开云体育

Gemini 2 response to pulses that are smaller than minimum

Edward Beshore #71777 Hi - In another thread, Paul Kanevsky said "The minimum duration will be based on minimum step size of the mount. Step sizes are: 0.375 arcsec/step for Titan,? 0.5625 for G11, and 1.13 for GM8. ? "If the guiding rate is 0.5x, then the movement for 50ms pulse is 0.375 arcsecs. A Titan can move that much, but for a G11 or GM8 the step is too large. For a G11 at 0.5x rate the minimum pulse will be 75ms and for GM8 - 150ms My question is, how will Gemini respond when it gets a pulse duration that is shorter than the minimum movement for a given gear/servo combination? Will it? ? a) ignore the pulse? b) issue the minimum pulse? ? Likewise, will all longer pulses be "rounded up" to the shortest combination of steps corresponding to the requested pulse duration or truncated down? ? Reading this thread made me go back and look at the guide logs for my GM811 and I am wondering if I am seeing the result of ?a "quantization" error that arises from the issuance of small pulses.
Stephen Migol All Messages By This Member #71783 There is the additional question of what happens with the guiding software.? If it sees that the shorter pulse change hasn't been made, it will make the request for movement based on the next measurement.?? If the hardware and the software will queue and sum pulses, then there's a potential for overcorrection. I think that the hardware will ignore the shorter than minimum pulse. Stephen
Paul Kanevsky All Messages By This Member #71784 There's no queuing of pulse guide commands. If a request is too small to result in movement, the mount will ignore it. Should the error continue to grow and exceed the minimum correction size, the pulse guide command will be processed and the position corrected. In effect, this acts as a low-pass filter, ignoring smaller, quick corrections which may sometimes be a good thing (ignoring seeing effects or small worm imperfections / dirt, for example).? Regards, ? ? -Paul toggle quoted message Show quoted text On Thu, May 20, 2021 at 01:22 PM, Stephen Migol wrote: There is the additional question of what happens with the guiding software.? If it sees that the shorter pulse change hasn't been made, it will make the request for movement based on the next measurement.?? If the hardware and the software will queue and sum pulses, then there's a potential for overcorrection. I think that the hardware will ignore the shorter than minimum pulse. Stephen
Brian Valente Member Profile All Messages By This Member #71785 >>> I think that the hardware will ignore the shorter than minimum pulse. Yes, and Paul did a good job of summing it up One other thing to keep in mind is if your guiding is issuing these tiny pulses that are below the Gemini threshold, chances are your min moves are too small. you are almost certainly chasing seeing?at that point toggle quoted message Show quoted text On Thu, May 20, 2021 at 10:36 AM Paul Kanevsky <yh@...> wrote: There's no queuing of pulse guide commands. If a request is too small to result in movement, the mount will ignore it. Should the error continue to grow and exceed the minimum correction size, the pulse guide command will be processed and the position corrected. In effect, this acts as a low-pass filter, ignoring smaller, quick corrections which may sometimes be a good thing (ignoring seeing effects or small worm imperfections / dirt, for example).? Regards, ? ? -Paul On Thu, May 20, 2021 at 01:22 PM, Stephen Migol wrote: There is the additional question of what happens with the guiding software.? If it sees that the shorter pulse change hasn't been made, it will make the request for movement based on the next measurement.?? If the hardware and the software will queue and sum pulses, then there's a potential for overcorrection. I think that the hardware will ignore the shorter than minimum pulse. Stephen -- Brian? Brian Valente portfolio
Edward Beshore #71819 Hi Brian, I was mostly concerned about the Dec axis with my GM811. The default minmov for PhD2 in Dec is 0.2 pix. I am using my 480mm Pronto (206.265/480 = 0.43"/μ?plate scale) and a Lodestar 2 (8.6?μ?pixels) for my guidescope setup which yields 8.6 μ * 0.43"/μ?= 3.7"/pix. With a 0.2 pix min move that is 0.2 * 3.7" = 0.74" in Dec. Seems like I should be setting my Dec min move to ~1.15 " / 3.7 "/pix = ?0.32 pix? Given Paul's response, it sounds like the Dec move commands would be getting ignored until the error gets past the minimum step. In the final analysis, that may be fine, since the error will eventually move past that and a correction will make it through but seems a bit pointless to set a minmov to a value that the mount can't accommodate. What do you think? Cheers, Ed
Michael Herman All Messages By This Member #71820 Ed, Two things I suggest you evaluate: You have the option of using ASCOM pulse guiding (sending commands to the Gemini to move) or using the ST4 port (the autoguide camera sends electrical signals to the Gemini). I've always read they work identically, but I think it's worth it for you to do a comparison.? I tried both mothods and found good results from either way.? But since you are experiencing some problems perhaps one method is superior and minimizes your issue. The other factor you might try is this: eliminate the SLW movement if your DEC uses this spring system.? You only need to adjust the worm for smooth jam-free rotation into the ring gear, then tighten down the bottom two worm block mounting bolts so the DEC assembly cannot shift in the pivot hole (left block) and cannot shift in or outward (right block).? This will cut the hysteresis time needed for those motions. Compare your autoguide results.? Use the best method for real practice and don't be concerned about theory.? (Again, friction matters here and for lowest friction of the worm to ring gear I suggest CRC Brake and Caliper Grease ...that has the best technical ratings and I feel the smoothest worm behavior (no chatter) on my G11T RA worm.) All the best, Michael toggle quoted message Show quoted text On Sat, May 22, 2021, 12:23 PM Edward Beshore via <ebeshore=[email protected]> wrote: Hi Brian, I was mostly concerned about the Dec axis with my GM811. The default minmov for PhD2 in Dec is 0.2 pix. I am using my 480mm Pronto (206.265/480 = 0.43"/μ?plate scale) and a Lodestar 2 (8.6?μ?pixels) for my guidescope setup which yields 8.6 μ * 0.43"/μ?= 3.7"/pix. With a 0.2 pix min move that is 0.2 * 3.7" = 0.74" in Dec. Seems like I should be setting my Dec min move to ~1.15 " / 3.7 "/pix = ?0.32 pix? Given Paul's response, it sounds like the Dec move commands would be getting ignored until the error gets past the minimum step. In the final analysis, that may be fine, since the error will eventually move past that and a correction will make it through but seems a bit pointless to set a minmov to a value that the mount can't accommodate. What do you think? Cheers, Ed
Brian Valente Member Profile All Messages By This Member #71821 Hi Ed >>>seems a bit pointless to set a minmov to a value that the mount can't accommodate. What do you think? I agree, but as you pointed out, there's no harm in doing that. I'd rather just stick with the recommendations of the guiding assistant and let it set the min move. If you ever move to G11 DEC you won't have to remember adjusting that.? Your imaging scale of your primary OTA may also dictate how important are these small differences.? Brian

Previous Topic Next Topic