Hoping someone with more experience than me can explain this to a beginner.

Does this mean the min-move for both axes in PHD2 should be set to 0.56" (converted to pixels of the guide cam ?)
What practical effect does it have?

I assume it doesn't limit tracking to 0.56" in RA/DEC axes, or does it ?

Tony (GuitsBoy),
Sorry I was wrong in my PEC description. The sensor is on the big pully with an extended shaft (. The other end directly connects to the worm.
I have used 2 designs where I changed the pulley for a 1/4" shaft (including bearings with 1/4" inside dia) and? used a stiff direct connect to the worm.
Or another design with a pulley with a 5mm shaft?(bearings for 5mm shaft) where I use a coupler to translate to the worm shaft.
The 1/4" design does not allow any flexibility so does not work with a spring load design.
Since I use a MaxPCB, I have to use a EthernetAddOn (OnStep Wiki) which has encoders inputs and PEC input which have to be 3.3v translated to 5v.s for the
PEC sensor and the encoders (AstroDevices G11 sensors).
Sorry all this is probably needless information. But I am seriously interested in you design from issues I have had trying to implement a DIY spring loaded design.

Attached is the rockmover modified gear reduction images with the PEC sensor. I had to lower the case over the large gear to prevent interference with the AstroDeices encoders on the G11 axes.

thanks again? Brent?

开云体育

Thanks Brent -

If the PEC sensor is directly reading the worm shaft itself, it would be on the large pulley which attaches to the worm itself, and not the motor.? If the PEC sensor can be divided by the appropriate ratio in the config (ie. 1:3.75) then you could have it on the smaller pulley which is connected to? the motor.? I would think having it on the motor would have greater accuracy, but you would have to find a motor with an extra long shaft.?? Putting it on the large pulley side would probably be easier, since you can simply buy a 5mm shaft in whatever length you desire.? Either way, it seems like you would just have to drill through the cover to pass the shaft through.? Sadly I'm a little unfamiliar with how it actually works in OnStep.? I don't run native PEC at all, and only occasionally use the Predictive PEC algo in PHD2.

As for tensioning the belt with this bracket, its very simply.? Just gently hold the motor to one side to tension, and snug the screws with the other hand.? I didnt find rockmover's brackets to be terribly difficult to tension, but since the cap screws are inside the bracket, it might be tougher to get the hex key in? there.? It's certainly tougher to initially put the screws in.

There's a slight chance the clouds might hold off for an hour or two after sunset, in which case I might be able to test the brackets tonight.? Otherwise, wednesday is? the next earliest reasonably clear sky here in NY.

Thanks,
-Tony

Thanks GuitsBoy,

I apologize, I did not read your first post detailed enough as I was too excited to see the simpler approach over rockmover.
When I read it again I realized it was stabilized. I agree that the idealer pulley probably is not needed. It was just hard to tension
the belt with rockmovers design.
As for the PEC, I believe that as long as the large pulley shaft (attached to the motor) extends out your cover
a PEC sensor (if you want I could send you the url) double backed to your cover and attached to the shaft. Then the correct?
ratio added to the config.sys for PEC.
Thanks for the additional images?

Brent

开云体育

Hi Brent -

We're getting close, but I still want to star test and make sure it runs well.? I already discovered one oversight.? The big pulley was off center by 1mm, which means it could potentially rub against the inside of the cover.? I just aborted a 17 hour print half way? through once I realized.? I'm re-printing the latest revision now in PETG with 100% infill, and will use the new bracket and cover on the RA axis.?? For now, I'll use the off center bracket on the DEC axis with a big pulley I had previously turned down slightly to reduce the OD for use with the rockmover brackets.? The RA is the more important test, IMO.? Once I know it works, Ill send over the STL files so you can tweak however you want.

The brackets do use two bearings to support the big pulley shaft.? It does not rely on the coupling for stability.? The bearings are separated by 15mm, so the shaft is well supported and stabilized.

Personally, I don't think a tensioner is really necessary.? If the GT2 belt were to stretch to the point where a tensioner would take up slack, the teeth would no longer be a perfect mesh and would likely be contributing error.? But Ive used GT2 belts for a few years now and I've never noticed any stretch at all.? Seems like an unnecessary added complexity in my opinion, but I'm sure you could rig something up if you really want it.

Where would you want to put the PEC sensor??? Are you talking about putting it on the shaft?? Or integrating it into the big pulley?

Here's a few photos of my progress on the brackets:

Thanks,
-Tony

There are a number of reasons to change out the gearbox from Losmandy. I have the original ones with plastic tabs, their modified version and the 25:1 McLennan. I built my observatory early January in subzero conditions using heat plankets to set the concrete pillar etc. Having an observatory has been a game changer for me with not a single imaging night missed this year. In those winter months my rms error was 0.8 down to 0.5 arcsec rms. This was with the stock gearbox and the one piece worm on the Ra and original worm on Dec. Even with these good tracking numbers my RA was always 30% worse than Dec. So room for improvement I feel. When you look at the FFT response the 32s error was quit high at 0.2 arcsec rms or higher. Very low 76s error and the is some 240s as might be expected. The high 32s error in my mind a definite feature of the Losmandy gearbox they use and is ultimately the limiting factor on performance, assume one knocks down the 76s error. PEC does a reasonable job knocking down the 80s and 240s errors for me.

If you go over to the McLennan gearbox, the 32s will go away due to a different internal gearing arrangement. But you may introduce a significant 9.3s error if you do not reamer out the pinion gear accurately enough to fit the 1/8th inch shaft of the high torque motor. This does not get quided out and a noticable sinewave will be seen in the tracking error graph. Further, I was not impressed with the build quality of the 25:1 ratio gearbox in that there was much play in output shaft relative to the bushings, potentially making worse the 80s and 240s error. You may well trade one set of problems for another.

If you are going to the trouble and expense of changing out the gearboxes, the other option might be to use either a 50:1 or 80:1 gearbox. Early indications are there is significant benefit to doing this in terms of significantly lowering the rms racking error. Presently, I have only experience with the 125:1 gearbox, but a PEC indexing limitation seem to preclude using this high a ratio. Moreover, rpm limitations will drastically reduce slew rates. In the past some have worried that reaming out the very small pinion gear to 1/8th inch might be a problem. I did not find this to be the case.?

In conclusion I would defiantly suggest going to the 50:1 gearbox, but the sweet may well be the 80:1 ratio.

开云体育

Hi!

Here's the description of the improved feature in a discussion thread:

I hope this is useful. I also use this routine at times, althouhg with the NCP visible. The process is very straightforwads and well explained in the software interface, I think.

Best,

Magnus

Hi David,

Was just looking through your log with PECPrep and this looks like a guided run. For a good mount analysis you need to restart PHD2 and do unguided output. Here you must exit PHD2, not just close it. Then restart and select a star, but uncheck use guided output, and also make sure the Gemini PEC box is unchecked. Run about 30-40 min. and exit again when you are done. That way you can see how the periodic error and other anomalies look uncorrected. Sorry if this was not explained earlier.?

Just looking at your? PHD2 screen shot, your scatter plot looks well centered and the total error was around 0.5 if I am reading the fine print correctly. But the RA seen here looks like it is see-sawing, suggesting a PEC curve may be needed. But this is hard to say with one screen. You need to observe over time. Please see attached doc on how to look at logs. If you are seeing a 76 error a PEC curve does not help as the period will not repeat and overlay the worm period; with each cycle the peaks and valleys between PE and 76 sec. will always differ. Many here have redone their worm blocks and bearings to change this error as Chip said. But I think you said this was a new mount so you probably don't want to rebuild and experiment which will void warrantee. I don't have the spring loaded worms so I can't advise on which adjustments might help here. Suggest you stay in touch with Brian V. on this.

PEMPro is pretty intuitive to use if you follow the Mount Wizard in the upper drop down menu. I think this software is still free for the first month or two of use. You can check out the Losmandy video on the subject for a good tutorial. PEMPro will let you upload the PE correction curve right to the mount and you can see if guiding is improved with PEC Off vs PEC On. With a good PEC curve PHD2 doesn't have to work as hard to make adjustments. But if your 76 second error is large this still presents a challenge.

John

开云体育

Thanks for the details. My view of the polar region is limited, and I can’t get below about 45? altitude around the celestial equator so drift alignment is difficult. ?This KStars procedure is very promising.

On Sat, Jun 19, 2021 at 02:39 PM, Dwight Fujita wrote:

Probably late to the party, seeing as Brian Valente started uploading tutorial videos on the subject.? Thought I'd share here in case others might find the information useful.

HI Dwight

you're not late to the party at all :)

There are too many options to cover them all, but it's good to hear about this one.

if you happen to know and can recommend a good tutorial on this approach (or if you have the time and inclination, record one yourself) we'd love to include it in additional resources "roundup" that i think i will do as well

Think of it as your car engine oil change. How often, how far, how hard do you drive it. Water contamination or corrosion is what to look for whether it's always out or stored indoors. Always good to know what's under the hood anyway and how it ticks. Doesn't hurt to understand and helps with using it.

On Fri, Jun 18, 2021 at 03:30 PM, alan137 wrote:

Please adjust the worm properly and don't try to "power through" a stall.

That's the whole point behind the idea of changing the gear ratio. Easier on the motors and smoother running.

On Sat, Jun 19, 2021 at 12:48 PM, <pcboreland@...> wrote:

Chip, Bearing grade I think has a lot to do with the tolerances that affect the seating in the one piece worm (oxymoron) blocks a relative to the relative worm shaft, and if there is play.? At least I found this to be the case. Preloading is key and I did read Mark's paper on various ways this can be done. Is this the one you are referring to or is there another document? Interesting you are arguing for a looser fit of the bearing and the shell. I need to read this document I think. It's important that we stand on the shoulders of other who have been through this time loop before.?

Peter

Peter,

The ABEC bearing grades have nothing to do with external dimensions. The ABEC grade is a standardized set of internal bearing tolerances. The normal commercial bearing grade is ABEC-3

If your new bearings were supposed to be ABEC-5 and they fit no? friction into the bearing blocks those bearings are probably junk.

??
--

Chip Louie Chief Daydreamer Imagination Hardware?

? ?Astropheric Weather Forecast - South Pasadena, CA?

I use the my main scope for the alignment procedure, fl=1000mm on a 3.8um sensor, or about 0.78"/pixel.? I've done it with my guide scope (253mm, 2.9um, 2.36"/pixel) and it also works fine.

I can't see the pole from my imaging location, so I usually point my optics at the equator.? Azimuthal adjustments are usually good the first time, but I sometimes need to retry dialing in the altitude adjustment.? Usually after the first try my misalignment is <5'.? After the 2nd attempt it's usually ~1', so the 3rd time is usually to verify the result.

Sensitivity to polar misalignment is different depending on what part of the sky you point at, and in my case, the procedure ends in the western sky.? My guess is that flexure and gearing backlash throw off positioning of the reference star when I make adjustment to the alt-az knobs.

The alignment procedure gives repeatably good results after I've correctly set my alt-az knobs, so I'm pretty sure the problem occurs after the 3rd image capture and when I make adjustments.

My guess is that this would work much better if I could point my scope at the pole.? Even with iterating the procedure, it's still faster than drift aligning in PHD2.

Thanks,
Dwight

Hi Edward.? Yes, I understand.? My issue is that one side can be pulled out along the axis of the coupler while the other side is tight with no play.

------------------------
Jim W
Phoenix, AZ. USA

Losmandy GM811G

Jim, In case it wasn't clear, an Oldham coupler is *supposed* to slip along those groves yet not have play when the shaft is twisted. The purpose of an Oldham is to allow connection of two shaft that are not perfectly aligned along their axes of rotation.?
--
Edward

开云体育

Thanks for the tip, Dwight.?

I use it too, on a Pi4b with an OnStep controller on my G11S.? Why do you need multiple iterations?? I use it near the NCP and 1 iteration works, it's a one shot method.? However one time I was off by quite a bit and the line was too long to fit in the FOV so I went as far as I could and repeated.? Is that why you iterate too?

OK

------------------------
Jim W
Phoenix, AZ. USA