Poor DEC backlash after belt mod (old EQ6)

[rnobleeddy]

Had a bit of a nightmare belt mod with my EQ6 but after a lot of work, got it back together and there was just about enough clear sky to see how it guided. The answer is worse than before!

Just wondering if anyone has any additional ideas for this large backlash in DEC (which amounted to 12s!) :

The easy things to try are:

- make sure the worm shaft is not too tight (or too loose, but it isn't!)

- ensure the pulley tension is sufficient

After that, and most likely, I guess its the worm gear itself meshing with the shaft? There is no free play by hand - can this be true and lead to such bad backlash?

The RA gear was worn and had some minor damage on a few teeth, so I swapped RA/DEC over but I didn't do too much to file the burrs/rough bits on what is now the DEC cylindrical gear. It's not as tight as I'd like but I just got to the point where there was no free play by hand and it didn't bind on a full rotation. It would go a lot tighter and not bind most of the way round, but would bind in one spot.

Does anyone know if the binding tends to be caused by small defects on the gear itself (which I can file) or manufacturing tolerances relating to roundness etc of the whole part? If it's the former, I can take it apart and work on filing it. If the latter, I guess I'd do better swapping the RA/DEC cylindrical worm gears back?

In general, if one gear has some damage and a replacement is not easy to find, is it better on the DEC or RA axis?

Edited November 26, 2021 by rnobleeddy

[rnobleeddy]

To reply to myself, this pretty much came down to belt tension.  From my experiments there was a (relatively) forgiving range of worm end tightness and worm meshing in which good results were achieved, but insufficient belt tension was a killer.

I'm unconvinced I'm any better off than before the belt mod - DEC backlash still measured between 0.8 and 1.9s but that's within the range that PHD2 appears to be able to mop up with backlash compensation.  I didn't measure this before though! 

[wimvb]

Just read your post. You should be able to measure if the belt tension is ok. With my AZ-EQ6 (belt drive), I consistently had a 10 s period in my guiding (you need 1 s guiding exposures to see this). I analysed the guide graph and a frequency analysis showed a pronounced peak at 10 s. It turned out to be caused by the belt being to loose. Tightening the belt improved the guiding. The next hurdle was adjusting backlash. I spent two evenings on this. The trick is to take it in small steps (1/32 of a turn of the grub screw), and use the guiding assistant to measure backlash after each adjustment. Also check for binding after each adjustment.