¿ªÔÆÌåÓý

  1. Losmandy_users
  2. Topics

My Experience with Gemini L6.XX

 

I am running a Losmandy Titan mount, and quite a few months ago I upgraded from Gemini 2 L5 to L6.02.? I flashed my encoder ports with the PIC kit, and installed the latest Gemini ASCOM driver, and updated the ASCOM platform.? After installing L6.02, I ran PEMPro and uploaded a new PEC curve, and re-calibrated PHD2.
?
My guiding got excellent.... somewhere around a 20 to 30% reduction in guiding error.? My plate solving got incredibly accurate for me.... easily within 10 pixels per axis.? I also noticed my PHD2 guiding calibrations improved... most all of the step plots were on the axes; whereas, they weren't before.??
?
However, after a week or two, the guiding stayed good, but the plate solving got inaccurate again.... I couldn't get closer than 30 pixels for RA, and 250 pixels for DEC.? It stayed that way for weeks.
?
I finally re-flashed L6.02 (later discovered I could have done an HGM Reset).? The plate solving went extremely accurate again, immediately.... less than 10 pixels per axis.? But, after a few weeks, the plate solving went south again.... back to no better than 30 pixels RA, and 250 pixels DEC.? The guiding still stayed great.
?
So, now, this past Friday, I flashed the mount with L6.12.? The plate solving went great again.... less than 10 pixels per axis, and the guiding was great.? So..... that's 3 out of 3 flashing events where the result was the plate solving got great.
?
I got the release notes from the Losmandy website for the various versions.... L6.02 to L6.12.? I did not see any notes regarding my situation, but I did notice one of the release notes for one of the versions post L6.02 was not very specific, it said, "Many Gemini fixes...."
?
I'm running WIN 10 on my PC, and connect to mount via Ethernet.? I'm wondering if Microsoft Updates could have somehow corrupted L6.02?? I'm keeping my fingers crossed that some improvements were made and L6.12 hangs in there for me.
?
I look forward to hearing from others on this issue.
?
Regards,
?
Dan
?
?
 

Hi Dan,
?
When you see the plate solving went great, and then got worse, are you referring to the initial centering of the object is within 10 pixels at first, then goes to more like 250 pixels in DEC off center? Hard to imagine that anything in L6.x could cause this.
?
This sounds like it could be that some DEC backlash worm-gear adjustment may be needed. Do you have TVC set to 0? What does your PHD2 backlash calibration look like?
?
Do you do meridian flips, and can you tell if the larger centering error occurs just on one side of the meridian?
?
Do you build a pointing model before doing plate solves? If so, how many stars, and are you building it on both side of the meridian?
?
What is your optical train? Refractor? Reflector? Any chance of mirror flop or possibly some focuser sag?
?
Regards,
?
?-Paul
?
On Tue, May 27, 2025 at 07:42 PM, <dkfcpalfd@...> wrote:

I am running a Losmandy Titan mount, and quite a few months ago I upgraded from Gemini 2 L5 to L6.02.? I flashed my encoder ports with the PIC kit, and installed the latest Gemini ASCOM driver, and updated the ASCOM platform.? After installing L6.02, I ran PEMPro and uploaded a new PEC curve, and re-calibrated PHD2.
?
My guiding got excellent.... somewhere around a 20 to 30% reduction in guiding error.? My plate solving got incredibly accurate for me.... easily within 10 pixels per axis.? I also noticed my PHD2 guiding calibrations improved... most all of the step plots were on the axes; whereas, they weren't before.??
?
However, after a week or two, the guiding stayed good, but the plate solving got inaccurate again.... I couldn't get closer than 30 pixels for RA, and 250 pixels for DEC.? It stayed that way for weeks.
?
I finally re-flashed L6.02 (later discovered I could have done an HGM Reset).? The plate solving went extremely accurate again, immediately.... less than 10 pixels per axis.? But, after a few weeks, the plate solving went south again.... back to no better than 30 pixels RA, and 250 pixels DEC.? The guiding still stayed great.
?
So, now, this past Friday, I flashed the mount with L6.12.? The plate solving went great again.... less than 10 pixels per axis, and the guiding was great.? So..... that's 3 out of 3 flashing events where the result was the plate solving got great.
?
I got the release notes from the Losmandy website for the various versions.... L6.02 to L6.12.? I did not see any notes regarding my situation, but I did notice one of the release notes for one of the versions post L6.02 was not very specific, it said, "Many Gemini fixes...."
?
I'm running WIN 10 on my PC, and connect to mount via Ethernet.? I'm wondering if Microsoft Updates could have somehow corrupted L6.02?? I'm keeping my fingers crossed that some improvements were made and L6.12 hangs in there for me.
?
I look forward to hearing from others on this issue.
?
Regards,
?
Dan
?
?
 

Thanks for your response, Paul.? Let me try to address all your questions:
?
I have an Edge 14 SCT on the mount, with the 0.7X reducer.? I have a QHY 600 mono imager, taking visible light through an ONAG.? I typically run plate solving through the luminescent filter, which is an Antlia.? ? I'm solving with ASTAP through SGP.? Normally the first plate match may have a pretty good amount of error, but successive plate matches get successively closer.? I have the solver set to do up to 7 plate matches.? Normally it would stop in 2 to 4 plate matches, and that would be pixel error of less than 15 on both axes (when the mount has recently been flashed).
?
When I said the 30 and 250 pixels, it could never get better than that.? It would start large, and successively go smaller, but really couldn't get closer than 30 and 250.
?
The mount does have large DEC backlash, and I assumed that backlash was the reason for the DEC error being larger.? However, Sunday night, I did some plate solving, and I saw DEC go as close as 1.8 pixels of error.? The solving finally finished at 5 and 7 pixels of error for RA/DEC.? I have spent plenty of time trying to adjust out DEC backlash, and have it operating the best I can.? I know for a fact, that the Titan mount is known to have huge backlash, as one of the authors of PHD2 told me that on the phone.? The Guiding Assistant does show a large backlash.
?
I used to do modeling with the mount, but don't do much of that anymore.? I sometimes go to a bright star, center it, and do a modeling/alignment..... that's about it.
?
No possibility for mirror flop:? I have the SMFS focuser, and the primary is locked down.? I don't think there's focuser sag with the SMFS.? Meridian flips did not cause this.... those are rare for me to do.... I'll do a 2 to 3 hour session per night, and plan it so a flip is not needed.
?
TVC is at the default setting.... don't recall... that's either 0 or 5.? I tried increasing that setting for my autoguiding one time, and the guiding went terrible, so I leave it at the default setting.
?
My question is how could the plate solving all of a sudden get very accurate when I flash the mount with L6.XX, without that being the result of the flashing?? So.... for 2 out of 2 prior flashings, the solving after several weeks reverted back to the more inaccurate state, and that would happen all of a sudden.... I try to solve, and it gets no better than the 30/250.? It stays that way for weeks, until I flash again, and the solving then goes very accurate.
?
It just makes me wonder if MS Updates have somehow corrupted Gemini, or there's something else going on.
?
Thanks.
?
Dan
?
?
 

On Tue, May 27, 2025 at 09:31 PM, <dkfcpalfd@...> wrote:
The mount does have large DEC backlash, and I assumed that backlash was the reason for the DEC error being larger.? However, Sunday night, I did some plate solving, and I saw DEC go as close as 1.8 pixels of error.? The solving finally finished at 5 and 7 pixels of error for RA/DEC.? I have spent plenty of time trying to adjust out DEC backlash, and have it operating the best I can.? I know for a fact, that the Titan mount is known to have huge backlash, as one of the authors of PHD2 told me that on the phone.? The Guiding Assistant does show a large backlash.
Thanks for the information, Dan. Backlash could still be the cause of this, since backlash only plays a role when switching direction of the DEC motor. Depending on how the initial or successive plate-solve/re-center operations go, you may or may not be switching DEC direction, which may result in a large backlash... or not :)
?
I also have a Titan mount, and backlash is something that's not that hard to adjust out. You need a little baclkash, but a very large amount is counter-productive. Don't know if PHD2 authors actually know anything about the Titan, but having operated one for years, I can tell you that backlash can be reduced to an arbitrarily small amount. You just need to make very small adjustments using the wing-head screws to push the worm into the gear. I usually have the worm adjusted with some?
?
It just makes me wonder if MS Updates have somehow corrupted Gemini, or there's something else going on.
Can't imagine how MS update could affect such a complex process as plate solving/recentering. Usually it's really obvious when MS breaks something -- things just stop working altogether :)
?
It sounds like there's enough sag/slop somewhere in your system that causes the large error. If the mount guides well, I can't imagine why re-centering after a plate solve would not bring the mount to the correct position within a few arcseconds of the desired position. Double-check you plate solution settings, maybe there's something there you can adjust. With the large focal length of your system, I have to assume there are not that many stars in the FOV to make plate solves precise and predictable.?
?
 

I had considered DEC backlash as having negative impact.? However, the symptoms of the problem..... all of a sudden appearing...... when I go along with countless plate solves inside 10 pixels on both axes, then all of a sudden, the error on both axes goes much larger, and stays that way for weeks, until the firmware is re-flashed, then it's instantly back to less than 10 pixel error...... made me think there may be some issue going on besides backlash.
?
Well, thanks for your help.? I think if more people had the problem, it may be easier to diagnose.
?
I also explored the idea of static electricity in my dome somehow corrupting the firmware.... who knows.
?
Overall, I'm happy with the upgrade, so I'll just continue on.
?
Thanks.
?
Dan
 

On Wed, May 28, 2025 at 07:55 AM, <dkfcpalfd@...> wrote:
I had considered DEC backlash as having negative impact.? However, the symptoms of the problem..... all of a sudden appearing...... when I go along with countless plate solves inside 10 pixels on both axes, then all of a sudden, the error on both axes goes much larger, and stays that way for weeks, until the firmware is re-flashed, then it's instantly back to less than 10 pixel error...... made me think there may be some issue going on besides backlash.
?
Well, thanks for your help.? I think if more people had the problem, it may be easier to diagnose.
?
I also explored the idea of static electricity in my dome somehow corrupting the firmware.... who knows.
?
Overall, I'm happy with the upgrade, so I'll just continue on.
?
?
It's not a problem I've seen myself or heard from others. There could be other issues here, including something slipping (like the coupler, or the encoder disk), or the encoder ticks not being counted correctly due to a cable or a poor connection.
?

1 - 6 of 6