开云体育

  1. Losmandy_users
  2. Topics

Got another Losmandy

 

I picked up a G11G on Cloudy Nights from 2015 with HT motors (non-tucked-in), HD folding tripod, Gemini-2 level 5 and polar scope, no SLW, for $2100.? The mount looks like new, it was well taken care of by the previous owner.? I upgraded the Gemini-2 to level 6, got a PicKit3, flashed the motor controllers with v3.0 and checked the 4x checkboxes.? I just tried it with Ekos, everything works beautifully.? ?I'll try to get a first light tonight.
?
The previous owner loved the mount, but he had moved to a place where he had no good space to set up a telescope, so it didn't get used.? I picked it up from Arizona with a detour to visit an old friend, almost 23 hours of driving without air conditioning but well worth the drive.
?
My current mount is a G11S from 2018 with OnStep.? One reason for getting the new one is to compare a standard Losmandy with servo motors with OnStep that uses steppers.? Comparing the price of Gemini II with motors with the $2100 that I paid, I think this is an excellent deal.? I can always sell it easily should I want to.?
?
Another reason for getting it is to have a more "mobile" mount.? I have an observatory that is a converted lean-to shed against a wall next to the kitchen area from where I can only see the Southern skies.? I noticed that I could see most of the sky including Polaris from the patio in front of my kitchen window.? Since I don't like to move the mount in my observatory to that spot and back all the time, it is nice to have another one.? It is also handy for star parties.? I got a Telegizmo cover to keep it set up outside for New-Moon weeks.
?
 

First light was last night.? I replaced the battery of the Gemini after seeing many of low battery warnings - it was about 0.3V instead of 3V.? I used the polar scope for polar alignment.?
?
First problem I noticed, after doing some gotos and parking back to CWD, was periodic messages on the hand controller.? Every 3 seconds there was a log message that the goto had finished and a stop sign was displayed briefly each time.? Turns out, I had set automatic meridian flip in Ekos that generated an attempt to do flips in this situation; it would not happen away from CWD.? Not a big deal, I turned it off and can turn it back on once on target.? Or leave it on but avoid parking to CWD for no good reason.
?
Then, the gotos were not working.? After checking all parameters and verifying they were all fine, I restarted without Ekos just using Gemini and the gotos worked fine.? The problem was that I have to start and initialize the mount before starting Ekos.? I had noticed something similar before with my AVX running under Ekos.? Somehow with OnStep I can get away with it.? After restarting staged that way, the gotos worked fine with Ekos and I was in business.? Next time I will double-check that this is indeed the cause because I don't quite understand why starting Ekos and Gemini simultaneously would cause issues.
?
The autoguider calibration curves looked very good with even zero to no backlash in DEC - I had tightened it.? Using M31 as a target, guiding error was 1" to 1.5" total RMS, about the same as what I get from OnStep and much larger than I like.? I tried all kinds of autoguider control settings but there was little difference compared to the defaults.? When shutting down, I noticed that I forgot to tighten the DEC clutch. The gotos were fine nevertheless, so I don't think it mattered much; in fact, the error was often less in DEC than in RA.? Clearly, I have some work to do and examine some logs.? Thank goodness for BlurXTerminator, it handles high RMS values quite well!
 

I took a look at the error spectra from the data in the log file.? There were 3 periods in the session with different guiding parameters:? camera exposure = 2,1,1 and camera delay = 1,0,1.? This resulted in average sampling intervals of 4.8, 2.9 and 3.6 seconds.? These intervals include the camera exposure + processing, pulse duration and delay for the next exposure after the control action.? The pulse duration average is mostly below 300 ms, so the sampling interval is reasonably constant.? The RA error spectra look like this:
?
Each of them shows the 76 second worm bearing peak and the 32 second McLennan gearbox peak, more so with the (2,1) and (1,0) cases than with the (1,1) case.? So, an exposure of 1 second with camera delay of 1 second works best to guide out the periodic errors, although of course the samples were taken in different positions, so who knows.? Altogether it doesn't look too bad compared to the off-peak errors.? So maybe it's a matter of atmospheric conditions, polar alignment, camera noise and guiding parameters.? I have not yet touched the RA worm blocks but I did for DEC, and DEC showed no special peaks.? So, I think I can fix the worm bearing peak by re-aligning the blocks.? Not much I can do about the gearbox periodic error, I think.?
?
The test image for the above spectra is below, just another M31, for the record (53x2 min with an ASI2600MC, ED127CF, flats, biases, darks).? Note that BlurXTerminator was used to correct some flaws and sharpen the stars - the focuser was a bit skewed and caused oval star shapes on the right side.? Altogether I think this system will work well.? In terms of RMS, I usually get the same numbers with my G11S/OnStep.
?
 

M31 looks great, Henk!?

What is the vertical axis scale in the spectrum plot? Arcseconds? If yes, that gearbox error seems excessive. Might be worth trying to swap gearboxes.
?
You could likely fix the 32sec error if you program Gemini PEC function with 2x worm cycle.
?
I believe Ekos has an adaptive PEC function, similar to PHD2's PPEC, no? Do you use it?
?
Regards,
?
? -Paul
 

Thanks Paul,
?
I scaled the FFTs such that the RMS of the data is the average of the spectrum.? This is done by dividing the FFTs (in Scilab -- probably the same as Matlab) by the square root of the number of data points.? So yes, it corresponds with RMS in arc seconds, and the expectation values of the spectrum do not depend on the number of data points.
?
I was using Ekos' adaptive PEC that is indeed identical to PHD2's PPEC.? I have it turned on with a period of about 4 minutes (sidereal calculations).
?
I will try swapping the gear boxes but the screws feel awfully tight, and I have to be careful not to damage the screws.? Thanks for the hint.
 

Paul, about the peak of the gearbox that is almost 6 arc seconds in the first graph:? This does not mean that the time domain data oscillates with that amplitude.? The proper way to look at it is to integrate the power over the peak, divide it by the total power to get the fraction of RMS that the gearbox is responsible for.?
?
To get an idea of what it looks like in reality (using a raw FFT spectral analysis, which is a bit crude), I zeroed out the off-peak Fourier components and inverse-FFTed it, the plot looks like this:
This is definitely significant, but it clearly does not have the 6 arc second spectral peak value for an amplitude.? Doing the same for the 76-second peak, I get this:
The RMS fraction of the 32-second peak is 0.4 and that of the 76-second peak is 0.22.? ?These values do not change much for the 3 parameter settings.? The RMS value is about 0.95 for all scenarios, so that is also not affected much by the parameter settings.? This was for RA.?
?
I noticed that for DEC that the RMS values are 0.77, 0.9 and 1.1.? This could mean that DEC is affected by the parameter settings, or that the atmospheric and other environmental conditions changed.??
?
Lastly, I mentioned the McLennan gearbox.? ?This is a mistake, I have the Losmandy gearbox.
 

Ok, that makes more sense. In that case, the errors are not large at all. Time domain peak errors are important in figuring out guiding/tracking issues. RMS values and averages are valuable, but when looking for specific faults, you would want to see the large and fast excursions, as these are harder to correct, to PEC out, and to guide out.
?
Regards,
?
? -Paul
?
On Thu, Oct 3, 2024 at 12:59 PM, HenkSB wrote:

Paul, about the peak of the gearbox that is almost 6 arc seconds in the first graph:? This does not mean that the time domain data oscillates with that amplitude.? The proper way to look at it is to integrate the power over the peak, divide it by the total power to get the fraction of RMS that the gearbox is responsible for.?
?
To get an idea of what it looks like in reality (using a raw FFT spectral analysis, which is a bit crude), I zeroed out the off-peak Fourier components and inverse-FFTed it, the plot looks like this:
This is definitely significant, but it clearly does not have the 6 arc second spectral peak value for an amplitude.? Doing the same for the 76-second peak, I get this:
The RMS fraction of the 32-second peak is 0.4 and that of the 76-second peak is 0.22.? ?These values do not change much for the 3 parameter settings.? The RMS value is about 0.95 for all scenarios, so that is also not affected much by the parameter settings.? This was for RA.?
?
I noticed that for DEC that the RMS values are 0.77, 0.9 and 1.1.? This could mean that DEC is affected by the parameter settings, or that the atmospheric and other environmental conditions changed.??
?
Lastly, I mentioned the McLennan gearbox.? ?This is a mistake, I have the Losmandy gearbox.

1 - 7 of 7