Best Z Probe for Duet 3 Mini 5+

zapta

@pfn said in Best Z Probe for Duet 3 Mini 5+:

If it isn't obvious, I too have a clicky probe (annex quickdraw in this case).

attach/detach, I called it picked/drop

Thanks, will give it a try, including attach/detach between each probing.

BTW, I do homing, auth leveling and mesh probing with hot bed, since it can also impact the results.

This post is deleted!

PaulHew

My Euclid PCBs arrived today, in the UK.
Have everything ready apart from the magnets, unless I want to spend more on delivery than the magnets from USA.

I have run BLT originals on my RailCore / Voron2.4 and Ender3.
The RailCore struggles with back right corner with a BLT, nothing wants to stick to that area! Always too high.

Hoping the Euclid probe will alleviate that issue!

zapta

@pfn, I run 100 iterations of attach/probe/detach and got the message below

"G32 bed probe heights: -0.129, mean -0.129, deviation from mean 0.000"

Does it mean that I need to increase the Z value in my probe setting?

G31 K0 P500 X-2 Y19 Z9.1

Also, is there a way to get the 100 probing values or at least min and max?

echo "Home all"
G28

while iterations < 100
  ;echo "Probing " ^ {iterations}
  
  M98 P"/macros/clicky_pick"
  M98 P"/macros/z_current_low"
  M98 P"/macros/speed_probing"
  
  G30 P0 X127 Y106 Z-9999
  
  M98 P"/macros/speed_printing"
  M98 P"/macros/z_current_high"  
  M98 P"/macros/clicky_drop"

; ---------------
echo "Last probing"

M98 P"/macros/clicky_pick"
M98 P"/macros/z_current_low"
M98 P"/macros/speed_probing"
  
G30 P9 X127 Y106  Z-9999 S-1

M98 P"/macros/speed_printing"
M98 P"/macros/z_current_high"  
M98 P"/macros/clicky_drop"

echo "Clicky test done"

zapta

@arnold_r_clark said in Best Z Probe for Duet 3 Mini 5+:

The probe is not offset from the nozzle on either the X or Y planes, the switch is picked up from its dock and sits directly under the nozzle at a known distance below the nozzle hence Zero/Zero.

That's very neat. Thanks for the explanation. I thought that this is what you meant but didn't know how it can be designed.

pfn

@zapta your macro has a mistake it should be G30 P{iterations}

So you didn't get an aggregate result

The final G30 should be P100

Also, I'm a bit concerned, it says the probe result was 0.129mm lower than the homing probe. It could mean the mount is inconsistent.

With the correct P arguments, the result for G30 S-1 should list all probe values

zapta

@pfn, yes, you are right. Here it is after fixes. Not sure where the 0.109 offset comes from but my slicer is set to print at z offset of 0.1mm.

Is this a reasonable probe accuracy?

G32 bed probe heights: -0.120 -0.106 -0.110 -0.111 -0.114 -0.111 -0.110 -0.098 -0.111 -0.105 -0.100, mean -0.109, deviation from mean 0.006

var N=10

echo "Home all"
G28

while iterations < var.N
  echo "Iteration " ^ {iterations}
  
  M98 P"/macros/clicky_pick"
  M98 P"/macros/z_current_low"
  M98 P"/macros/speed_probing"
  
  G30 P{iterations} X127 Y106 Z-9999
  
  M98 P"/macros/speed_printing"
  M98 P"/macros/z_current_high"  
  M98 P"/macros/clicky_drop"

; ---------------
echo "Last probing"

M98 P"/macros/clicky_pick"
M98 P"/macros/z_current_low"
M98 P"/macros/speed_probing"
  
G30 P{var.N} X127 Y106 S-1 Z-9999

M98 P"/macros/speed_printing"
M98 P"/macros/z_current_high"  
M98 P"/macros/clicky_drop"

echo "Clicky test done"

pfn

@zapta the 0.1 comes from the probe location, I assume your homez is at a different location. So z=0 at your homing location results in an offset at this x/y

In terms of repeatability, it's pretty good. Bltouch advertises 0.005mm stdev. I got 0.002 on my ir sensor. without unlocking, I got 0.002 on the microswitch. Probing speed can also affect consistency, I normally probe with F100.

All things considered being under 0.01mm is good enough, unless you're pushing like 0.04mm first layer heights... babystep resolution is like 0.01mm at most (and more often 0.04 or 0.02mm)

zapta

Thanks @pfn, I will keep it then. I find this Clicky probe simpler than the Voron V2.4 stock end switch + inductive probe and so far it seems to be reliable, never missed a drop or pickup and the probe stayed in place when I turned the machine over to access the electronics.

BTW, since you mention speed 100, I reduced the M558 speed to F500:100 (was F120) and got slightly lower deviation from mean but I don't know if it's statistically significant.

G32 bed probe heights: -0.140 -0.141 -0.141 -0.151 -0.149 -0.140 -0.146 -0.136 -0.146 -0.143 -0.150, mean -0.144, deviation from mean 0.005

To answer the original question here, Duet Mini5+ and a Clicky probe work well for me.

pfn

@zapta lol, whoops, that was a typo, I meant I normally probe at F300.

Yeah, I'm not sure it's very different to warrant going slower. Don't get rid of the original voron endstop switch tho! Do this!

https://youtu.be/8P1tNT4el8Y

zapta

@pfn, I am using dual speed F500:100, I don't waste much time lowering the Z.

As for the video, that's a nice trick, having a fixed reference point that can home both the nozzle and the Clickey switch. Personally like minimalist systems so throw away whatever I can. I calibrate the Z probe offset with babysteps, setting initial value, than finding a good babystep distance and substracting it from the initial z offset. Once this is set, I do final adjustments, as needed, with the slicer's z offset which is easy to adjust.

As for the trick above, I think it's accurate only if the clickey and static switches trigger at exactly the same time. Otherwise you incur error base on the current travel locations of each of the two switches.

Anyway, I love my Clickey probe and thanks to the test you suggested I also know it's reasonably good.

sinned6915

@paulhew Thanks for your support!
There are many in the Railcore community that are VERY happy with it. From the mounts that we have designed and dock options. We have one that attaches to the corner of the extrusion if you don't have extra nuts in the extrusion channels.
Scan the QR code and enjoy! We are always available on the Discord channel too.

Pickett

@zapta thanks for the answers bro Ill be looking into making my own clicky probe as Ive got so many endstop switches laying around!!! I just ordered a six dollar filament runout sensor and I just thought about it and I opned that filament runout sensor and sure enough all it was internally was a mechanical endstop switch with a roller wheel on the tip of the metal contact bar!

sinned6915

@zapta @engikeneer

That is how we do our switch proving in our heated and non-heated chamber. We are over 250,000 deploy-probe-stow cycles on the test printer: batches of (X,Y) fixed coordinates, at random Z dive heights, and various speeds. After a bunch of cycles, we maually check the switch, magnets dock, etc...

zapta

@sinned6915 said in Best Z Probe for Duet 3 Mini 5+:

We are over 250,000 deploy-probe-stow cycles on the test printer

Does it matter what micro switch we use? Are the all about the same or do some models result in tighter accuracy?

pfn

@zapta

As for the trick above, I think it's accurate only if the clickey and static switches trigger at exactly the same time. Otherwise you incur error base on the current travel locations of each of the two switches.

This isn't actually a problem, only one of the 2 switches is aligned to activate. There is a static offset of the mechanism, but it's constant.

zapta

@pfn said in Best Z Probe for Duet 3 Mini 5+:

This isn't actually a problem, only one of the 2 switches is aligned to activate.

I think this is exactly the problem, At the time that the main switch activates, the other switch is not necessary at its activation point (e.g. fully pressed or fully released) which introduces a an error. But anyway, you can always fix it latter with babysteps.

BTW, it just occurred to me that now that the inductive probe is gone, I can mount on that small clickey bracket an accelerometer. It's a good position, close to the nozzle, and with convenient route for the wires.