So, I recently performed a hypertune and beltmod on my HEQ5-Pro telescope mount. Here's my report on the Hypertune and the Beltmod.
Hypertune:
I have a 2021-ish HEQ5-Pro mount acquired used from a CN user some time ago. Before my experiments I was getting ~0.7 RMS on good nights, and ~0.95 on bad nights with poor seeing and wind. During times of absolute best of everything (seeing, alignment, location) I could get long streaks at ~0.55 RMS but still long term averages not there. My DEC backlash was floating around 2600. I'm using a 6" F4 TPO Newtonian with a Pentax K-5 camera and a 30mm F4 mini guidescope and a 224MC guide camera and a miniPC on the top rings. From what I understand, this RMS was about as good as it gets for a gear driven HEQ5-Pro. Of course the mount is not broke so I want to fix it. My goal in the indefinite future is to upgrade in time to a 8" F4 Newtonian, so I want to push this mount as far as I can at 6" so I feel comfortable about bumping up to 8".
I reviewed almost all the major YouTube videos and CN posts that I could find for the hypertune and found them all to be fairly accurate. Here's the bearing parts that I used:
- https://www.amazon.com/dp/B082385W52
- https://www.ebay.com/itm/334273566061
- https://www.ebay.com/itm/115362639421
I decided on ceramic/metal bearings for the worm gear because we're running at low speeds and also we're applying axial loads on the bearing outer ring. The other bearings purchased were what were recommended or the highest quality variant available.
In taking apart the mount, the worm gear bearings were in bad shape as noted in various tutorials and very much needed to be replaced. The worm gear and corresponding brass gear were filthy and had lots of shavings from use. These were cleaned to perfection using alcohol and a fine brush until the last round of cleaning left the alcohol clean and clear. I used Super Lube Silicone with PTFE to reasonably re-grease everything as suggested in tutorials.
My experience was similar to other commenters in that my impression was that besides the worm gear bearings, the rest of the bearings were just fine and really didn't need to be replaced. If I went back to do it again I would just clean/replace/readjust the worm gears for DEC and RA and leave the other bearings alone. In any case, the bearing replacement went fine. Adjusting the DEC/RA backlash was somewhat challenging. In sum, the worm gear bearings very much needed replacement. The other bearings seemed fine.
Beltmod:
I bought the cheap Chinese belt mod off Aliexpress. It was USD$106.46 after tax and shipping. The Aliexpress website is an abomination and has the most awful search engine ever. The cheapest beltmods are found with searching the phrase "Riem Mod Kit heq5" and still lists for about $81, pricing varies (a lot) by seller. The lowest price listing is in Dutch and hard to find in the search tools. No extra large spacer is needed between the mount housing and mount cover, which is a plus compared to the Rowan UK version for sale.
I have mixed feelings about the Chinese version, and so I'll review it part by part:
- Gear Removal Tool - fine. They sent me 2 gear removal tools. They are 3d printed plastic and metal. Mine worked fine. I used the same tool 2x and had no issues. I had a backup if needed. The brass gears are interference fit onto the shaft of the motors and you really need a tool to get them off safely. Some CN posts have expressed reservations about plastic printed parts. I had no issues.
- Main Large Gears - fine. They appear to be cast or 3-d sintered aluminum. I feel like I would want them to be better machined post-casting, but I have no insight into the tolerances/design or if this concern is valid or not.
- Rollers - fine. I ended up switching the rollers from the recommended sides to their other respective sides because I was having severe issues getting the RA roller to cooperate with the motor position. The belt would contact itself on a direction switch as the tension direction changed. I swapped rollers and everything worked much better on both DEC and RA.
- Belts - very bad. They started deforming and shedding metal wire after the first few hours of use. I purchased new ones from the UK. Other users have had similar problems with UK belts, so this might be user error in overtightening. The motor gears may also bear some responsibility as noted below. Nevertheless, I'm disappointed in the belts.
- Gears for Motor - sub-par. The original brass gears are an incredibly tight interference fit. But these new motor gears have 8 set screws to use to secure to the motor shaft. This means that it's very difficult to perfectly center the gear on the shaft. Tightening the set screws drives the gear off center and out of alignment with the axis of rotation. And, as I found out, the set screws can loosen during operation and so they need to be fairly tight. It took a few hours to figure out how to center the DEC gear. And despite my efforts my RA axis still has some slight visible wobble as it spins. (Also, I can only determine the wobble from the rotation of the top cap, which might not be perfectly centered or round, such that the gears may actually be centered on axis and the top wobble is misleading.) Anyway, this wobble may have contributed to my belt problems by slightly deforming the original Chinese belt over time, particularly if I had overtightened it. This centering and securing issue was the single hardest aspect of the belt mod. Sadly, the UK gears visually appear to have the same issue, but I've not seen complaints on CN about axially aligning the UK version. The belt motor gears really need to have a tight interference fit with one or two set screws at the bottom of the gear. But nobody does this.
The beltmod process is fairly straight forward and the various Youtube tutorials are on point. Adjusting the belt tension, DEC and RA backlash, and adjusting the worm gear locking rings was non-trivial. Stuff comes loose very quickly and self-adjusts over time. It took many hours in the dark to get things sorted out.
TWO CAVEATS:
First - I made a huge mistake and tried to disconnect the motors from the control board. DO NOT DO THIS. The solder joints on the control board are very, very weak and the connectors into the control board are very, very tight. I destroyed my control board trying to take off the motor connector and it was broken before I even realized it and i'm not a novice to connectors. It's not really necessary to take off the motors from the board. In retrospect if you do try and fail, then take the board to a cellphone repair shop and they can resolder the connector for less than the cost of a new board.
The point in taking off the motors from the board was to make it easier to center the motor gears. I should have left the mount disassembled during the hypertune and taken the board and all the guts out at once instead and partially installed the belt mod at that time. This would have been much cheaper.
Second - The screw holes on the motor for connecting the motor to the mounting bracket easily(!) strip. I stripped two motor holes trying to tighten them. It turns out the metal motors are made of soft iron that strips so very easily. The mounting bracket to mount holes seem much stronger. Fortunately the remaining two holes seems to be enough to keep the motor on the bracket securely. Also you can rotate the motor as needed relative to the bracket.
Big Picture Results:
Even with a deformed belt shedding metal wire, using the Chinese belts my DEC RMS dropped by at least 40%. This was a massive improvement on the DEC axis. This may be attributable in large part to replacing the worm bearings. With the Chinese belts, the RA axis was giving me issues, and was performing worse than it did when gear driven. Also, the deformities in the Chinese belt appeared to cause huge RMS spikes. Nevertheless, I was getting long 0.6 RMS bursts the first two nights and 1300 or so DEC backlash, and this with deformed belts and incomplete adjustments. During the times when it was running well (even with the deformed belts) it was performing, net, about 10% better total RMS than it was before. However, many frames had to be discarded due to large guiding errors.
I ended up buying UK belts, and after I replaced the Chinese belts with UK belts, everything was much better. Using PHD2 defaults I was getting good gear numbers the first night. By adjusting PHD2 and using the PEC correction algorithm I was able to get 8+ hours of imaging across two nights at ~0.63 total average RMS. If this was gear driven I would expect around ~0.75, so we're seeing a substantial improvement. I expect that with some more experimentation I might be able to get to ~0.55 averages for the entire night. Things that helped were:
- PEC algorithm for RA
- Reducing MnMo very aggressively
- Setting the exposure duration to 0.2 or 0.5
In sum, if I were to do the belt mod again:
- I would replace only the worm gear bearings, clean everything, and regrease. This saves like $75.00.
- I would still get the Chinese belt mod but I would also get extra belts from the UK at the same time. I'm not convinced that I would avoid the axial centering issue with the Rowan beltmod version. Belt failure seems to be a common issue across all versions of the mod.
- I would take the board and motors out as a unit during the hypertune and get the motor gears perfectly aligned and tested without disconnecting them from the main board.
- I would be more careful with the motor to bracket screw torqueing so as to not strip the holes.
Pricing:
- Belt Mod: $106.46
- All New Bearings: $111.30
- New Control Board: $199
- 2 UK Belts: $36
Total: $452.30
And if I were REALLY to try again, I would just sell the HEQ5-Pro and get one of the new $1K or less strain wave mounts from Aliexpress (e.g., Juwei 17 or 14) that have same or better performance, greater payload, less weight, and directly compatible with NINA without an EQMOD cord. Awkwardly, doing so would have been about the same net cost (strainwave mount + tripod + lifter) and resulted in (I suspect, but cannot prove) equivalent or slightly better performance than a beltmod. Strainwave is clearly the future.
