开云体育

I have been trying to find out the sub-second dynamics of my G11S with OnStep for which a guide error log is sampling too slow.? The only way I could think of was to create star trails using the hand controller and analyze them.? Speed variations result in variations of light intensity.? This can easily be inverted and integrated to retrieve the deviations from the nominal trajectory.? Its power spectrum then tells the oscillations of which some may be attributed to the stepper motor.?

For a focal length of 755 mm (ES127CF+FF) and a pixel size of 3.8 mu (ASI2600) at nominal RA tracking speed (relative to the sky) the stars move at 14.6 pixels/s so we can analyze up to about 7 Hz.? I took some star trails at various speeds, rotated them vertically in Gimp then imported them in Scilab for analysis.? I used an order 10 polynomial trend removal to account for incident angle, dust motes and other low frequent disturbances, and the median for background removal.? The DEC position is done by a weighted average of pixel values.? This is pretty crude in some ways, but I don't think it is complete nonsense.

The RA 1x relative speed is shown below.? The 700x100 pixel image on the left is the star trail (DEC is horizontal, RA vertical).? The center column has the DEC (X) and RA (Y) errors relative to the nominal trajectory.? The RA plot also shows the 10-th order polynomial trend-removed version.? Their power spectra are to the right.? DEC shows errors tapering down at 1 Hz, RA tapers down at 0.5 Hz.? Since RA is moving at 2x tracking speed, maybe this translates to errors up to 0.25 Hz at tracking speed, if it is due to the stepper motor.? Take it FWIW.

About the number of microsteps, at 64 microsteps my Nema17 micro-steps at 107 Hz.? I don't know if it makes a lot of difference with 32 microsteps at 53 Hz or 16 microsteps at 27 Hz.? The mechanical telescope system limits the bandwidth quite a bit, so any errors above 8 Hz or so get smoothed out quite a bit.?

I am interested in stepper motor related errors, which is why I did this.? Based on the above (if it can be believed), the highest frequency peak at tracking speed is below 0.25 Hz, and the magnitude looks minor compared to the rest.? Just my 2 cents.

Enter in the hardware tab and the Gemini.Net should be available in the roll down selection per below. The Properties button allows you to set what interface you are using to the Gemini.Net setting and I use the Ethernet setting. If you are using the ethernet connection then SharpCap does not have to be on the same PC that the Gemini.Net driver is on.

For Stellarium and others you may want to see if it supports ASCOM's Alpaca server so that you can also run the Stellarium on another system to the Gemini.Net driver. Alpaca is a newish server/client model for ASCOM that does not require .Net (i.e. install Alpaca Server on the PC with the Gemini.Net driver and as long as the Stellarium has the Alpaca client then it can communicate to the Server syste with Gemini.Net - see

Ken

Thanks Ken. This is very helpful!

Question... in the sharpcap settings page, any idea which tabs I would need to enter info into? Maybe the hardware tab? I can look at each tab if/when I get the mount; just curious. Also, I think the mount needs to be connected to the PC for it to show up in sharpcap, correct?

John,

I use the ASCOM Gemini Telecope.Net driver - (Under the Ascom section)
??????? Video here -

SharpCap just needs the ASCOM Gemini Telescope.net installed and you can set it up in the Settings area -
Stellarium - /g/Gemini-II/topic/trying_to_control_gemini_2/82816694?p=

Good luck,
Ken

Hi all,
I just joined the group. I am considering buying the G11G mount, to be controlled by my PC via a USB cable (I don't use asi air). I use Sharpcap and Stellarium.? Can someone please indicate the correct PC control software I would need to download (Gemini II)? And are there instructions on how to use this software for mount control?? Is it compatible with Sharpcap and Stellarium, and how well do they work together- any issues??

If there are any introductory tutorial videos on PC control of the G11G, using sharpcap and stellarium, can you send a link, or let me know where I can find one??

Thanks!

On Sun, Dec 31, 2023 at 04:52 PM, Michael A. Covington wrote:

On 2023-12-31 18:06, Brendan wrote:

You can increase torque by using a larger to smaller belt hub, which
is the same as any large to small gear ratio.
--
Brendan

Yes, but it remains the case that torque is not the same across the
entire microstep cycle, leading to slight speeding up and slowing down.?
We're not talking about lacking the torque to move the load.

Michael A. Covington, Ph.D.
Consultant, Covington Innovations, Athens, Georgia, USA

Micheal,

There is a fair amount of starting friction that has to be overcome on top of the normal load forces. When the mount is moving steady state the standing friction is lower overall. When you look at the available.torque for a micro stepped motor it is easy to see why it is desirable to start a static load using full steps or at least much lower micro step values to get the drivetrain moving. I expect that the static load never actually goes away completely with stepper motors drivetrains because their motion is not actually continuous like a servomotor driven by a small DC motor.?

The reason Losmandy can use such small DC servomotors is the high reduction in the Losmandy drivetrain. So stepper motors need to be run with lower microstep values and still need significant reduction to work smoothly. For the G11 I think 5:1 is the minimum for a stepper motor drive. The standard gearbox for the Gemini servomotors is 25:1 but this would yield a very slow slew for a stepper motor driven G11 with most of the OnStep controller boards because they don't have fast enough clocks to generate the required step pulses.?

--

Chip Louie Chief Daydreamer Imagination Hardware?

? ?Astrospheric Weather Forecast - South Pasadena, CA?

On Sun, Dec 31, 2023 at 12:26 PM, Michael A. Covington wrote:

Microstepped stepper motors do not have equal torque at all microsteps.?

And Michael's description becomes likely one of a motor that lost position during micro-steps while synchronizing position as the motor passes the next pole position (full torque).? Requiring either a larger stepper motor or mechanical transmission reduction (belts or gears).? ?Allowing more torque, more resolution, or a mixture of torque and resolution.

It should have a rough PHD2 graph with moments of hope.? ?It's very interesting Michael could see the issue visually.?

Thanks Michael and Chip,
Doug

This is from an old Hackaday article about microstepping stepper motors. It is old but the rules and torque losses are immutable laws of the universe and explain the problem using very high microstep values. You can easily validate this information using a stepper tester. If we had an imaginary 0.9 degree (400 step)? 100 oz/in NEMA 17 stepper even at only 32 microsteps the stepper could at best only provide 5 oz/in of torque. At 256 microsteps less than 1 oz/in torque. As you can see this is not very reliable. A high torque standard NEMA 17 size frame normally delivers 59 oz/in torque. Obviously some mechanical reduction is in order as well as some mode switching and stepper speed ramping.? ?


--

Chip Louie Chief Daydreamer Imagination Hardware?

? ?Astrospheric Weather Forecast - South Pasadena, CA?

On Wed, Jan 3, 2024 at 12:36 PM, @darkairmouse wrote:

Hi Chip,
Understood.
Despite the not-so-good RA guiding the accuracy of the pointing was very good.?

Yue

So getting good GOTO results I think tells us that the mechanical basics are solid and all you really need to do is fine tune your autoguiding. Leave the mount hardware alone for now and look at the what your autoguiding corrections are doing. Are you autoguiding using a OAL or piggybacked guide scope? If you are looking at images from an SCT and using a piggybacked guide scope what is the working focal length of the SCT and guide scope??

Some years ago I tested autoguiding results using both an OAG and a piggybacked scope on my pre-update classic 2-piece worm block G11 which had the Belleville spring disc mod installed. I had solid results at prime focus of my 10" f/6.3 LX200 SCT both piggybacked and using the OAG with the same QHY 5-II mono camera. Auto guided PE was well under 0.5" RMS - within 0.46" - 0.38" PE RMS, this is about at good as this mount gets. I had the chance to retest after Scott did the tucked in servo motors and spring loaded OPWs and the results were virtually identical on the now all stock new spec drivetrain. The difference is that the reworked mount is now ALWAYS in this range or better if the seeing is good.?

I cannot repeat the exact same setup using the latest L6 firmware because several years ago I switched to ZWO cameras and I am now using the ASIAIR Plus controller in place of the laptop and full suite of ASCOM apps like PHD2. That said the ASIAIR did show a slight autoguiding improvement over the normally slightly lower performing simplified ASIAIR PHD2 autoguiding.?
?
--

Chip Louie Chief Daydreamer Imagination Hardware?

? ?Astrospheric Weather Forecast - South Pasadena, CA?

Hi Chip,
Understood.
Despite the not-so-good RA guiding the accuracy of the pointing was very good.?

Yue

On Wed, Jan 3, 2024 at 05:59 AM, @darkairmouse wrote:

Ok finally got a chance to test my 811G last night, and honestly, not as I expected. Not as bad as 'disappointed' but not as what I 'expected'.
The DEC turned out smooth, no issues at all however the RA..... jumping up and down on PHD2 graph, worse than my eq6-r pro.
I did everything was as usual i.e. I did what I'd do if that's my eq6-r pro and PHD2 didn't report any issue/problem after calibrating but still the RA curve showed ugly and I can see that from the final 20-min sub.
Anyways, now I adjusted the RA adding very little backlash, waiting for next clear night to see if this helps.

Yue

Yue,

The Losmandy GM811 is not an EQ6 and you need to do things a bit differently to find the mount's sweet spot. Every Losmamdy mount is capable of a slightly different different minimum PE value. The most important things to check are related to constraint of the worm between the bearing blocks and the alignment of the gearbox and worm before the Oldham grubs are glued and tightened. After that it is all about spring force to push the worm into the worm wheel and the worm lift bolt adjustments.?
?
--

Chip Louie Chief Daydreamer Imagination Hardware?

? ?Astrospheric Weather Forecast - South Pasadena, CA?

“Nudge mount out of safety limit on slew” which is set in ASCOM driver dialogs, not in Gemini screens.
--
Edward

Ok finally got a chance to test my 811G last night, and honestly, not as I expected. Not as bad as 'disappointed' but not as what I 'expected'.
The DEC turned out smooth, no issues at all however the RA..... jumping up and down on PHD2 graph, worse than my eq6-r pro.
I did everything was as usual i.e. I did what I'd do if that's my eq6-r pro and PHD2 didn't report any issue/problem after calibrating but still the RA curve showed ugly and I can see that from the final 20-min sub.
Anyways, now I adjusted the RA adding very little backlash, waiting for next clear night to see if this helps.

Yue

On Mon, Jan 1, 2024 at 07:29 AM, WayBack wrote:

Michael,

Micro step position does vary but I'm not sure it represents a significant issue; this not accumulating during guiding.? The issue appearing like corrected backlash.? And it cannot accumulate in physical position between motor poles (true steps).

I always felt power consumption and the related issue of producing less torque for the power were my issues.? The benefit of being open loop becomes significant only if the weight and power allow for this.

With torque levels at tracking quite small, as you illustrate here, the variation cannot be large, or it shows up in guiding.? I've never seen anyone ask David Partridge?but the RMP of our motor during tracking must be quite low and the ability of our motor to arrive exactly on count at a moment in time also a variant.? Still there too, all just esoteric when compared to the tracking RMS achieved.? Hence the importance placed on achieving Level 6 functionality.

Doug? ?

I did not write the firmware or design these systems but my understanding of the way these system work is that they are quite different.?The way an open loop stepper drive system works vs the closed loop of the Gemini 2 servo motor drive is quite different.

A stepper motor controller like the OnStep (not the stepper driver IC which can be quite sophisticated) blindly sends the step or microstep pulses calculated and has no idea if the stepper motor has actually completed the full step count or lost steps. This is an inherent problem with an open loop stepper drive system especially a marginally powered system for the load. This means that an open loop stepper drive system cannot ever know the mount axes position with certainty. Encoders on the drive motor itself or the movement axes allows the mount controller to know with reasonable certainty the mount's position. Certainly there are encoder equipped stepper motors but they are not being exploited widely yet.?

So open loop stepper motor drive mounts using a worm screw can basically only successfully step and move the mount or lose steps and physical position but the mount controller's model is not aware and has to assume the mount? position is correct which puts the slew off target requiring the mount be recentered. Depending on the how many steps were lost and the application it may or may not matter.?

The closed loop system with encoders on the DC motor drive shaft as used in the Gemini 2 allow the controller to just run the motors and watch the encoder count slowing, then stopping or assuming the selected movement rate for tracking. Positional data for the mount is generally? known and the mount position can be corrected in the case of an under or over shoot of the desired tic count. The Gemini sends a known amount of current to the motors and watches the resulting movement through the tic counts. If the tic count is not accumulating as predicted vs time more or less current is sent as required to match the desired movement rate.?

With the rewritten PIC code and revised and tuned L6 primary controller firmware the Gemini 2 L6 firmware now has a resolution of 0.14 arcseconds. This is getting into the pretty good range IME and is 4x better resolution of the direct stepper drive G11 mount. But if you added a 4:1 reductio with a pulley you would have very similar theoretical resolution though it would still only be a guess being open loop. Adding encoders on the stepper to close the loop and correct any missed microsteps and now you have something.? ?

--

Chip Louie Chief Daydreamer Imagination Hardware?

? ?Astrospheric Weather Forecast - South Pasadena, CA?

Yep, safety limits will also do that, and I didn’t think of that option either so, double doh! Lol

You have a couple of options to take the mount off the safety limits. You can use the hand controller to manually nudge the mount away from the limit (with the diamond buttons, in the right direction away from the limit). Or you can enable an option that I don’t remember off the top of my head where it is, so a park command will nudge the mount away from the limit before issuing the command itself. I believe that option is in the gemini.net driver but may also be available on the HC. Once that option is enabled NINA should be able to park the mount, even if it’s at the safety limits.

Doh! Thanks, Jonathan. I didn't even look at the HC this morning. I just went out and checked and the mount had reached a safety limit.

Anyone,

I'm guessing here but once that limit is reached, does it completely lock out control via software until reset by the HC (i.e. is this why I couldn't park from NINA)?

Thanks,
Joel