Flaky - maybe dying - skr

Software / Firmware SKR Boards
1

I had a z-max switch die on me recently - crazy, but it did - so I have spent the last 2 days swapping it out. Yes. 2 days.

A bit of that was me being slow, or having to print a new switch carrier that has more adjustment, but a lot of it has been thanks to what appears to be a gremlin in my SKR. I have a BTT 1.2 running the dual endstop firmware.

Unfortunately it’s just randomly doing %#|+ instead of what I tell it to do. Here’s an abbreviated list of issues I experienced this afternoon:

While at Z home, send G38.2 Z0, get an “ok” response immediately with no movement. (Happened many times). Verified with the LED on the SKR that the probe wasn’t touching. Jogging Z down would eventually ’fix’ this.

While homing Z with the new endstop switch, the endstop would trigger, Z would back off and approach slowly for the second try. Endstop switch would trigger, but Z stepper would drive straight through and clunk clunk clunk. Again verified with leds that switches were triggering and being detected, but ignored.

Using the TFT to jog Z out of the way, I set it to 10mm and tapped Z- twice - it drove straight from home into the spoil board. Triple checked after this happened, and yup, it was set to 10 mm. Thought I was crazy, but it happened again.

I issued an M114, and it responded with the current position. Then 2 seconds later it did it again. And again. And again. Every few seconds printing out the current location, and preventing me from sending any other commands until I cut power and rebooted it.

Oh yeah - at one point it just randomly decided to cut or reduce power to ONE of the Z steppers, letting that side drop to the spoil board.

There were other random issues and weird quirks, but I can’t remember them all…

Needless to say, I need a beer, a hug, and probably a new controller. Unless anyone has any ideas?

2

There are many threads here about an issue with SKR Pro 1.2s having some that ship with a bad endstop circuit design. On these, the LED circuit causes intermittent or total loss of endstop triggering.

I suggest you repeatedly trigger your enstops manually, and each time you trigger them, pull an M119. Also note what the LED for that endstop does. report those results here.

If your board has a problem, you’ll see it where you trigger the endstop, but M119 doesn’t show it. The LED may or may not show it triggered.

Repeating this a bunch of times is one way to identify if you have one of the flakey boards.

It’s fixable with a pigtail, soldering resistors, or other workarounds. Let’s see if this is your problem before trying any next steps.

2 Likes
3

Yeah, that sounds a lot like the endstop issue. It can be quite intermittent, so having it recognize the first contact and then drive through the second is definitely something that is possible (ask me how I know…)

This issue can also show up after changing the circuit resistance of a previously working circuit, such as lengthening the wiring, changing the switches, etc. (ask me how I know…)

Pigtails w/resistors between Sig and VCC added to the wiring connectors is a fairly quick and easy fix (if you have a crimping tool and Dupont M & F connectors). Soldering resistors on the underside of the board between Sig and VCC is also quick and easy, but will void any warranty (not an issue if he board is a bit older, I suppose). Busting the LEDs is supposed to work as well, but that is a destructive methodology, not everyone’s cup of tea…

If you are raising the Z past the travel limit, it is not uncommon for one side to shut off and then drop to the spoil board (ask me how I know…) Once you get within 12mm or so from the top of travel, switch to 1mm steps.

That sounds like it might be a loose motor wire connection.

Have a beer and recuperate from the frustration. I’ll send you a hug over the Internet. Try the endstop resistors, check for loose wires, and modify your moves to avoid hitting the end of the travel limits, especially on the Z moves. The controller may be at fault, but odds are that the above mentioned steps will fix most of the issues.

4

Yes, my SKR was one of these and I had to add the resistors. I’ve verified that M119 and the LEDs still report correctly.

These are nothing like the issues I had before fixing the board though - before the resistor fix, endstops just didn’t work period. Also, they don’t account for any of the drive-into-spoilboard or one side of the gantry suddenly falling…

5

The weird thing isn’t that it happened, it’s that it was 100% consistent. When it was failing, it would ALWAYS trigger correctly the first time and trigger but ignore the second time.

I wasn’t raising Z, I was lowering it, and power wasn’t cut - it drove to the board. I’m guessing that it did something similar to the repeating M114 issue I saw in the terminal where it just got stuck repeating the last command it received.

Thank you. After sleeping on it, I think I might be placing an order for a jackpot today. I built this thing in large part for fun, and this isn’t fun anymore :stuck_out_tongue: