[LinuxCNC/linuxcnc PR#481] Feature/spindle sync jitter fix

未分类 bolang 5个月前 (10-15) 22次浏览

Issue #481 | 状态: 已关闭 | 作者: rene-dev | 创建时间: 2018-08-12

标签: for-discussion-only


fix for #164


评论 (7)

#1 – SebKuzminsky 于 2018-08-13

I made a throw-away integration branch containing this PR merged into master, and I ran @robEllenberg’s recipe (from #164) for showing the Z-velocity jitter (using the sim/axis/axis.ini config). Here is the result:
!spindle-sync-jitter-481-saa

For reference, here’s the result from the current tip of master with the quantized spindle encoder commits from this PR cherry-picked in (same config):
!spindle-sync-jitter-quant-encoder-saa

So, while I agree that master has a lot of Z-velocity jitter during this spindle-synced move, I don’t see the proposed PR making the situation better. Am I missing something here?


#2 – samcoinc 于 2018-08-13

It has been a while since I have played with this branch but I think there is an extra pin that needs to be set.

This was from rob
– Add a HAL pin, motion.spindle-tracking-gain (range 0.0 to 1.0) to
control how aggressive position tracking is. 1.0 is stock behavior, 0.0 is
no position tracking.

IIRC I was playing with .5
sam


#3 – SebKuzminsky 于 2018-08-13

There is no new hal pin in this PR.


#4 – samcoinc 于 2018-08-13

hmm – this is what I had been testing
feature/spindle-sync-overhaul-2.7-rebase

from robs..
https://github.com/robEllenberg/linuxcnc-mirror/tree/feature/spindle-sync-overhaul-2.7-rebase

Are there a lot of differences? The last commits on robs are from Nov 07, 2017
This is confusing (for me)..

sam


#5 – SebKuzminsky 于 2018-08-13

This is confusing for me too. @rene-dev can you explain this PR to me?


#6 – rene-dev 于 2018-08-13

whoops, I think I picked the wrong branch. It was supposed to be the real fix mentioned above.


#7 – rene-dev 于 2018-11-05

according to rob, there is more to it. He will collect


原始Issue: https://github.com/LinuxCNC/linuxcnc/pull/481

喜欢 (0)