Re: Yet another DIY spring loaded worm - ultra minimalist approach
Sorry about the duplicate / inverted pics. After this point, just tighten left block snug / flush and right block just off snug / flush.
.
|
Re: New G11G Looking for final tweaking advice
On Sat, Jul 10, 2021 at 01:54 PM, <pcboreland@...> wrote:
Nike,
Is this a new or old mount?
That backlash number doesn't surprise me. I worked on and this on and off for months. I spring loaded my worm and decided to just lock things down instead. I get the mesh as tight as possible by rocking the axis back and forth. Don't forget to load the bearings so the worm does not move laterally. I do this with the cover off so I can see if there is any lateral movement in the worm. Then I replace the cover. You may find the motor stalls so you will need to back off the work/gear mesh pressure with the adjustment screw. You should do a full 180 swing to see there is not a stall spot. That said, with all that attention to detail, my backlash is 2500ms, which allows me to guide OK. I've now traced the backlash problem to the slop in the needle bears which is significant and explains I think the whole problem. Also, I found the same issue in the Ra axis.?
Peter
Mount is just a few weeks old. I was hoping not to have to start manually messing with adjustments especially with the new improvements that have been done on the mounts.?
?
|
Re: New G11G Looking for final tweaking advice
Nike,
Is this a new or old mount?
That backlash number doesn't surprise me. I worked on and this on and off for months. I spring loaded my worm and decided to just lock things down instead. I get the mesh as tight as possible by rocking the axis back and forth. Don't forget to load the bearings so the worm does not move laterally. I do this with the cover off so I can see if there is any lateral movement in the worm. Then I replace the cover. You may find the motor stalls so you will need to back off the work/gear mesh pressure with the adjustment screw. You should do a full 180 swing to see there is not a stall spot. That said, with all that attention to detail, my backlash is 2500ms, which allows me to guide OK. I've now traced the backlash problem to the slop in the needle bears which is significant and explains I think the whole problem. Also, I found the same issue in the Ra axis.?
Peter
|
Re: Changing the servo motor encoders from 256 to 512 or 1024. Any reason not to do this?
I moved this discussion to its own thread:
/g/Losmandy_users/topic/needle_bearings_and_axial/84141331?p=,,,20,0,0,0::recentpostdate%2Fsticky,,,20,2,0,84141331
|
Re: New G11G Looking for final tweaking advice
On Sat, Jul 10, 2021 at 12:39 PM, <pcboreland@...> wrote:
Nick,
I looked at your log file. Several things jump out at me.
1. I noticed you have backlash compensation turned off.? Have you run the Guiding Assistant to determine your Dec backlash. I notice there are times when it takes quite a few pulses to make the correct and the pulse width is small at 20ms.?
2, Your pixel scale is 6.45 "/px. By comparison mine is 1.24 "/px.? At this low resolution I'm not sure one can put too much into the guiding numbers, your numbers will look super good as you can not really detect true movement.? But others here are way more experienced on this front and I'm sure I'll get shot down for saying this.. Personally, I think the guide resolution ideally ought to be under 2x the imaging resolution.?
3. Your 240s error is at 4" to 4.5" peak-to-peak give or take. But that is with running the Predictive PEC algorithm. You might what to do a run using say the Hysteresis algorithm with Mount PEC off to get a baseline value. I reduced my native error down to just under 3 arc-sec peak-to-peak by installing a rigid coupler in the Ra. This way the motor shaft and worm are better aligned.? But your number is already low and it guides out really well to around 0.15", so is under control.
4. You have a very large low frequency component at 1100s or so. Not sure what might be causing this?
5. Those spikes are the biggest issue for me, and the elephant in the room is the 76s error, which is surely giving rise to those large transient moves in Ra. PHD2 is only able to reduce it by 2/3rds. There is a lot out there on this problem. With the OPW I do not see this error but I have an equally large one at 80s. Fortunately, I think I can build a PEC curve using PEMPRo to eliminate it. At least that's what I hope.
Overall, this are pretty good results with the caveat that things might look different at a higher guiding pixel scale. Time to take some pictures!
Peter
Thank you so much Peter
1. Guide assistant said my backlash was 3500 ms so it did not recommend turning it on. It recommended guiding in one direction which did not work well. Brian V said that PHD cannot accurately measure BL on the spring-loaded worms, so not to enable the comp
2. Completely agree. I got a small guider setup for my old star tracker as I did not want to have too much weight. I am considering upgrading but mostly only if it will improve my images.?
3. I did a baseline here and it was I believe 7.5 to 7.5 peak to peak (it's in my other thread here). That seems a bit more than expected but Brian V said it was smooth and would guide out as you said
4. I have no clue on the 1100. I tried to get my cabling neat and tidy and nothing snagging etc. I'm on a driveway so cars do sometimes pass by but I am sure they are not coming at exactly 1100 second intervals :)
5. Yeah the spikes worry me but I dont have much recourse here as I am not wanting to muck with the mount and I don't really have a clue where they come from (maybe its the 76-second issue ?)
I'm not letting this stop me imaging, but I dont want to image with poor results. The spikes are my main concern. I will adjust the weight back a bit next time to see if it can help at all
I'm going to look at upgrading the guider. I think I can keep the scope the same and get a better camera and not have to significantly change my attachment to the scope
I'm working on an image now but may have picked something too tough for my gear and skies. Going to give it a few more nights and then consider moving onto a new target
Thanks so much for your feedback !
EDIT (sorry)
This is a brand new mount so I do have all the latest upgrades etc on the hardware side
|
Re: New G11G Looking for final tweaking advice
Nick,
I looked at your log file. Several things jump out at me.
1. I noticed you have backlash compensation turned off.? Have you run the Guiding Assistant to determine your Dec backlash. I notice there are times when it takes quite a few pulses to make the correct and the pulse width is small at 20ms.?
2, Your pixel scale is 6.45 "/px. By comparison mine is 1.24 "/px.? At this low resolution I'm not sure one can put too much into the guiding numbers, your numbers will look super good as you can not really detect true movement.? But others here are way more experienced on this front and I'm sure I'll get shot down for saying this.. Personally, I think the guide resolution ideally ought to be under 2x the imaging resolution.?
3. Your 240s error is at 4" to 4.5" peak-to-peak give or take. But that is with running the Predictive PEC algorithm. You might what to do a run using say the Hysteresis algorithm with Mount PEC off to get a baseline value. I reduced my native error down to just under 3 arc-sec peak-to-peak by installing a rigid coupler in the Ra. This way the motor shaft and worm are better aligned.? But your number is already low and it guides out really well to around 0.15", so is under control.
4. You have a very large low frequency component at 1100s or so. Not sure what might be causing this?
5. Those spikes are the biggest issue for me, and the elephant in the room is the 76s error, which is surely giving rise to those large transient moves in Ra. PHD2 is only able to reduce it by 2/3rds. There is a lot out there on this problem. With the OPW I do not see this error but I have an equally large one at 80s. Fortunately, I think I can build a PEC curve using PEMPRo to eliminate it. At least that's what I hope.
Overall, this are pretty good results with the caveat that things might look different at a higher guiding pixel scale. Time to take some pictures!
Peter
|
Re: New G11G Looking for final tweaking advice
Here is a link to two RAW subs - now we are thinking about "can we see the effect in the images"
This was something David suggested I do to see the effect
You can see in the 90 second sub, the FWHM is lower and the eccentricity higher compared to the 900 second sub
Obviously the 900 is getting more exposure to dim light, but the question David proposed to me was, is the 900 sec sub just smearing the error around a bit, bloating stars and making them appear more round and less eccentric (I measured a bunch of subs and the pattern held for all subs)
Or is the 900 sec sub just bringing in more light, so getting a better image that is more "correct" ?
I don't really know how to answer that.
Some of it may be the duo-band filter, which aren't known for being super high quality filters. I may bite the bullet and remove it and tweak the spacing and see if I get different results.
Part of it is I am not sure what expectations are reasonable and what aren't
|
Re: New G11G Looking for final tweaking advice
On Sat, Jul 10, 2021 at 09:51 AM, Chip Louie wrote:
This issue is part of what the Belleville disc springs resolves by taking up all backlash axially. the worm axels are sized as they are for the real world and ease of maintenance and replacement. These bearings are consumables and as such are replaced as they get dirty and noisy. For visual users none of this makes any difference at all as you can barely even measure a difference in PE using PHD2. But it is there if you want to "fix" it.? ? ?
--
Chip Louie Chief Daydreamer Imagination Hardware?
? ?Astropheric Weather Forecast - South Pasadena, CA?
Thanks Chip. I only want to fix it if it will help my images and I have no idea if it will. Also, not being mechanical, I am more likely to do harm than good most likely
|
Re: G11 Spur Gear Removal to Access Gearbox
Peter,?
What do you mean by "hammer action"?
--
Chip Louie Chief Daydreamer Imagination Hardware?
? ?Astropheric Weather Forecast - South Pasadena, CA?
|
Re: New G11G Looking for final tweaking advice
This issue is part of what the Belleville disc springs resolves by taking up all backlash axially. the worm axels are sized as they are for the real world and ease of maintenance and replacement. These bearings are consumables and as such are replaced as they get dirty and noisy. For visual users none of this makes any difference at all as you can barely even measure a difference in PE using PHD2. But it is there if you want to "fix" it.? ? ?
--
Chip Louie Chief Daydreamer Imagination Hardware?
? ?Astropheric Weather Forecast - South Pasadena, CA?
|
Re: G11 Spur Gear Removal to Access Gearbox
Brett,
I'm sure you have solved this problem already, so this is for others with similar issue if stuck Allen key screw. I've found a driver works the best. You can buy driver hex key sets easily online. The hammer action does the trick without stripping the key hole.
Peter.
|
Re: New G11G Looking for final tweaking advice
So, specifically here is what I would like to correct (it's not stopping me imaging right now, but I am tweaking software and balance each time etc)
This was over a single 15 min sub, so it isolates the data very well
The spikes down to -2.5 worry me Anything within 1" area which is most of the graph, I am not worried about
I noticed anecdotally it might have happened more often when the counterweight approached horizontal last time, so I moved my weight down a bit (East) but that seemed to have no impact. It's not a very big weight though
|
Re: Yet another DIY spring loaded worm - ultra minimalist approach
After two weeks of wonderful and badly needed monsoon clouds/rain over Tucson (heresy I know), I finally got a chance to test this simple hack when a part of the sky cleared near zenith. It turns out it worked very nicely with the guiding and motor loads. My DEC backlash is down 75% and RA guiding RMS is <1" in poor seeing. I decided to change the position of the DEC band to be in the same 45¡ã position as the RA. This should give a bit more stability along the worm axis.
BTW - the reason you don't put a band on the block near the motor is due to that hole not having the adjustment clearance of the outer block.
Here are few pictures that I took to document the procedure (note high-tech labeling on the band from the local produce section of the grocery store).
|
Re: New G11G Looking for final tweaking advice
This is an interesting read on what looks like a significant issue /g/Losmandy_users/topic/g11_fix_source_for/21183709?p=,,,20,0,0,0::recentpostdate%2Fsticky,,,20,1,40,21183709"
Does anyone know of a source for undersized needle bearings to fix a
weakness in my Losmandy G11 mount? Specifically, 1.247" ID / 1.500 nominal
OD. For the benefit of those who may encounter the same problem, I have
described it below:
With a nominal vertical load, the RA and/or DEC shafts on my G11 mount tend
to 'hang' about .002"-.003" below the axis centerline when the clutches are
relaxed. When the clutches are tighten, the shaft end plate initially
'locks' to the nylon clutch surface slightly below center. When the axis is
rotated far enough that the weight overcomes the grip of the nylon surface,
the axis 'shifts' into a 'more' aligned position. This is normal behavior on
four different mounts that I have examined. Besides introducing tracking
errors, this misalignment distorts the worm/wheel spacing by a small amount,
and causes the backlash to vary as the axis is rotated through 360 degrees.
Changing the load on the mount, or loosening and retightening the clutches
can completely upset prior backlash adjustments. If you are not adjusting
your worm/wheel spacing for absolute minimum backlash, you may never
encounter this aspect of the problem.
The needle bearings on both the RA and DEC are set back about 2" from the
'business end' on each axis. As an experiment, I pressed an additional
1.250" ID needle bearing into the opening so that the shaft was supported
nearly to the end of the opening. Additionally, the shafts I have seen
measured 1.2470" dia instead of an optimal 1.250". I wrapped some .0015"
stainless steel shim stock around the undersized shafts as a temporary fix
to remove the bearing slop.
Well, it was like night and day. That simple change eliminated the
'shifting' on each axis, improved tracking and GOTO pointing accuracy
dramatically, there was no more slop in either axis, and best of all, the
backlash can be adjusted to virtually zero through a full 360 degrees of
rotation.
Anyone else have a similar experience?
Neal
"
|
Re: New G11G Looking for final tweaking advice
On Fri, Jul 9, 2021 at 10:11 PM, John Kmetz wrote:
Each new worm new has a surface profile on the tooth faces right off the CNC machine, just as the rings gears do. Some have tried to use polishing compounds in the contact zone and other methods to wear things down a bit, but I don't think this was entirely successful. Letting the mount run over time is the best way to break in the gears as the surfaces will smooth down and mesh better over time.?
I have new worms running in both axis. I have no expedience to know how things change as they wear or how long that takes. To what extent are the one PHD2 tracking spikes a function of the worm surface? Does tracking improve or get worse as they wear in? Since I have a few worn worms I thought I might lightly buff one while spinning in an electric drill. I have to assume a polished worm has less friction/stiction.? Another question I have is what effect do the needle bears have on the system. Do they have a frequency profile? I do have some axis movement. My rig is rather old, and I suspect these have never been renewed. However, it looks like a difficult job, not to be taken on lightly.? Peter
|
Re: Changing the servo motor encoders from 256 to 512 or 1024. Any reason not to do this?
On Fri, Jul 9, 2021 at 06:10 AM, Michael Herman wrote:
Alan Fang has that GM811 version and he is also investigating its behavior (or he is happy and his system is stable and in use).?
?
?
Alan, while visiting his parents in southern California, stopped by the humble Losmandy building and swapped out his G8 DEC axis for a G11.? They were very nice to give me a credit towards purchase of the new axis.? So now I am a G11 owner. In the end, I had the G8 axis (while using a 40 lbs 10" F4 Newtonian) working "ok, but not trouble free".? Which means that on a day with good seeing, I could crank out many good frames in a row, but will occasionally have to cancel a frame before completion due to excessive DEC movement 30 seconds after a dither move, etc.? Given that it has a "resolution" of 1.x arc-sec, I would consider that it is good only up to 1000mm focal length and about 40 lbs.? The g11 has a worm wheel twice as big, so I expect all errors to be half as much, and a much better baseline to work from.
|
Re: New G11G Looking for final tweaking advice
David,
You may be going over some well worn paths to better worm performance which sound familiar. Searching through the forum might provide more insight. If you had a deep nick in the worm, like the outer edge of a tooth, you would probably see a sharp spike every 240 seconds, or every time the worm goes round once and hits the bad spot. If you weren't seeing this effect, the nick was probably not that significant. The worm and ring gears don't ride on the points, but a bit further in on the teeth faces. If you try to polish the worm by hand, you most likely will not be able to evenly remove metal from just the right spots and you could knock it out of round, creating further errors. Each new worm new has a surface profile on the tooth faces right off the CNC machine, just as the rings gears do. Some have tried to use polishing compounds in the contact zone and other methods to wear things down a bit, but I don't think this was entirely successful. Letting the mount run over time is the best way to break in the gears as the surfaces will smooth down and mesh better over time.?
If one worm is doing less than desirable guiding, swapping with the Dec worm is a quick test for possible improvement. The Dec worm has less effect on motion as it rocks back and forth to make corrections while guiding, and really only spins completely around while slewing, and wears more slowly. The RA worm is constantly turning and therefore creates most of the periodic errors. Once you have things apart, you may want to try the Bellville washer modification to reduce the 76 second error. But now you are talking about disassembly and modifications which are not under warrantee.? But it sounds like you may have gone that far already.
Hope you get to where you want to be.
John
|
Re: New G11G Looking for final tweaking advice
On Fri, Jul 9, 2021 at 04:57 PM, <pcboreland@...> wrote:
On Fri, Jul 9, 2021 at 08:06 AM, Nick Ambrose wrote:
I do see fairly often, RA going to +/- 1.5-2" deviations (usually it is well inside +/-1" on my graph
Nike,
It's those transients (or just one) that I've found are the real problem with loosing what would be great frames shooting long exposures. I'm working hard to eliminate them with success! You know nice perfectly round stars with a little brother or sister just of to the right.?
You may already have said already, but what is image scale is your guiding system??
Peter
Right now I have a very forgiving 2.2 arc sec/pxl, so in theory any blip I suppose under (approx) 1.1" deviation would expect not to cause any issues ? I am pretty happy with my images but I can't help but think these must have some kind of an impact
|
Re: New G11G Looking for final tweaking advice
Peter,
Have no Idea what the RA blips are, maybe the worm nick?? I'm thinking the DEC blips might be a bearing glitch like the RA's 76 second error. They seem to be evenly spaced. Will have to time them.
Never noticed the worm nick till adjusting it. Spinning the gears while adjusting the bolt next to the spring, the worm stuck a little bit in the same spot. Took off to clean, re-grease and examined the worm, found the nick.
?
|
Re: New G11G Looking for final tweaking advice
On Fri, Jul 9, 2021 at 09:51 AM, David Malanick wrote:
Have a new RA worm on order because I noticed a nick right in the center of the worm that was causing a slight binding when meshed.
Under a magnifying glass I used 2500 grit sandpaper to fix a bulge caused by it but it's still noticeable.?
Then I looked through the magnifying glass and saw it was a little rough, along with the whole worm.
Took some ultra fine automotive final mirror polish on a buffing wheel and polished the worm all around for a few seconds.
Will replace that repaired worm with the new one when it comes and check the results.
The nick in the worm seems to be guided out by PEC.
I'm wondering if polishing the worm had any affect in getting the results I'm seeing?
David, Buffing the worm sounds like an really interesting idea. Also, I've found that nicks are so easily created just installing the worm. Brass is so soft. I'm at about 0.65 arcsec without PEC. I too know I can do better, Those little Ra wonderings in your tracking curve, what do you attribute them too? Peter Peter
|
Greg
Chip Louie
David Malanick