Autobed level servo twitches every 2 minutes. Video.

[rhonal89]

Hi there. I have a problem my servo keeps moving every 1 or 2 minutes nonestop. do anybody knows what could it be. I have switch the servo already and still the same thing. I have a video to show the problem

[Roxy]

There is a problem where the servo can twitch. There are a couple of things that can be done about that. But the movement you have is much more than a twitch. I haven't seen that much movement before.

[cub22908]

FWIW, I hesitate to post this as it makes no sense as to why it solved the same problem but here it is. I used the same microswitch you have as it is the recommended switch and had the same movement as you except it was downward instead of up. Mine moved down probably because with my servo I had to use 10 as the down position and 100 as up which is backward from what is generally used for some reason. I went through two of the recommended switches due to very unreliable bed leveling behavior, I had to watch it like a hawk. I finally used the z switch that came with the makerfarm ( the one with an arm ) and after many prints I have had perfect reliable auto bed leveling and no movement like yours (knock on wood) during the print. With the recommended switch I had to unplug the servo after bed leveling to stop the movement. No more! Like I said - FWIW.

[misquamacus]

I've been waiting for this to come up here, had this problem last Summer. No one on any of the forums I frequent had a solution, several diagnosed it as RF interference, but none of their suggestions fixed it. I limped by for a month until it caused a particularly catastrophic print failure (ripped the probe off and almost started a fire). After weeks of Googling I still couldn't find any answers, but then I located this post. That is your answer, ABL has worked like a dream ever since. If it does work for you, be sure to thank jamesdanielv; that's the only real solution to this problem I've ever seen.

[Roxy]

I've been waiting for this to come up here, had this problem last Summer. No one on any of the forums I frequent had a solution, several diagnosed it as RF interference, but none of their suggestions fixed it. I limped by for a month until it caused a particularly catastrophic print failure (ripped the probe off and almost started a fire). After weeks of Googling I still couldn't find any answers, but then I located this post. That is your answer, ABL has worked like a dream ever since. If it does work for you, be sure to thank jamesdanielv; that's the only real solution to this problem I've ever seen.

Good find! And here is some more info: https://github.com/MarlinFirmware/Marlin/issues/1885 This thread references a thread at the Arduino group where the root problem is discussed: http://forum.arduino.cc/index.php?topic=312869.30

If this 'twitching' is happening on every print, this may not be the issue. (But it could be!) What is happening is there is a 'bug' (The Arduino group says it is not a bug) where the detach() function doesn't check the state of the control signal going to the servo before disconnecting from it. If it is disconnected while in a high state some bad things happen. First, the last pulse of the pulse train going out to control the servo never ends. Or, from another perspective the control line stays high instead of being low like (some people claim) it should be. There are no standards for how these R/C servo's are supposed to do things. Depending upon the brand (and how much money you spend) you are going to get different behavior.

This is marked as a 'bug' and Marlin is going through a bug fix phase right now. I suspect in 2 or 3 weeks this will be ironed out. But for right now, here is something else you can try instead of switching servos. You can try changing the retracted and engaged angle numbers by one or two or three in each direction. Probably, you will be able to find a number where it is stable. I had to do that with my cheap eBay Tower Pro servo (that cost me $4.75 with shipping from China).

[rhonal89]

I guess is what you just posted. It doesn't happen on every print sometime I'll would print 2 or 3 items next print the servo would be crazy. Sometime the servo arm would move more than what you just sow on the video. Thanks for the info. And autobed level sometime don't work look at adamfilip he created a thread also. I posted some video. But I can deal with that stop and restart until it works.

Good find! And here is some more info: https://github.com/MarlinFirmware/Marlin/issues/1885 This thread references a thread at the Arduino group where the root problem is discussed: http://forum.arduino.cc/index.php?topic=312869.30

If this 'twitching' is happening on every print, this may not be the issue. (But it could be!) What is happening is there is a 'bug' (The Arduino group says it is not a bug) where the detach() function doesn't check the state of the control signal going to the servo before disconnecting from it. If it is disconnected while in a high state some bad things happen. First, the last pulse of the pulse train going out to control the servo never ends. Or, from another perspective the control line stays high instead of being low like (some people claim) it should be. There are no standards for how these R/C servo's are supposed to do things. Depending upon the brand (and how much money you spend) you are going to get different behavior.

This is marked as a 'bug' and Marlin is going through a bug fix phase right now. I suspect in 2 or 3 weeks this will be ironed out. But for right now, here is something else you can try instead of switching servos. You can try changing the retracted and engaged angle numbers by one or two or three in each direction. Probably, you will be able to find a number where it is stable. I had to do that with my cheap eBay Tower Pro servo (that cost me $4.75 with shipping from China).

[cub22908]

But for right now, here is something else you can try instead of switching servos. You can try changing the retracted and engaged angle numbers by one or two or three in each direction. Probably, you will be able to find a number where it is stable. I had to do that with my cheap eBay Tower Pro servo (that cost me $4.75 with shipping from China).

This makes sense now for me. When I switched microswitches my extend/retract numbers changed a little bit which is probably the reason my movement problem went away! I still had other crazy things during the auto bed leveling sequence that were also solved when I got rid of the recommended microswitches.

[rhonal89]

I received the servo voodoo28 recommend and am still getting the same servo movement. Exempt this time is down. Servo is backward. I did couple of changes to the page roxy posted and no luck. Could it be something with marlin 1.0.2. That's the only firmware I have try other than makerfarm. I try roxy one and when it flashed servo twitches like crazy. I guess bad configuration on my part on that firmware. Also changed the value by one and still happens.

[TopJimmyCooks]

What worked for me: Getting rid of the towerpro servo and getting a better quality 9G servo, and rearranging the wires to be differently bundled together.

[rhonal89]

Am looking for a better servo. Can you link me to the one you bought if possible. Thanks.

What worked for me: Getting rid of the towerpro servo and getting a better quality 9G servo, and rearranging the wires to be differently bundled together.