Precisionpiezo Problems

arhi

@stuartofmt said in Precisionpiezo Problems:

To get better insight into what is happening - I changed to analog.

questions

what type of cable are you running from sensor to duet
how are you configuring the threshold? only in that one line or?

for e.g. for me I was having terrible problems running it analog trough flat cable, then I switched from flat cable to coax cable (I used shielded balanced audio cable) and it worked like a charm

as for the th, mine bed.g

; called to perform automatic bed compensation via G32

G29 S2 ; delete compensations
M561   ; clear any bed transform
G28    ; home

; wait 1 second for everything to stabilize
M400
G4 S1 
M400

; change threshold value
G31 P{sensors.probes[0].value[0] + 3}
G29 S0

what's important there is this: G31 P{sensors.probes[0].value[0] + 3} and I run the same line before I home Z and any other time I want to use piezo, what it does is it reads the current value and set the threshold just above it (+3)

The only issue ATTM is that RRF3 do not probe at the speed you set for probing but it slows down when it's "close to the trigger" and as you are constantly close to the trigger value, as PPiezo do not change the analog value with distance like IR probe does, the probe move is much slower than anticipated. This is a problem 'cause PP likes to be probed "fast" as if you just very slowly touch the probe it will never trigger... dc42 said he will change this behavior so that we can turn off this "slowing of the probe speed" but they are working on more important features now so..

so all in all - switch to a coax cable, configure threshold dynamically, increase probing speed and the issues will go away.

engikeneer

@stuartofmt I think that arhi's way is probably the 'best' way (coax cable, analog with adjusting the threshold on the fly). However, I have an old PP Orion which doesn't have the analog option so have to use it in digital mode. My config for ref is:

; Z-Probe
M574 Z1 S2                               ; Set endstops controlled by probe
M558 P8 C"!zprobe.in" R1 H2.5 F400 T3000 A10 S0.01              ; Set Z probe type to effector and the dive height + speeds
G31 P450 X0 Y0 Z-0.12 ;                  ; Set Z probe trigger value, offset and trigger height>>>

A few things that I found I've had to work on with my Orion to get it working well:

setting the trim-pot for use in digital mode (tbh I did this the first time I had it 2 years ago and have barely touched it since). When I didn't set this right, I either got lots of false triggers if I even walked near the printer, or it barely triggered at all.
check your wiring to the probe doesn't put a load on the board. I found that I had to strip the outer sheath off my cable ~2inches before the connector else the load on the board from the bent cable caused repeatability issues
I tend to drop the Z accel & jerk settings during home to get a smoother motion. I think this helps with preventing false triggers a bit, but remember to set it back! I set all my accels & jerks in a macro so I can recall it whenever to set everything back to default
using the stock V6 mounts provided with the orion worked well. However, I have now switched to a Chimera so had to design my own which took several iterations to get right.
I also had an issue with some flex in my x-gantry because some things weren't tightened down properly. It's worth double checking if you've done a bunch of other mods too.

Your heatmap looks like youre getting some definite false triggers. I'd recommend using the A & S parameters in your M558. With a 10um S, mine only ever has to probe twice, if I go down to 5um repeatability, it's usually still only two or three probes, occasionaly four so doesn't add that much time for the added piece of mind & better accuracy

arhi

@engikeneer said in Precisionpiezo Problems:

setting the trim-pot for use in digital mode

they are really crap quality, if you don't get it in first few tries they ware out (they are rated for 20 turns only) and then they lose connection with temperature and vibration ... you can get it to work much more reliable if you replace that POS with some decent multiturn pot (you can google around a lot of ppl did it)

it is important you set everything properly, the tightness of the 4 screws clamping the disks is important, if you tighten them too hard it will not work if you get them too loose it will not work, so you need to get it just right

stuartofmt

@arhi and @engikeneer -- Thank you both very much.

Based on @arhi's comments I will (1) leave off using analogue for now (was mostly trying it as a diagnostic) and (2) replace the pot with something decent (this is brand new and I have already seen the instability with temperature - not good).

I switched back to digital and used the combination of the A and S parameters that @engikeneer mentioned to do some more investigation. VERY strangely - I noticed sequential runs at 10um and 5um resulted in false triggers at almost exactly the same places on the bed - even after making a pot adjustment (less sensitive). At this point I suspected something physical .....

I have my printer in an enclosure and the Bowden tube rubs on the top cover .... So I took the cover off and "tada" all except one of the false triggers disappeared. It shows up in the same spot each time. So likely wiring wiggle or similar .....

Now that I'm down to just one false trigger (and I know where to look) - I may be able to see what is happening by staring intently .....

Getting closer ..... I had high hopes for this probe after a real BLTouch failed on me (connector separated from PCB). Seems it's all a bit hit and miss.

stuartofmt

I spoke too soon. Up till now I had not had either of the heaters on.

I just did a couple of runs at print temperature and both failed to successfully complete. Failed after 10 probes with S0.01. Happened at different places.

Not sure which way to go at this point - more or less sensitivity on the pot. More or less tension between top and bottom mount. Yep, I know, change one thing at a time .... this is disappointing ...

jens55

@stuartofmt, just FYI, I gave up and went back to the BLTouch. Yes the BLTouch is less accurate but certainly accurate enough and it is a hell of a lot more consistent. You can use it no matter what the phase of the moon is!

lui2004

i think we must wait for a better mount System for Ender 5 stock and Micro Swiss Hotend.

i hope anyone can design a better one

arhi

@lui2004 said in Precisionpiezo Problems:

i think we must wait for a better mount System for Ender 5 stock and Micro Swiss Hotend.

I use PPiezo on many different printers with many different mounts and all of them have identical issues. When finally configured to work it works awesome, but every time I need to remount the head I know there will be at least a week of getting PPiezo to work.. few times I "had enough" and moved to bltouch

lui2004

i still search good mount for my ender 5 if you have one for me i can test it

arhi

@lui2004 said in Precisionpiezo Problems:

i still search good mount for my ender 5 if you have one for me i can test it

I'm very happy with final iteration of my mount on my ender5... It does require a linear rail on the X as everything I tried with the original X slide I didn't like so I added a mbn12h on the X and designed a carriage (this is 4th iteration with bmg, I also have one with flex3drive g5 but I moved g5 to my e3d toolchanger so running ender5 with bmg+e3dv6)
https://www.myminifactory.com/object/3d-print-x-carriage-mbn12h-bmg-121644

lui2004

linear rail is not the problem the problem is i have the micro swiss all metall hotend

arhi

@lui2004 said in Precisionpiezo Problems:

linear rail is not the problem the problem is i have the micro swiss all metall hotend

if you look at the mount the hotend is on the other side of the ppiezo disk so existing plate holding hotend should work with that mount

stuartofmt

I am making progress !!!

Having thought about it (i.e. went to bed) I decided to increase the F value from 400 to 600. I.e. faster approach speed. It looks to have got rid of the multi-probe not converging issues and can now see what looks like a reasonably good mesh. A little "tuning" of the bed leveling nobs and I should be good to go.

Will report back as I continue to debug. As @arhi pointed out - time (and some sort of change) will tell .....

@lui2004 - I think the mounts that I pointed to (in the other thread) are probably robust enough and have the advantage of being quite compact and using otherwise stock ender parts. At least I do not think the mounts are the source of my initial problems. Note that I printed my (slightly modiefied) top mount in PLA for rigidity and the rest in PETG for temp resilience.

stuartofmt

I spoke too soon

The precisionpiezo is fine in theory but, in my view, misses terribly in practice. Having just spent another few hrs when, I thought I had it working, only to have it go sideways again ..... I am literally pulling the plug.

I do not give up easily but the hours spent on this are not worth any claimed precision because the repeatability is simply not there. Its a good idea that simply does not live up to the claims (implied or otherwise).

Yes - I was warned but had to find out for myself .....

I have joined @arhi and @jens55 - just ordered a replacement for my trusty BLTouch (that I broke being ham fisted). I will now gladly put up with the probe being offset from the nozzle and not being able to probe the entire bed because it simply works "good enough". The precisionpiezo certainly does seem to be a case of "better is the enemy of good". Hopefully in some future iteration it will be worth revisiting.

I wonder if PP will refund me for the Orion board? Other than having been "clamped" it's effectively unused and good as new

stuartofmt

P.S. Since I have my eye on a Zesty Nimble - I also doubt that the PP could cope with the physical "noise" of a co-located extruder ....

arhi

@stuartofmt said in Precisionpiezo Problems:

co-located extruder

I used it with flex3drive G5 (zesty is a clone of flex3drive) and had no issues with it. Basically, while you are probing you are not extruding so no matter what extruder you have or not have there makes zero difference.

@stuartofmt said in Precisionpiezo Problems:

I have joined @arhi a

I still have PP on some of my printers (ender5 is one of them), I hate having to set it up every time I rebuild the head but these days it's just the matter of finding the proper tension of the screws holding the top and bottom part together as since I replaced that cable and went with good coax cable and analog output I don't have any issues after it is properly tensioned.

stuartofmt

@arhi - Thanks for the clarification. At $35 for a new BLTouch and my currentl hourly rate working on PP (about 35c per hour) .... well - you get my drift

stuartofmt

I'm curious now as to the difference between Zesty Nimble and flex3drive but that's a different thread

arhi

@stuartofmt said in Precisionpiezo Problems:

@arhi - Thanks for the clarification. At $35 for a new BLTouch and my currentl hourly rate working on PP (about 35c per hour) .... well - you get my drift

yes, that is why I'm not using it for new printers ..it's a great idea but...

@stuartofmt said in Precisionpiezo Problems:

I'm curious now as to the difference between Zesty Nimble and flex3drive but that's a different thread

in 99.9% cases, this question leads to a flame war and lock of a topic so I'll restrain myself from commenting. I heavily tested both and can give you my own opinion (not affiliated with any of them) in private chat if you are interested