GRBL missing steps when jogging and de accelerating
Hope de accelerating is some sort of word and this is close enough place to post this.
I have currently V1.1f GRBL motion slowing down is issue when jogging, not when program is running. Didi use google and many other search parameters and seems like this was issue on GRBL 0.8
Lets say 1mm jog step is needed, machine performs ok. IF I decide to hit keyboard for second mm when machine or GRBL is slowing down, machine can lose most of this second mm. How much is missing depends what is left from previous mm. Hope somebody figures out what I´m trying to describe, might do short video to demonstrate this. My machine can run hours without losing any steps, but jogging ruins everything, steps might be missed before Zero point and tool change jogging will escalade thing to worse. I have homing switches, but those wont help if program zero is wrong all ready.
Re: GRBL missing steps when jogging and de accelerating
What version of grbl? What sending program? What version of that sending program?
Does the position report correctly report the machine position even though it didn't go where you wanted? If so then it is not lost steps, but the implementation of jogging in the sending program.
Re: GRBL missing steps when jogging and de accelerating
GRPL 1.1f
Sender programs tested:
GRBL panel 1.0.9.17 (latest) steps are missed during jog, at certain spaces described at first post
UGS 1.0.9 steps are missed during jog, at certain spaces described at first post
GRBL Controller 3.6.1 not supported new jogging system or new coordinates, but seems to work ok when jogging.
Position reads exactly what it should, but machine goes slightly more or less...
This feels to me it is not related to sender programs, but I might be wrong. Thinking to mess with accel setup and lets see what it will do or do not.
Re: GRBL missing steps when jogging and de accelerating
No help, set accel to 10 and machine moves reeeaaalllyyy smooottthhhhlyyy, but hitting jog more at same direction on wrong moment still result same problem. Tested with another computer, my desktop is known to be sometimes at relaxed state but no luck.
Forgot to add: this happens on all axes and both direction, so this behavior rules out most of the obvious errors.
GRBL missing steps when jogging and de accelerating
Are you using a real Arduino Uno board or a clone. Some grbl users have noticed some clone boards having serial communications issues. Mostly due to problems with the ch340g serial interface chip or others that are used instead.
If you are still having problems, I would try a authentic Arduino board.
Re: GRBL missing steps when jogging and de accelerating
Yes, few topics down, there seems to be CH340 related issue and CH340 usb driver pops up when connecting to CNC, DOH! Luckily I have few genuine arduinos, so let´s see what happens.... If this is not the issue, there are not much more to test.
Re: GRBL missing steps when jogging and de accelerating
And this was not the issue, great. Maybe I have to make arduino cotrolled keyboard thingy to slow down my jogging buttons. It is quite safe to press jog button again, if cnc is completely stopped. Pressing jog button when cnc is slowing down movement and all goes to pieces.
Thing I might still try is adjust micro stepping, something must be wrong if this is only instate GRBL is not working
Re: GRBL missing steps when jogging and de accelerating
Aaaannnnddd Tadaaaa! setting micro stepping to half does absolutely nothing at all :tired:
hmmmm
Re: GRBL missing steps when jogging and de accelerating
Post your $$ config settings. I use grbl almost everyday and even have modified and compiled custom versions for my laser. This is strange that it does this since I use grbl panel and hit the jog keys multiple times with no loss of position.
What is the machine specs?
Driver model, voltage, leadscrew or belt etc.
Re: GRBL missing steps when jogging and de accelerating
I might found culprit, but not yet jumping up and down and act like idiot :idea: Not even saying what it might be until vigorous test is complete.
CNC is chinese origin 3040 all ballscrew with Finnish improvements
drivers are A8somethinsomething those usual's
Everything runs on now 12V except VDF spindle
Grbl 1.1f ['$' for help]
$0=10
$1=25
$2=0
$3=0
$4=0
$5=1
$6=0
$10=3
$11=0.010
$12=0.002
$13=0
$20=1
$21=0
$22=1
$23=1
$24=100.000
$25=1500.000
$26=35
$27=3.000
$30=1000
$31=0
$32=0
$100=800.000
$101=800.000
$102=800.000
$110=1500.000
$111=1500.000
$112=1500.000
$120=40.000
$121=40.000
$122=40.000
$130=280.000
$131=370.000
$132=48.000
ok
Re: GRBL missing steps when jogging and de accelerating
I see nothing out of place with the config settings. Your steps/mm and max rate places the microcontroller at 20,000 steps/sec. This is under the maximum 35,000 steps/sec the microcontroller can do. Plenty of headroom there.
Maybe try setting $1=255. This prevents the stepper drives from being disabled.
Not familiar with those stepper drivers or performance characteristics.
12volts is rather low to be driving the motors for the speed your spinning the motors. Try upping the voltage to atleast 24 (only if the drivers can accept that). That will push the torque curve to a more acceptable level. Use atleast 8 microsteps, this smoothes low speed motor vibrations. Helps with motor resonance too.
Re: GRBL missing steps when jogging and de accelerating
Quote:
Originally Posted by
james_III
Position reads exactly what it should, but machine goes slightly more or less....
I'm not sure what you mean by this. Lets say you have an axis zeroed and a 1mm jog increment selected and you jog once then jog again before the first finishes, does the position read 2mm, or does it read one point something?
The reason I ask is because V1.1 has a jog cancel command that if implemented in the sender, will cancel the remainder of the first jog once you command the second jog. The result would be that the machine would only move a part of that first jog and then the full second jog. Result, the machine moves one point something mm.
Re: GRBL missing steps when jogging and de accelerating
Quote:
Originally Posted by
109jb
I'm not sure what you mean by this. Lets say you have an axis zeroed and a 1mm jog increment selected and you jog once then jog again before the first finishes, does the position read 2mm, or does it read one point something?
The reason I ask is because V1.1 has a jog cancel command that if implemented in the sender, will cancel the remainder of the first jog once you command the second jog. The result would be that the machine would only move a part of that first jog and then the full second jog. Result, the machine moves one point something mm.
Does it? I guess grbl panel doesn’t cancel mid jog when hitting jog multiple times. It just keeps sending them into the grbl queue. If you press hold or reset it will cancel your jog and halt right away.
This is how I am used to using grbl. I wouldn’t want it to cancel mid jog. I use hold if I want to stop right away.
Re: GRBL missing steps when jogging and de accelerating
Yep. Here is the description of the jog cancel real-time command:
Quote:
0x85 : Jog Cancel
Immediately cancels the current jog state by a feed hold and automatically flushing any remaining jog commands in the buffer.
It depends on how the GUI sender is programmed. Personally I don't use the jog cancel command during incremental jogging in my GUI, but I don't know how other GUIs are programmed.
Re: GRBL missing steps when jogging and de accelerating
Ya I wouldn’t want it done that way too. I would want it to move exact incremental amounts all the time.
Re: GRBL missing steps when jogging and de accelerating
Quote:
Originally Posted by
jfong
I see nothing out of place with the config settings. Your steps/mm and max rate places the microcontroller at 20,000 steps/sec. This is under the maximum 35,000 steps/sec the microcontroller can do. Plenty of headroom there.
Maybe try setting $1=255. This prevents the stepper drives from being disabled.
Not familiar with those stepper drivers or performance characteristics.
12volts is rather low to be driving the motors for the speed your spinning the motors. Try upping the voltage to atleast 24 (only if the drivers can accept that). That will push the torque curve to a more acceptable level. Use atleast 8 microsteps, this smoothes low speed motor vibrations. Helps with motor resonance too.
You are correct, setting $1 is the command to turn things to good again.
Quote:
Originally Posted by
109jb
I'm not sure what you mean by this. Lets say you have an axis zeroed and a 1mm jog increment selected and you jog once then jog again before the first finishes, does the position read 2mm, or does it read one point something?
The reason I ask is because V1.1 has a jog cancel command that if implemented in the sender, will cancel the remainder of the first jog once you command the second jog. The result would be that the machine would only move a part of that first jog and then the full second jog. Result, the machine moves one point something mm.
Yes, position reads 2mm.
Setting stepper drivers to power at all times fixes this problem completely, but not sure why? Problem is on all axes on both directions, it is simple not possible to be driver issue. Setting GRBL $1=255 and then back to $1=25 machine will work ok for several minutes, this makes no cense. If reset is pressed problem will occur immediately. One occasion happened when machine was on idle for hours and when started to jog, it did jog perfectly. Somehow this complete thing must be time related and setting $1=255 cancels this timer. I wish this idle timer was minutes and not microseconds.
Now with the video.
https://youtu.be/a72EpwglWnc
It is really nice to change tools without worry and no need to find sloppy parts from machine, seems like I have been barking wrong dog (chair)
GRBL missing steps when jogging and de accelerating
If $1=255 fixes the issue then the driver probably requires extra time to be enabled(wake up) before the first step pulse is sent. If there is a data sheet for the driver chip, there might be some specification for the enable wait time.
Btw the $1=255 was a wild guess on my part!!! Wouldn’t of thought it was the issue.
Re: GRBL missing steps when jogging and de accelerating
Drivers are A4988 DMOS Microstepping Driver with Translator And Overcurrent Protection and as far as I can understand on/off time is 20-40 micro seconds.
https://www.pololu.com/file/0J450/a4...translator.pdf
I think this idle delay time is not that important, because I feel like idle does not have time to kick in before everything goes wrong and second movement is called. I did test this before and did this again, some jogging with double time 80 ja max time 254 aaaannd results are boring. Digging timings even further needs oscilloscope, but seems like my cnc is stuck with all time powered stepper.
Re: GRBL missing steps when jogging and de accelerating
I thought you were a different stepper drive. I use a a4988 in my laser y axis driving a 50oz-in nema17 and grbl-lpc. I have never seen the missing jog-step issue. I know lots of 3d printers use that driver with chip enable activated.
I do know the a4988 is susceptible to over heating and it will loose steps. Active cooling with a direct fan cooling on the heat sinks is recommended. Doesn’t sound like your problem though.
There is a big difference in quality between some of the clone a4988 and the original pololu a4988 drivers. You may want to try one and see if that fixes your problem.
Re: GRBL missing steps when jogging and de accelerating
Yes I might test better drivers someday, but I have only one bad driver sofar of these cheapo types. I have few of those running long and hot, meaning 2 3Dprinters, laser table engraver thingy and this problem engraver +"few" camera timelapse thingies. Of course there is always plan for more, maybe next CNC will get better drivers. Too bad I can´t test this issue with my lasertable, all electronics would be identical, we did move and that darn thing wont fit stairs, DOH!