Duet 2 Wifi Z-probe issue

Sazabi

Hey there.

I'm having some weird Duet 2 wifi behaviour with diy Z-probe and need your help. My diy z-probe is servo and ir-sensor with flag, so-called BFPTouch. Servo works just fine. But IR sensor ports fail often. I've tried
zprobe.in, after some time z-probe was always read as triggered, I've switched to ext.29, same thing, works for a time and fail. It worked quite a lot, for a few months. Now set it to exp.39 yesterday, printed about 9 hours. Today I've tried to run printer again and bum, same error. Sensor and servo power in parallel and it is 5V. Other than that, my old bu trusty Duet works just fine. I cant get what causes it. It can work for months, and can work just for few days on that port.

; Z-Probe
M307 H7 A-1 C-1 D-1 ; Disable Heater 7
M574 Z1 S2 ; Set endstops controlled by probe
M950 S7 C"^exp.heater7"
M558 P5 C"^exp.39" H5 F360:120 T6000 ; Set Z probe type to switch and the dive height + speeds
G31 P600 X0 Y80 Z1.000 ; Set Z probe trigger value, offset and trigger height
M557 X10:300 Y10:300 S40 ; Define mesh grid

infiniteloop

@Sazabi

My diy z-probe is servo and ir-sensor with flag, so-called BFPTouch.

My Z probe is based on the ideas of the BFPTouch. That one has some drawbacks, for example, it is risky to put PLA printed parts in direct vicinity of the hotend. The second problem is to operate the sensor at 5V just to save one wire. All signals are expected by the MCU to be in the 0-3.3V range.

Following your description, you seem to face an electrical problem, i.e. something is wrong with the signal from the light barrier. According to the BFPTouch’s BOM, you should use an optical endstop TCST2103 - is this what you describe as ”ir sensor”?

My mod of the BFPTouch uses a bare light barrier fed with 3.3V from the Z probe header. This allows me to evaluate the signal via Z_Probe_In in analog mode, permitting me to fine-adjust the trigger level. I never had to test various I/O ports of the MCU for appropriate digital signal detection. And due to the separate 3.3V power line, the servo cannot induce irregularities into the sensor’s power input.

Sazabi

@infiniteloop rewired endstop for separate 3v from duet board same day. Had no issues till yesterday. After powercycle, started to work again. This morning still no issue. Can't get it why it happens.

infiniteloop

@Sazabi Did you try to run the BFTouch using Z_Probe_In in analog mode? Then, you can call deployprobe.g in plain air and actuate the pin manually. Since DWC monitors the analog value in ”Z-Probe” (range 0 - 1000), you get an idea what to chose as your trigger level.

Here’s a snippet from my config.g, your values will be different. The trigger level is set with the P value in G31:

; Z-probe:
M558 C"zprobe.in" P1 F200 T10000 H5 A9 S0.003 R0.0	; mode 1 (analog in), F=probing speed, T=travel speed, H=dive height
							; 				A=repetitions, S=tolerance, R=recovery time
G31 T1 P500 X0 Y9 Z1.30					; configure probe physics; P=trigger value, X/Y offsets, Z=probe offset
							;		NOTE: less Z-offest = nozzle higher above bed

Sazabi

@infiniteloop said in Duet 2 Wifi Z-probe issue:

T1

Today I had probe issue again. Tried what you suggested. Printed one part and stuck again. When I disconnect probe, zprobe value on web interface slooowly drops. I did tried few things with few board restarts. Than I've connected probe back and it worked as expected. Sometimes, despite on m401 or m402 (deploy/retract) I still have probe triggered. I see led on sensor turning on/off, so I dont think it's sensor issue. More like something with my duet2 board.

infiniteloop

@Sazabi

Sometimes, despite on m401 or m402 (deploy/retract) I still have probe triggered.

When retracted, the state of the probe is irrelevant, you simply cannot probe anything because the pin is kinda locked. So let’s concentrate on the readings you get in the ”deployed” state.

I see led on sensor turning on/off

Did you actuate the pin manually as I proposed? If so, what readings do you get in DWC?

I dont think it's sensor issue. More like something with my duet2 board.

That’s too early to tell. First, report the range of DWC readings on multiple touch actions (with your fingertips), separately for triggered and untriggered states.

Printed one part and stuck again.

If you encounter mixed results or intermittent failures, have a look at the probe’s mechanics. The pin must move easily, it must not be tilted under any circumstances. What about the spring? Can it possibly get stuck on its way? Carefully inspect all tolerances, Follow the assembly instructions on post-printing: ”Sand the flag and clean the main channels if needed.“ Well, in some cases, to file is superior to sanding.

BTW: what kind of filament did you use for the BFPTouch - PLA, PETG, or some other material?