fix high-power bursts during deceleration #64
Conversation
The problem was most visible when engraving small shapes at low power.
|
Big thanks for figuring that out. I'll test this out as well and hopefully it works on the laser too. I also agree about opening up the group and hopefully that will happen. |
|
Hmmm...that's annoying. I'd suggested more or less that same fix a while I guess they must have somehow screwed up the change. Oh well, hopefully this time it's a fix! Many thanks! -- Steve Martin Renold wrote:
-- Steve |
|
For the record, the fix was tested on our laser too. All those random black dots are gone. |
|
looks legit, thanks |
fix high-power bursts during deceleration -- I think adjusted_rate should always be positive so checking for underun is good.
|
BTW, mailing list is now ask-to-join with many old-time members having the permission to add new members. |
The problem was most visible when engraving small shapes at low power.
I haven't actually laser-tested this yet but used an arduino to monitor the PWM output of the firmware. I have triggered an oscilloscope on the large pulse width and was able to reproduce the glitches. Now they are gone and the duty stays below 5%.
The problem was already discussed and narrowed down by "jondale" and others in the google groups: https://groups.google.com/forum/#!topic/lasersaur/1i36ZNAvRFo (I'm not posting there because I don't like the idea of paying or asking for permission to submit a bugfix. Please consider opening the group up for everyone.)
An alternative fix would have been to just make the integer signed. But I'm not sure if the full range is used or not, and whether the other uint32_t in this file should be converted to signed, too.