|
Group,
I have been trying to use a 14 year old GM8 mount (non-spring loaded worms; non-tucked motors) for variable star analysis. For this analysis, it is not essential that guiding is perfect, but I’d like to get it better than I’m currently achieving. This is with a small refractor of 740mm focal length.
The crux of the issue is that guiding in Dec seems to be reasonable (say 0.3 RMS”) but RA guiding is significantly worse (say 1.5 RMS”). I’ve attached a screen shot showing what the RA guiding looks like (blue is Dec; green is RA).
I am using KStars/Ekos for imaging and guiding but I get similar results if I use PhD.
The RA seems to bounce around somewhat chaotically. I do have pretty bad seeing where I’m located, but the fact that the Dec doesn’ bouce around as much indicates this in more than bad seeing.
I have been working on this problem off and on for some time. I have tried.
1. Stripping the mount and completely cleaning/relubbing it 2 or 3 times
2. Swapped the worm and worm blocks between RA and Dec axes
3. Swapped the gear boxes between RA and Dec axes
4. Swapped the motors between RA and Dec axes
5. Shimmed the RA gear box to correct a slight misalignment at the Oldham coupler
6. Tried lubricating the Oldham coupler
7. Shifted the counterweight around to make it East heavy and East light
8. Tried a different power supply
None of this has made a significant difference. I still get poorer guiding in RA.
Anyone have any ideas on what is going on and how I might fix this?
Thanks,
Bill
--
Bill Tschumy
Otherwise -- Longmont, CO
www.otherwise.com
|
Hi Bill, it looks like to me that the autoguider isn't calibrated properly. Corrections are often insufficient in RA, and sometimes overshoot a quick jump. The first step is to measure your periodic error during tracking (capture a few cycles of tracking error while not guiding). See what tracking looks like without autoguider corrections. Then, a guide log may help diagnose what's going on while guiding is engaged. PHD2 also has a number of diagnostic/calibration tools that may help figure out what's going on.
?
Regards,
?
?-Paul
?
On Sat, Sep 28, 2024 at 11:03 AM, Bill Tschumy wrote:
Group,
I have been trying to use a 14 year old GM8 mount (non-spring loaded worms; non-tucked motors) for variable star analysis. For this analysis, it is not essential that guiding is perfect, but I’d like to get it better than I’m currently achieving. This is with a small refractor of 740mm focal length.
The crux of the issue is that guiding in Dec seems to be reasonable (say 0.3 RMS”) but RA guiding is significantly worse (say 1.5 RMS”). I’ve attached a screen shot showing what the RA guiding looks like (blue is Dec; green is RA).
I am using KStars/Ekos for imaging and guiding but I get similar results if I use PhD.
The RA seems to bounce around somewhat chaotically. I do have pretty bad seeing where I’m located, but the fact that the Dec doesn’ bouce around as much indicates this in more than bad seeing.
I have been working on this problem off and on for some time. I have tried.
1. Stripping the mount and completely cleaning/relubbing it 2 or 3 times
2. Swapped the worm and worm blocks between RA and Dec axes
3. Swapped the gear boxes between RA and Dec axes
4. Swapped the motors between RA and Dec axes
5. Shimmed the RA gear box to correct a slight misalignment at the Oldham coupler
6. Tried lubricating the Oldham coupler
7. Shifted the counterweight around to make it East heavy and East light
8. Tried a different power supply
None of this has made a significant difference. I still get poorer guiding in RA.
Anyone have any ideas on what is going on and how I might fix this?
Thanks,
Bill
--
Bill Tschumy
Otherwise -- Longmont, CO
www.otherwise.com
|
Paul,
I have calibrated and re-calibrated many, many times. ?I have also tried a variety of different pulse lengths for calibration. ?If calibration is off then there must be some other fundamental problem that is causing it. ?I’ve attached the Calibration Plot that Ekos generates. ?It doesn’t look particularly bad to me.
I do use a somewhat longer pulse length (around 1250 ms) because if I do shorter the Dec calibration looks worse.
Sometimes the Dec calibration dots can be a bit squirrely although the ones here look fine to me.
I should also mention that I’ve tried adjusting the RA worm play ad nauseum and that doesn’t seem to help.
Bill On Sep 28, 2024, at 9:22?AM, Paul Kanevsky <yh@...> wrote:
Hi Bill, it looks like to me that the autoguider isn't calibrated properly. Corrections are often insufficient in RA, and sometimes overshoot a quick jump. The first step is to measure your periodic error during tracking (capture a few cycles of tracking error while not guiding). See what tracking looks like without autoguider corrections. Then, a guide log may help diagnose what's going on while guiding is engaged. PHD2 also has a number of diagnostic/calibration tools that may help figure out what's going on.
?
Regards,
?
?-Paul
?
--? Bill Tschumy Otherwise -- Longmont, CO www.otherwise.com
|
So the calibration looks good. How about an unguided tracking log over a few worm cycles?
?
This is with Gemini-1, correct? How is your autoguider connected to Gemini?
?
Regards,
?
? -Paul
On Sat, Sep 28, 2024 at 12:44 PM, Bill Tschumy wrote:
Paul,
?
I have calibrated and re-calibrated many, many times. ?I have also tried a variety of different pulse lengths for calibration. ?If calibration is off then there must be some other fundamental problem that is causing it. ?I’ve attached the Calibration Plot that Ekos generates. ?It doesn’t look particularly bad to me.
?
I do use a somewhat longer pulse length (around 1250 ms) because if I do shorter the Dec calibration looks worse.
?
?
?
Sometimes the Dec calibration dots can be a bit squirrely although the ones here look fine to me.
?
I should also mention that I’ve tried adjusting the RA worm play ad nauseum and that doesn’t seem to help.
?
Bill
On Sep 28, 2024, at 9:22?AM, Paul Kanevsky <yh@...> wrote:
Hi Bill, it looks like to me that the autoguider isn't calibrated properly. Corrections are often insufficient in RA, and sometimes overshoot a quick jump. The first step is to measure your periodic error during tracking (capture a few cycles of tracking error while not guiding). See what tracking looks like without autoguider corrections. Then, a guide log may help diagnose what's going on while guiding is engaged. PHD2 also has a number of diagnostic/calibration tools that may help figure out what's going on.
?
Regards,
?
?-Paul
?
--?
Bill Tschumy
Otherwise -- Longmont, CO
www.otherwise.com
?
|
Paul,
I should be able to get an unguided log tonight.
Yes, Gemini 1. ?The autoguider is connected to the main camera, which in turn is connected to a Raspberry Pi 4. ?The RPi4 also has a USB cable coming out that goes into the RS232 port on the Gemini 1 (via a USB to serial converter). ?Pretty sure I’ve tried the autoguider plugged directly into the RPi4, but I can try that again to see if if there are any differences.
Bill On Sep 28, 2024, at 3:09?PM, Paul Kanevsky <yh@...> wrote:
So the calibration looks good. How about an unguided tracking log over a few worm cycles?
?
This is with Gemini-1, correct? How is your autoguider connected to Gemini?
?
Regards,
?
? -Paul
On Sat, Sep 28, 2024 at 12:44 PM, Bill Tschumy wrote:
Paul,
?
I have calibrated and re-calibrated many, many times. ?I have also tried a variety of different pulse lengths for calibration. ?If calibration is off then there must be some other fundamental problem that is causing it. ?I’ve attached the Calibration Plot that Ekos generates. ?It doesn’t look particularly bad to me.
?
I do use a somewhat longer pulse length (around 1250 ms) because if I do shorter the Dec calibration looks worse.
?
?
?
Sometimes the Dec calibration dots can be a bit squirrely although the ones here look fine to me.
?
I should also mention that I’ve tried adjusting the RA worm play ad nauseum and that doesn’t seem to help.
?
Bill
On Sep 28, 2024, at 9:22?AM, Paul Kanevsky <yh@...> wrote:
Hi Bill, it looks like to me that the autoguider isn't calibrated properly. Corrections are often insufficient in RA, and sometimes overshoot a quick jump. The first step is to measure your periodic error during tracking (capture a few cycles of tracking error while not guiding). See what tracking looks like without autoguider corrections. Then, a guide log may help diagnose what's going on while guiding is engaged. PHD2 also has a number of diagnostic/calibration tools that may help figure out what's going on.
?
Regards,
?
?-Paul
?
--?
Bill Tschumy
Otherwise -- Longmont, CO
www.otherwise.com
?
--? Bill Tschumy Otherwise -- Longmont, CO www.otherwise.com
|
Paul (and others),
I’ve attached a screenshot showing the unguided results over about 18 minutes. ?You can sort of see the periodic error which I think is 478 seconds (about 8 minutes) for the GM8. ?I’m not sure what this tells us, but hopefully it will be useful.
I also noticed a curiosity last night. ?Ekos has a “Plate solve and center target” function. ?It will automatically repeat this until the error is less than some amount (30 arcsec in my case). ?Lately it seems to be having more trouble zeroing in on the target. ?Here is a screenshot of what it did. ?Rows 8 through 18 are its trying to zero in. ?Note that the declination zeros in nicely. ?However, the RA seems to oscillate by overshooting the center each time (assuming I’m reading this correctly). ?It oscillates between a positive error and a negative error.
It seem kind of surprising it would be overshooting in this case.
Back to the guided results I posted initially, it sort of feels like something is slipping. ?It tries to make a correction, but doesn’t get there (or possibly goes too far). ?The most likely thing I know of that could slip is the connection of the motor shaft and the worm shaft to the Oldham coupler. ?However, I’ve verified several times that the set screws are tight. ?I don’t see how that connection could be slipping. ? Is there any place else in the system where slippage could occur?
Bill
toggle quoted message
Show quoted text
On Sep 28, 2024, at 3:41?PM, Bill Tschumy <bill@...> wrote:
Paul,
I should be able to get an unguided log tonight.
Yes, Gemini 1. ?The autoguider is connected to the main camera, which in turn is connected to a Raspberry Pi 4. ?The RPi4 also has a USB cable coming out that goes into the RS232 port on the Gemini 1 (via a USB to serial converter). ?Pretty sure I’ve tried the autoguider plugged directly into the RPi4, but I can try that again to see if if there are any differences.
--? Bill Tschumy Otherwise -- Longmont, CO www.otherwise.com
|
It's nearly impossible to read the PE chart due to small scale. What I was interested in seeing was any sudden jumps, and especially those that don't repeat with the worm cycle (which requires more than one worm cycle to be recorded). PE looks OK from a very high level, but I really can't see the details in the image to be sure.
?
As to what else can be slipping:
?
1. Oldham coupler, as you already checked
2. Clutches
3. Encoder ring inside the servo (you can try swapping DEC and RA servos to see if the problem goes over to DEC)
?
Additionally, poor connection to the encoder(s), noisy or intermittent cable connection, or encoder damage / dirt on the encoder optical ring can all cause issues that look like something is slipping. I had one servo that was misbehaving. Turns out a small spider crawled through the external connector and put up some spiderwebs inside the motor that caused all kinds of strange problems :) Dirt or damage in the gearboxes can also cause erratic behavior.
?
For most of these, try swapping the components between RA and DEC to see what makes the problem switch to the other axis.
?
Regards,
?
? -Paul
?
On Sun, Sep 29, 2024 at 12:02 PM, Bill Tschumy wrote:
Paul (and others),
?
I’ve attached a screenshot showing the unguided results over about 18 minutes. ?You can sort of see the periodic error which I think is 478 seconds (about 8 minutes) for the GM8. ?I’m not sure what this tells us, but hopefully it will be useful.
?
I also noticed a curiosity last night. ?Ekos has a “Plate solve and center target” function. ?It will automatically repeat this until the error is less than some amount (30 arcsec in my case). ?Lately it seems to be having more trouble zeroing in on the target. ?Here is a screenshot of what it did. ?Rows 8 through 18 are its trying to zero in. ?Note that the declination zeros in nicely. ?However, the RA seems to oscillate by overshooting the center each time (assuming I’m reading this correctly). ?It oscillates between a positive error and a negative error.
?
?
It seem kind of surprising it would be overshooting in this case.
?
Back to the guided results I posted initially, it sort of feels like something is slipping. ?It tries to make a correction, but doesn’t get there (or possibly goes too far). ?The most likely thing I know of that could slip is the connection of the motor shaft and the worm shaft to the Oldham coupler. ?However, I’ve verified several times that the set screws are tight. ?I don’t see how that connection could be slipping. ? Is there any place else in the system where slippage could occur?
?
Bill
On Sep 28, 2024, at 3:41?PM, Bill Tschumy <bill@...> wrote:
Paul,
?
I should be able to get an unguided log tonight.
?
Yes, Gemini 1. ?The autoguider is connected to the main camera, which in turn is connected to a Raspberry Pi 4. ?The RPi4 also has a USB cable coming out that goes into the RS232 port on the Gemini 1 (via a USB to serial converter). ?Pretty sure I’ve tried the autoguider plugged directly into the RPi4, but I can try that again to see if if there are any differences.
?
--?
Bill Tschumy
Otherwise -- Longmont, CO
www.otherwise.com
?
|
I resent the image to your email, bypassing Google. ?Hopefully that will be readable.
As to the slipping.
1. Yes, I have checked the coupler 2. Clutches are tightish. ?I can make them move with come moderate pressure on them, but I can see how guiding changes could do it. ?I have previously tried this with clutches really tight and really loose. ?No real change. 3. As I mentioned in my original message, I have trid swapping the motors and well as the gearboxes (and as well the worms). ?No real difference.
It is a mystery.
On Sep 29, 2024, at 10:19?AM, Paul Kanevsky <yh@...> wrote:
It's nearly impossible to read the PE chart due to small scale. What I was interested in seeing was any sudden jumps, and especially those that don't repeat with the worm cycle (which requires more than one worm cycle to be recorded). PE looks OK from a very high level, but I really can't see the details in the image to be sure.
?
As to what else can be slipping:
?
1. Oldham coupler, as you already checked
2. Clutches
3. Encoder ring inside the servo (you can try swapping DEC and RA servos to see if the problem goes over to DEC)
?
Additionally, poor connection to the encoder(s), noisy or intermittent cable connection, or encoder damage / dirt on the encoder optical ring can all cause issues that look like something is slipping. I had one servo that was misbehaving. Turns out a small spider crawled through the external connector and put up some spiderwebs inside the motor that caused all kinds of strange problems :) Dirt or damage in the gearboxes can also cause erratic behavior.
?
For most of these, try swapping the components between RA and DEC to see what makes the problem switch to the other axis.
?
Regards,
?
? -Paul
?
On Sun, Sep 29, 2024 at 12:02 PM, Bill Tschumy wrote:
Paul (and others),
?
I’ve attached a screenshot showing the unguided results over about 18 minutes. ?You can sort of see the periodic error which I think is 478 seconds (about 8 minutes) for the GM8. ?I’m not sure what this tells us, but hopefully it will be useful.
?
I also noticed a curiosity last night. ?Ekos has a “Plate solve and center target” function. ?It will automatically repeat this until the error is less than some amount (30 arcsec in my case). ?Lately it seems to be having more trouble zeroing in on the target. ?Here is a screenshot of what it did. ?Rows 8 through 18 are its trying to zero in. ?Note that the declination zeros in nicely. ?However, the RA seems to oscillate by overshooting the center each time (assuming I’m reading this correctly). ?It oscillates between a positive error and a negative error.
?
?
It seem kind of surprising it would be overshooting in this case.
?
Back to the guided results I posted initially, it sort of feels like something is slipping. ?It tries to make a correction, but doesn’t get there (or possibly goes too far). ?The most likely thing I know of that could slip is the connection of the motor shaft and the worm shaft to the Oldham coupler. ?However, I’ve verified several times that the set screws are tight. ?I don’t see how that connection could be slipping. ? Is there any place else in the system where slippage could occur?
?
Bill
On Sep 28, 2024, at 3:41?PM, Bill Tschumy <bill@...> wrote:
Paul,
?
I should be able to get an unguided log tonight.
?
Yes, Gemini 1. ?The autoguider is connected to the main camera, which in turn is connected to a Raspberry Pi 4. ?The RPi4 also has a USB cable coming out that goes into the RS232 port on the Gemini 1 (via a USB to serial converter). ?Pretty sure I’ve tried the autoguider plugged directly into the RPi4, but I can try that again to see if if there are any differences.
?
--?
Bill Tschumy
Otherwise -- Longmont, CO
www.otherwise.com
?
--? Bill Tschumy Otherwise -- Longmont, CO www.otherwise.com
|
Thanks, Bill. This one I can read :)?
?
PE is kind of high, around 20 arcseconds peak-to-peak. Most Losmandy mounts I’ve seen have PE of around 10 arcsecs p2p, often less.?
?
What concerns me are the large, fast jumps of many arcseconds at a time. These are hard to guide out, as the error grows quickly and autoguider often can’t keep up. You could try increasing guide speed to 0.8x to see if that might help (recalibrate the guider, of course, once you do it). It may also be worth swapping the worms between DEC and RA to see how much better the PE might be with the other worm. I had a G11 where my PE was around 10 arcsecs p2p, but once I swapped worms, it went down to 4 arcsecs. Amazing improvement and a fairly quick and free fix :)
?
Regards,
?
? -Paul
?
On Sun, Sep 29, 2024 at 12:39 PM, Bill Tschumy wrote:
I resent the image to your email, bypassing Google. ?Hopefully that will be readable.
?
As to the slipping.
?
1. Yes, I have checked the coupler
2. Clutches are tightish. ?I can make them move with come moderate pressure on them, but I can see how guiding changes could do it. ?I have previously tried this with clutches really tight and really loose. ?No real change.
3. As I mentioned in my original message, I have trid swapping the motors and well as the gearboxes (and as well the worms). ?No real difference.
?
It is a mystery.
?
On Sep 29, 2024, at 10:19?AM, Paul Kanevsky <yh@...> wrote:
It's nearly impossible to read the PE chart due to small scale. What I was interested in seeing was any sudden jumps, and especially those that don't repeat with the worm cycle (which requires more than one worm cycle to be recorded). PE looks OK from a very high level, but I really can't see the details in the image to be sure.
?
As to what else can be slipping:
?
1. Oldham coupler, as you already checked
2. Clutches
3. Encoder ring inside the servo (you can try swapping DEC and RA servos to see if the problem goes over to DEC)
?
Additionally, poor connection to the encoder(s), noisy or intermittent cable connection, or encoder damage / dirt on the encoder optical ring can all cause issues that look like something is slipping. I had one servo that was misbehaving. Turns out a small spider crawled through the external connector and put up some spiderwebs inside the motor that caused all kinds of strange problems :) Dirt or damage in the gearboxes can also cause erratic behavior.
?
For most of these, try swapping the components between RA and DEC to see what makes the problem switch to the other axis.
?
Regards,
?
? -Paul
?
|
Bill Tschumy
Paul Kanevsky