Hi all,
This is my first post here. I'm not an English speaker, so I apologize in advance if there are any language errors in the post.
There's little practical information online about improving the quality of the Chinese-made EQ8-R. Or Orion HD110 - popular in US, essentialy the same mount. There's no before-and-after comparison anywhere.
(Of course, I'm not sponsored by any manufacturer; I paid for everything myself, and I'm only writing this because someone might find some useful information. I'll also post it on Cloudy Nights, maybe my writing will help someone)
Article Summary:
After installing the Shibumi Astro Engineering kits, the EQ8-R mount achieved a total RMS of 0.17px/0.42" (RMS RA 0.13px/0.32" RMS DEC 0.11px/0.26"). This was on the first day after installation, without any guiding adjustments. The best result achieved in the next session was a total RMS of 0.12px – but on Skyguide (not PHD2).
The cost-effectiveness is certainly debatable, as the combined value of a brand new EQ8 and the Shibumi kits is similar to the cost of a used, higher-quality mount. However, once have one, mastered the installation, you don't want to switch to another system, secondly, reselling the old one is usually not profitable and you won't get as much money back as you'd like, and thirdly, these better mounts rarely appear on the my market (Besides, buying expensive second-hand stuff is always a risk.…)
But you can buy an old Orion HD110 (or used SW EQ8) for a few dollars and DIY or send it to Shibumi for modification. The result is a premium-like mount at a fraction of the cost. If you do it yourself, it's even cheaper. And what a satisfaction to have done it yourself!
I'd rate the difficulty level as follows: the backlash adjusting kits – medium, (easy for mechanically inclined people); anyone can do them. Replacing the axles requires more, so it's difficult, (but for some mechanically inclined people, it's medium at best).
One more thing I must say – I've never seen customer service like Tony Owens's from Shibumi. We exchanged dozens of emails, and Tony patiently explained everything to me, clarified it, and offered suggestions. I usually got a response within 30 minutes. Even at 11 p.m. Tony – thank you very much.
History:
New (currently a year old), purchased and tested at local shop: teleskopy.pl. It was fitted with a C9.25 telescope / Starizona SCT corrector or Esprit 150 with flattener. Maximum setup weight: 18.5 kg. FL: 1450 or 1070 mm, respectively.
The mount worked with OAG or a 60/240 guider, with an average RMS of 0.51px/0.59", but only on the east side. After a meridian flip or around the zenith, it was bad, especially with the DEC. I thought I had a lot of play. I even adjusted it once according to instructions I found online, and it improved – but then the frosts came, and the play reappeared.
The idea:
The Internet suggested that Shibumi Astro Kits from Ireland manufactures precision components for the EQ8: complete worm drive modules, wheels, and axles, in both Heavy and Light Duty versions.
There are quite a few; those interested in upgrading their EQ8 can find everything on their website.
My idea was to order Anti Backlash kits for RA and DEC. Shibumi claimed the Light Duty kits alone would be enough, but after some thought, I added both axles. Made of steel. Mainly to avoid having to mess with the assembly again, increase its load capacity, and slightly improve accuracy.
Completion:
After about a week, I received the package from Ireland. Since it was winter and the workshop was cold, I did everything at home on a protected table.
📷 you can do it at home 2.jpg
I downloaded all the manuals; there's no point in describing the entire process here – it's all in the manuals. I'll just point out that almost everything was done with basic tools.
Shibumi suggests obtaining two specific torque wrenches, pointer-type (not click-type), 0-10 Nm and 0-0.5 Nm. The first could be purchased on a local auction site, while the 0-0.5 Nm one was only available on eBay.
One note: in all the mechanical systems I've worked on so far, you take the component, tighten it to a specific torque using a torque wrench, and that's it.
This is different!!! We compress the bearings with a nut turned with a pin wrench so that the shaft starts moving at 0.2-0.3 Nm (or other value specified by Shibumi for the given setup). This means we press down – and check the torque. If it's lower, we need to press a little harder. If the shaft torque is too high, we loosen the shaft nut slightly. A few iterations and it's done.
Since I didn't have that 0-0.5Nm wrench, I had to approach it differently:
📷 Torque checker.jpg
The solution above is a metal bar loaded with a weight suspended from a wire hook. By hanging it in the right place, you achieve a specific torque.
I first had to balance the bar by hanging it. (A nut tied with wire in the right place did the trick.) The weight is a bottle filled with water to the appropriate mass. In the example shown in the photo, the shaft started moving at 0.15 Nm. I don't provide the specific dimensions of the bar and weight; anyone can calculate the arm and weight mass themselves. (Use GPT Chat or Google formulas).
Disassembly:
Shibumi provides disassembly instructions, but only for the EQ8 version—not the EQ8-R. However, the assembly is similar; anyone who has done any mechanical work can handle it. The main differences are the lack of encoders and slightly different nuts tightening the axles.
Be careful with the very delicate ribbon cable connectors! They are small and require very careful handling, preferably with tweezers.
📷 mark plugs.jpg
I mark them all and carefully photograph them to avoid any problems during reassembly. (I still managed to insert one plug into the wrong socket...)
I clean the gears and examine them with a magnifying glass:
📷 broken teeth found2.jpg
I found a broken tooth, luckily in an area that does not engage with the drive.
Another note: if you're only installing the clearance adjustment modules, you can skip the next few paragraphs.
But if you've decided to replace the axles—like I did—there's a kind of „crescent moon” on each axle. It's used to adjust the Home position during assembly. This must be transferred to the new axle in the exact same position as before.
For the DEC, this takes the form of a nut (note: locked with small screws)—a straight line through its center is important. Also, pay attention to the direction of the bulge, as it can be installed backwards. It's worth aligning the line with one of the Losmandy saddle mount screw holes! And remember, where the top of the saddle comes out in the mount „Home” position…
For the RA, this is a flat plate also with a crescent moon, which we fit in relation to the locking screw holes in the RA axle. The straight line of the crescent moon must be somewhere halfway between the angle formed by the two locking screw holes. The description sounds intimidating, but once you get your hands on it, everything will become clear. Adjust it as precisely as you can – but you don't need to be super precise, using who knows what tools, measuring fractions of a degree.
Ultimately, the Auto Home mount should align relatively evenly, as if pointing north. Mine didn't do this perfectly brand new out of the box, with a few degrees of deviation visible – which didn't stop it from finding objects, and we all use Plate Solving anyway…
(Unfortunately, I mislabeled the crescents, and then had to set everything up from scratch, running the partially assembled mount on a table. This can be done too.)
The new axle kits include new precision bearings – following Shibumi's advice, I didn't replace the outer races of the RA bearings;
📷 Outer raceways left.jpg
I kept the old ones (those are the ones indicated with the screwdriver).
NOTE – the kit didn't include grease for the new bearings. You can use any lithium grease that doesn't age or thicken down to -25°C (try thickening the grease in your home freezer). Of course, the bearings are precisely fitted, so don't use force, a steel hammer, etc.; you can tap them lightly with a plastic or rubber mallet. They'll eventually fit. Freezing the axles helps – even in a home freezer.
Low-Duty Backlash Adjustment Kits:
Shibumi provides excellent manuals here, for both the EQ8 and EQ8-R versions. The entire process is described in detail, and installing them requires a set of Allen wrenches (longer ones are preferable, sometimes you have to reach deep). And, of course, good judgement (or a small torque wrench) to avoid stripping the fine threads in the aluminum housings. If we're not replacing the axles, just the kits, then the task is much simpler.
All this without major disassembly, but we do need to remove the motors. This can be done carefully without disconnecting the most of the electronics connectors. And importantly, the 0.2 Nm forces on the worm gear axles are already factory-set. No need for special torque wrenches (or my solution with a bar and weight).
Two notes here:
First, please be aware that the pressure is achieved through the elasticity of the aluminum, not the included springs! These are only used to facilitate the alignment of the entire block.
Second, it generally takes several attempts until you get everything aligned. A caliper and a little patience are all you need. As I said, the entire process is described in detail.
The only thing that can be unclear is how to properly adjust the block pressure. The manual shows a single photo of the indicator lines. I recommend contacting Shibumi and provide your setup weight, and they'll advise you on the correct setting.
For example, for a ~20 kg setup:
📷 Too much.jpg
This was too tight (arrow closer to the middle line), and this is below:
📷 proper settings.jpg
this was OK. Also note the gaps near the lines; you can see that the system is pressed harder in the top photo. Overall, the range of movement is limited.
After installing the sets:
📷 Motors installed.jpg
The belt and clamp arrangement is as above. The clamps are printed; I'd expect metal elements, but that's how it works.
After assembling everything, it's worth checking how the assembly adjusts itself to its initial position on a table.
Balancing the setup: Shibumi states that after assembling their kits, the setup balances evenly in both axes! Not heavy on the West, East, or South, just as even as possible.
After the first run, Shibumi asked me for PHD2 logs, they analyzed it and sent me a report, short summary below:
Observations:
1. Very low RA worm PE – sub 1” Pk-Pk.
2. The FFT suggests some cyclic timing drive error contributions but these are small (under 1”) and have no discernible effect on the RA motion as compared to the DEC motion.
3. Toothing errors on the RA worm wheel generate a bigger guiding error (8” Pk-Pk) than the worm. This is normal on hobbed worm wheels and is an artifact of hob design. As the error is sinusoidal and has a periodic time of ~13 minutes it is very easy to autoguide out.
Recommendations:
1. For short focal length imaging no further action is needed. The system should provide round images provided the finderscope-type guider is fixed without flexure to the imaging scope.
2. There is negligible worm periodic error, so no point in favouring self-learning PHD2 algorithms that attempt predictive periodic error compensation. I therefore recommend either Z-Filter (can be hard to get this to work well) or straightforward Hysteresis be used.
3. If there is any intention to move to longer EFL imaging, where pressure on accurate autoguiding is far greater, I suggest switching to off-axis autoguiding with IR filtration of the monochrome guider camera, which should be equipped with a big prism and adequate camera sensor size. This should be used with SkyGuide which does a better job of discriminating ‘mechanical’ noise of mount origin from ‘seeing’ noise that is atmosphere-related.
My results:
The mount worked with an Esprit 150 @ 1072mm, a 60/2400 guider, and an ASI290M guide camera.
Before replacing the axles and installing the Shibumi kits, best results were an RMS total of 0.31px/0.77". (RMS RA 0.22px/0.54", RMS DEC 0.22px/0.77") – but only before the meridian and not too high. Beyond it, and anything above 75° degraded the tracking to an RMS total of 0.49px/1.23". (RMS RA 0.39px/0.97" RMS DEC 0.30px/0.74"). All parameters from one session: NINA reported an HFR of 2.50 – normal seeing. (Sometimes it was far worse, but havent photos and can’t prove it.)
First run after the modification:
📷 1. PHD2 Zfilter on defaults.jpg
PHD2, Z filter algorithm, settings suggested by Shibumi. RMS total 0.30px/0.74" (RMS RA 0.25px/0.62" RMS DEC 0.16px/0.41") This was the first run after the rebuild, without any parameter tuning. I have no experience with PHD Z-filter, so I could probably squeeze a little more out of them.
To test, I changed the algorithm, and since Shibumi doesn't recommend the PPEC algorithm, I set the standard Hysteresis algorithm on both axes, with slightly improved parameters:
📷 2. PHD2 Hysteresis after initial tuning.jpg
I immediately achieved an RMS RA of 0.13px/0.32" and an RMS DEC of 0.11px/0.26" and a RMS total of 0.17px/0.42".
Seeing was average to bad, with NINA reporting an HFR of 2.6 to 2.9.
After an hour of use, I still had a similar result, with the RMS DEC dropping to 0.09px!!! Needless to say, I've never seen anything like it on this mount...
📷 3. PHD2 Hysteresis after an hour.jpg
In fact, I still have a long way to go; I just got it all working and tried it one evening.
Regarding guiding – as Shibumi recommends Skyguide from Innovations Foresight, full-frame guiding, preferably in NIR. I just installed it, I'm barely getting the hang of it.
First tests show a smoother graph and even better parameters: RMS total 0.12px
I've currently switched from the guider and PHD2 to OAG and SkyGuide, but as I'm having problems with Ascom communication between N.I.N.A., SkyGuide, and the rotator (Wanderer). I'll post the guiding results once everything finally starts working.
Clear skies!
Kuba Myśluk
Warsaw POLAND