I've found this thread on CN that might be of an interest.

I have the same issue. How did you solve the problem? Thanks.

You have to take the DEC cover off, straighten the belt and then tighten it so it has no more than 3 mm slack when you press it down in the middle. Having the proper tension and it running straight is key.
Nonetheless, this remains finicky. For instance, I can now get outstanding DEC guiding with an RMS of 0.5" early in the evening, only to see it fall apart and disintegrate to >1" after the meridian flip. That shows the mechanical limitations of the mount.
You will have to get a feel for how much tension the belt should have and whether the axle is running tight, but without excessive resistance, etc. A real problem IMO is that the DEC motor is drawing about 40% less power than the RA motor, which results in it losing steps whenever there is too much resistance.
Also, make sure that the mount is perfectly balance in DEC and that there is no cable drag. With those weak motor, you have to strive to eliminate any potential mechanical interference.
Hope that helps,
Jo

Thank you for your input. Did DEC belt tensioning eliminate the sawtooth pattern in the DEC guiding graph?

This is the DEC guiding graph I got.

For the most part. That sawtooth pattern is - I think - the result of mechanical tension that builds up as the mount fails to move in DEC until the tension breaks and is released, resulting in the overshoot in the other direction. When you adjust the tension, feel every moving part and determine whether anything snags, has resistance, etc.

Frankly, from looking at the simple construction, I am amazed that the mount can perform as well as it does (most of the time now) for me.

I think I got DEC pretty well under control now. Need to work on RA next:

Interesting thread - I'm trying to sort out my newish CEM25p dec behavior. Got the mount in September, but only added a guide scope in December. Never any issues with being able to move or wiggle either axis when they are locked down or with the tension knob backed out a reasonable amount - both RA and Dec are rock solid. I initially had very inconsistent guiding results - some nights quite acceptable (under 1' total RMS), other nights (dithering every 2 images) I would get good guiding for some dithers, terrible for others. Usually good/bad behavior was predictable from the calibration. Some times Dec calibration failed, sometimes the plot was wonky (much less total movement in the Dec direction than the RA), but every once in a while it looked 'normal'. I knew I had a backlash issue just from watching a star in the dslr live view. Speed 3 is the lowest setting where I can really tell what the mount is doing (at 10x zoom), and there was a noticeable (couple of seconds?) delay in reversing Dec direction.

Finally opened up the housing and the Dec belt was pretty slack. Tightened it - it seemed about 3mm of give. RA belt was already about like that, so I left it alone. At the same time I figured PHD2 out enough to run Guiding Assistant for the first time (so nothing to compare from before the first belt tweaking - my guess is it would have reported too much to compensate for). After the first tightening PHD2 reported 1450ms Dec backlash. It actually guided OK, but having read a CN thread where people had gotten theirs down to 3-400ms I got greedy and went back in and tightened the belt more. With very little play in the belt now, GA reported 2500ms backlash.

So - sounds like it is possible to have the belt too tight, correct?

Second question: Do we know exactly how the locking knob interacts in terms of backlash? I first backed it out enough for it to not stall at speed 9 (360 deg). At that point, it would still "hesitate" some when watching a star at speed 3, so I backed it out another 90 degrees and that was sorted. But I suspect speed 3 is still a lot faster than what the guiding pulses use, correct? So I should back it out a bit more. But is it best to just go "plenty further out" however that is defined ? At what point does there become too little pressure to hold the worm to the gear?

Thanks for any pointers - I feel that I am close to getting good consistently guiding - I hope.

Looks like you have already gotten a good grasp of the parameters. Fine tuning them is a matter of trial and error. Also make sure you are well balanced in DEC and that you have no cable drag.
The main drawback of the DEC stepper motor is its current limit, which is about 2/3 of the RA motor (for my CEM25P). That means that the motor is prone to losing steps if there is ANY excessive drag.
As you found out, overtightening the tension screw can induce such drag.
As for sufficient pressure on the worm gear: I now wiggle the mount in DEC while tightening the tension screw. Once I can no longer feel a discernible movement, that's when I tighten another half turn, then stop. I found that to work best in my hands. Depending on where the telescope is pointing, i.e. how much potential torque there is on DEC, I can now get the DEC RMS down to <0.5 on a good day.

All that being said, I think we have to be realistic and accept that a low end mount like the CEM25P is not optimal for telescopes with a focal length of >500 mm. I am using it with a WhiteCat (250mm focal length) and for that it is perfect.

Jo

PS: Yes, I think you can overtighten the belt. Also, make sure the belt is aligned with the gears and does not run in an oblique path over the gears.

Finally got clear sky for an evening. Lots of moon, so I don't even feel bad about having to spend it testing guiding. I started pointed east, within a half-hour of the meridian near the zenith. Ran GA twice and got 2490ms and 2775ms backlash. Not great, but it can probably guide OK with that.
But then I moved to the west, pointing at M45. Ran GA and got >21000ms. Tried various settings for the tension knob, but always >19000ms. So it looks like adjusting the dec mesh is my weekend project.

I think I have a decent idea what is supposed to happen. I will check the hinge tightness. But in terms of the mesh, it is basically: loosen the 3 hex screws along the hinge, lower the worm onto the ring gear. It should seat itself and in doing so move the hinge into the correct position. Then you just tighten the 3 hex screws to hold it there. I understand the part about moving the gears to find where it still binds and then re-doing the mesh there. Guess I'm just looking for confirmation that my view of how "adjusting the mesh" works is correct.

Did you watch the Paul Chasse ZEQ25 videos? I'm not sure what it is about putting a spacer between the tension screw and the hinge during the re-mesh accomplishes. My best guess is that it evens out the pressure along the hinge a bit. Did you do anything like that?

Thanks for this report! I have seen the same thing repeatedly, guiding accuracy on the same target shifting dramatically after the meridian flip. It can only be mechanical. I have not done much imaging over the last two months for personal reasons, but on the two occasions I did I thought I had that problem pretty much gotten down to a minimum.

I did not know about the ZEQ25 videos. Can you post the link?

Best, Jo

Sure: ZEQ25 Tune Part 1 That should link you directly to parts 2 and 3, let me know if it doesn't.

I don't think Paul is currently active on CN (astronewb), or I'd ask him why he says the metal ruler is the secret weapon in getting the perfect mesh. I guess the same principles apply to the CEM25P and ZEQ25. The upgrade kit included a new motor and a new worm gear, but not anything different with the hinge and worm housing.