I haven't investigated the specifics of upgrading, maybe I'll need to do both anyway.. if not then yes, I'll update the handset at some point but I've only ever plugged it in once to test the mount & check the PA routine. I'm under the impression there's nothing in the handset updates that are loaded into the mount?

The other thing I guess would be interesting to know is how others have EQMod set up. Mine is pretty much set the same as I do for my EQ6 and EQ5.

[Sp@ce_d]

Something else that has just occurred to me. I use a side by side set-up, so the puck is at 90 degrees. I wonder if this has any bearing here. I still have a triple bar on the mount, even though I've gone back to a single scope on the centre saddle to fault find. So the puck angle hasn't changed.  Hmmm

[Sp@ce_d]

Finally got a chance to have another (brief) go at this. So, my Motor FW was already at 02.10.04 but I've now updated it to 2.11 & the handset (I don't use) to 4.36.02.

Measured voltage to mount. 13.76V presented at mount connector, handset reports 14.6v.. Hmmm?  (I've run all my kit off one of these (preset to 13.8v) bench PSU's for some time)

http://www.maplin.co.uk/p/moonraker-375w-switched-mode-dc-variable-voltage-compact-bench-power-supply-n87jn

Was able to see enough stars through clouds for a while to (briefly) test guiding so:

Tested with new FW. No change.

Replaced EQDir (removed FTDI version) with my Hitec Astro one (prolific chipset) I use on the HEQ5. No change.

Plugged it into it's own dedicated powered hub. No change.

Wired up the ST4 cable to lodestar.. its never been used so a tight fit for now!... clouds rolled in.. had to stop there so, Not Tested Yet.

What I do seem to have now is a repeatable scenario. If I'm guiding pointing (scope) west, the RA will just carry on drifting off the graph from the start.

If I'm guiding in the opposite direction with the scope pointing east then DEC drifts off from the start.

I tried this 2 or 3 times on each side.

Now, then when pointing East, I was checking over the settings in Maxim and hit the set defaults on the advanced tab. I was already using the defaults by the look of it however, the guiding stayed on track for a while on the next run. I could see the pulses creating a proper graph (instead of a flat line with intermittent pulse) in the EQMod pulseguide monitor window. All the other times either RA or DEC is slammed hard up E or S. I have no idea if this is a clue or just coincidence!

It's almost as if its trying to correct in the wrong direction?

Just thinking about this while I'm typing. I must check the EQMod ini file. I wonder if there's possibly a funny with pier side reporting or something?

This is starting to drive me nuts, it was guiding ok apart from that weird spike. Maybe I'm looking too closely now & I've missed something basic..

[ollypenrice]

Regarding the orientation of your saddle plate, I'm sure I remember something about this. I think there's a software setting you can alter, though I can't remember the details. It's just possible that this has something to do with it.

Could you try guiding in completely different software? I use AstroArt. It's messy in terms of the number of windows you have open but once you have sussed out the parameters they can all be altered quickly and logically. You can swap RA and Dec round, make either guide the opposite way, disable one direction in Dec etc very easily and while guiding.

My brief experiment with PHD2 was a total failure with the mount being briskly driven in the wrong direction. I didn't look into this but just went back to AA. I found PHD1 worked fine when I used it, but that was on the Mesu which needs hardly any guiding at the best of times and never misbehaves (touch wood!)

Olly

[Sp@ce_d]

Thanks Olly

Yes the saddle orientation needs checking out. I'm going to trawl through the Yahoo group & pop a question out there I think.

I did try to run PHD2 but I forgot I had it set up with the old config & it grabbed the Atik instead of Lodestar & locked things up. Clouds were in after I sorted it. I have tried it before with SGPro but it never guided nearly as well as Maxim. I'll try it again. I'm also going to try V6 of Maxim as I'm using v5.24. Just in case there's an oddity there. I've run V6 before but its not as stable as 5 at the moment.

[Sp@ce_d]

Hmm.. looks like I've sorted it but I'm not sure what to make of this

I tried the ST4 cable.. it was worse & the errors were in the opposite axis to the pulse guiding!

So.. I pulled off the triple bar & re homed the saddle & stuck the Esprit back in normal orientation. Ran some 3 star alignment test with the handset & hooked it all up again ( I hadn't disconnected any cables apart from plugging the handset in. Guess what.. its been guiding away through the clouds all night. Tested in Maxim v5 & v6, I haven't touched any of the settings apart from switching between relays & telescope/ascom/pulse. Ran a  few calibrations & its now running using ACP's SmartGuiding. Heres a 1000s sub of Capella. It wont win any prizes but it's a sight for my sore eyes!!

[ollypenrice]

Great.

So it looks like the saddle orientation was the bug. If you want to go back to the dual rig then you can either find (and risk!!) the software parameter change or mount the dual bar to a standard standard dovetail attached at right angles. Most people would do the former but I confess that I'd do the latter! (I'm more competent with a pillar drill than with software...  )

Olly

[Zakalwe]

The original firmware didn't like side-by-sides if you were using the handset- it mucked up the auto-home feature. Some people removed the saddle and turned it 90 degrees.

I did experiment with s-b-s and it was easy to use with EQMOD by re-synching the encoders. The latest firmware now has a function to use a s-b-s arrangement.

[ollypenrice]

Ah yes, turning the saddle seems to be the obvious thing.

Olly

[Sp@ce_d]

I don't b***y believe it!.... (said in the voice of Victor Meldrew) ........ I'm on shorter subs than the tests on Sunday night through the clouds!!  Stars are like eggs again & the glitch Is back.. on DEC now. Only thing left to swap out is mount, laptop or me!

[ollypenrice]

OK, one test I don't think we've explored; how does the mount do if left unguided? I'm wondering if the PE is consistent and glitch-free or whether it gives any sudden movements as it does when guiding. Gina posted some impressively long unguided subs which is what made me keen to try an EQ8 myself. If you don't get any glitches unguided then it's the guiding that's generating them, which would be useful information to have.

Is it possible to ask PHD to produce a trace without giving any guide corrections? This can be done in AstroArt with which I'm more familiar.

Olly

[Zakalwe]

Is it possible to ask PHD to produce a trace without giving any guide corrections? This can be done in AstroArt with which I'm more familiar.

Good idea.

Disable the guiding output in the Brain in PHD. Then use PecPREP to analyse the PHD log.

http://eq-mod.sourceforge.net/docs/PECPrep.pdf

[tony210]

Hi - if this  is in RA and you are using EQMOD have you tried raising the steps in EQMOD to 0.5 - I have used an increased rate in DEC at 0.5 with great effect but on some 30 minute subs have questioned slight trailing in RA on my EQ8 - worth a try I think - best wishes Tony

[Gina]

I have a multi-scope rig on my EQ8 and get occasional guiding failure. I haven't had many hours of clear night sky in the last couple of months to do much testing, unfortunately, and when guiding has worked I've taken advantage and done an imaging run.

One thing I consider essential for a multi-scope rig with guiding on one scope only, is to have a really solid mechanical connection between all the scopes.  I'm using the MN190 with OAG and Lodestar for guiding and mainly PHD v1.14.  I have an Esprit 80ED Pro mounted on top of the MN190 scope rings using an aluminium bar 50mm x 10mm of appropriate length.  The rings attach to a 15" ADM dovetail mounted on the EQ8.  Between the rings I have a 450mm x 250mm x 10mm aluminium plate screwed to the dovetail with 6 M6 Allen head screws.  This takes a pair of ST80 scopes each with their own alignment plates of 200mm x 150mm x 4mm with M5 screw alignment adjusters.

Rather than clutter this thread with my own design I think I'll start a new thread for my rig.  Most of it is already covered in a couple of threads in the DIY Astronomer forum but I feel a new thread covering my final (well maybe) design would be better.

[Sp@ce_d]

Ok, thought I'd better provide an update as it's always unsatisfying reading an old thread that has petered out when looking for some answers to a similar problem.

Having said that, I can't exactly put my finger on what was causing this. However, this is what I've done.

Since I'd already been through reducing the setup to a single scope, swapping out the imaging kit (CCD's & Guider), all cables, checking for voltage drop etc. Even the network & switch were substituted during fault finding. I then swapped out the laptop & went back to my old notebook I used with the EQ6. I still had intermittent tracking spikes, not as frequent or as harsh but still there. At this point it really started to look like the mount was the only thing left but I had a nagging at the back of my mind that this looked like it was some kind of interference, some process that was cutting in or some anomaly/incompatibility with the chipset/USB drivers.  Both the laptop & notebook run on Win7 & use mobile chipsets. So I decided to try a totaly clean build on a desktop system and give in to Win8.1. Therefore, no residue from previous systems/setups and rather than copy my config files (Maxim, EQMod etc) across, as I did with the laptop, I setup everything from scratch.

I'm a bit scared to say this in case I jinx it but.... So far I've collected around 20 hrs of subs, ok maybe not great subs, the seeing hasn't been that good but they prove to have round stars & the guiding works! I've even got ACP more or less running the Obsy now too so the guiding is working automated.

Which brings me to one observation I made that may or may not have a bearing on what was going wrong before. When I was setting up ACP I noticed that it wasn't showing the correct pier side. Although I had configured EQMod to report it so that ACP shows the direction the scope is looking. EQMod was reporting the correct position but no matter which of the 3 settings EQMod was set to report (physical, pointing or dont care), it made no difference to ACP. It seemed to be stuck thinking it was on the other side of the pier. Now of course, trying to guide with the normal settings of reversing the axis on pier flip caused a similar run away to what I was seeing before. Eventually, flipping through the settings in EQMod several times I got it to pick this up and its been ok since. I just wonder if there was something going on before that was causing EQMod or ASCOM to loose sync & flip that setting when reporting to the host program... weird.