Regular massive spikes in guiding

The EQ5 has never been known for its tracking performance…

andrea tasselli:
The EQ5 has never been known for its tracking performance...

I got 0.5" out of the box but maybe I just got lucky?

You said regular – do some math, see if it really is regular, like a worm period multiple.  Also compare to a log of imaging, does it follow any event like a filter change (especially if using an OAG and using filter offsets without a stop/start guide).   Other possibilities include wildlife - a bird (e.g. owl) stopping by.  I've literally had a tree frog jump onto, and then off the OTA and you get a blip.  Cable snags and such have been mentioned, I had one of those "armored" USB cables (plastic rippled wrapper) and found it left a little vibration every time one of those ripples was pulled over a spot on the mount.  

Those are pretty big and always in DEC (right?).   DEC sudden excursions won't be caused by tracking errors, it generally requires something intervene and move the axis.  I guess it could be the mount responding badly to a small DEC move and making a giant one, but I'd look first for external causes.

I had my rig do similar at the beginning of the last session  in windy conditions.

After scratching my head while it did that few subs, I shut everything down and restarted and it recurred.

Then I doubled the guider exposure tine, restarted and all went normally the rest of the night.

I had a suspicion the cause is an oddity of some sort in the ASIAir guiding algorithm in windy conditions where it just goes berserk and slews in Dec, loses the guide star (I got the STAR LOST warning each time), then decides to go back to where it was and recovers  the guide star.

What’s interesting is that it’s a deliberate slew because the movement is larger and faster than is possible if it was using the guide pulses - it’s set to 0.25X sidereal and the pulse duration 200 milliseconds.

The ASIAir has current firmware. Mount is a CQ350.

Francesco Meschia:
That's the declination trace, it won't be anything related to the worm period simply because the dec motor doesn't have one – it isn't always running. Cable snags seem the most likely culprit.

Yeah, I think I said that before really paying attention to the color. Unless it's a MyT, with it's cross talk.    (Glad mine is gone). 

I do find it interesting there isn't much RA impact at the same time.  That's kind of a wimpy mount for a cable tug to have that much DEC impact without affecting RA at all.  It makes me wonder if something is literally sliding/moving as the OTA's angle changes. Notice that's over 7 hours of time.  If it's something with freedom to move (very slightly) only in DEC. 

Was the DEC axis very well balanced?  A related possibility is if as the angle changed the effective tension on the DEC axis overall changed and there's a tiny bit of give somewhere that occasionally slips?    or related, if there's a big temperature change over that time, dissimilar metals in (say) a dovetail plate and long OTA might result in periodic slipping in rings as one grows faster than the other (ok, that's grasping a bit at straws, but when dealing in arc seconds everything is made of rubber -- that's the saying right).

I would definitively check the belt on Dec.
Had a similar effect on RA and it turned out that the belt was too tight, deformed and slipped over the edge of the gears. Each time a tooth of the belt slipped over the edge, it kicked the axis. That explains also that some of these events coincide with dithering events, because specially at that time the Dec axis has to work.
As Christian mentions, if the teeth do not fit the gear very well and the belt is tightened too much, this can also produce this effect. Reduce the tension of the belt in that case. The teeth of the belt might slide more softly over the teeth of the gear. Or get a belt which fits better.

This is a close up look of what a typical spike looks like. The dec axis quickly moves off, the mount tries to connect but can't. I'm wondering if it's backlash

Backlash is usually shown by repeated pulses being unable to correct an excursion, so I get what you are saying, but the problem with that theory is that the excursion itself seemed so sudden and large.  I don't see how backlash per se could be the cause, though having backlash comp disabled may be the reason it took so long to recover (for a belt drive without encoders I expect you have substantial backlash innate in the system). 

It's certainly worth turning on compensation (unless it's off for a reason?) and running guide assistant to give it a starting point.

But I still think it likely there's something loose in the direction of the Y axis that periodically slips (or something akin to that).  Keep in mind 10" is probably so small if you were watching carefully when it happened  you couldn't see it by eye.  It's large as an error, but small as a physical motion.  The fact that the RA axis keeps on tracking well (if that's true in all these cases) seems to exclude some external impulse (wind, something bumping the mount, etc.).  I would still think it's something in the drive itself (where that last pulse instead of a slight motion did a very large one), or something loose where when you hit a certain balance or weight distribution it just slips or flexes a bit.

 I just read @Franchesco's note.  I'd agree (e.g. where I said "instead of a slight motion did a very large one"), except that it wasn't really trying hard to move in DEC before hand.  It may be worth looking carefully at each one, see what the last few intervals of DEC looked like before the spike.

Definitely the belt. I have a belt modded eq5 and get 0.5 to 0.7 no spikes.

Easy fix. Take of the side plate and touch the belt with your finger. If it's too loose it will flex more than a pencil width. If too tight it won't budge by more than 2 to 4 mm.

Adjust accordingly 

Happy guiding

…. except that you would think you have errors in both RA and DEC.   It was just a thought.  It could be that the belts, etc in above discussion are also suspect and it shows more on DEC due to more issues on that axis.

MaksPower:
If it has had the belt mod that means someone opened it up.

I would suspect one of the RA bearings has grit in it, and every time it rolls over that it pushes the axis up/down which will appear as a fast spike in dec.

I've mentioned this before - at the typical dimensions of the parts in these mounts an arc second corresponds to about 1 micron. YOU CANNOT SEE particles that small. Even if a part looks clean naked eye I would guarantee it isn't when examined under a loupe or a microscope.

If the mount has been opened by someone in surroundings that are less than scrupulously clean it is virtually certainly the worms and bearings are contaminated.

You could attempt to strip it down, degrease and clean, but by the time you reassemble the chances are it will be worse not better - unless you have a cleanroom environment similar to a semiconductor lab.

This is a major reason I will never buy a mount secondhand. Especially one that has been futzed with.

Actually, the latest version of the EQ5-Pro comes with belts as standard. My EQ6R-pro did as well.  I'm not sure how long that has been the case though.

I wouldn't discard it before attempting to fix it. As you say it's been used and it clearly needs tuning, which eventually they all do. When I started out in this hobby I did the same thing, bought an old HEQ5 PRO which I modded and tuned myself. It seemed like a lot to take on at the time, but it's not rocket science (thankfully). Adjusting backlash can be a bit cumbersome, but otherwise I found it a neat/teaching experience.