ABL and accuracy, which probe to choose?

I bought a genuine Prusa Super Pinda probe and it integrated really well with an EBB36 can board and the Voron stealth burner.

I’ve found it’s much more accurate due to detecting the stainless steel sheet, rather than the textured PEI surface most of us use. I found my BLTouch would produce occasional poor results if it measured a valley vs a peak in the texture.

It’s also good if you slightly damaged a smooth PEI sheet with a nozzle crash. It won’t get fooled by a sight depression in one place.

You do need to adjust z offsets in your slicer for each print surface type, but orca slicer makes that easy now.

PS, I’ve also tested this Super Pinda clone and found it to also work well enough to be reliable.

https://a.aliexpress.com/_mOVfHXs

1 Like

FWIW, I have tried inductive, capacative, light, 3DTouch, tap, and klickly. The most dependably reliable have been tap and klicky.

1 Like

How many iterations has your inductive probe typically needed? Mine often gives up after the fifth iteration.

I’m quite unhappy with my inductive probe, it’s more or less a kind of russian roulette if the first layer becomes usable…

Does anybody have experience with inductive probes of the following formfactor?

That would come in handy, because i could accommodate this type much easier (and closer to the nozzle) on my carriage.

Typically 1 to 3 attempts.

What probe are you currently using and what is that from the image?

@mykepredko thanks a bunch, I’ll get back to you. Right now, I’m short of time.

No, but I think this may help
Following mykepredko’s post about trianglelab M8 inductive proximity sensor. I stumbled over
gx-f_h_e_cata.pdf (2.8 MB)
Panasonic stopped producing probes the kind of trianglelab. Above is still active.

1 Like

Just refer to the datasheet @hcet14 provided.

Yes, it does!

Sadly the outcome of studying the datasheet makes me to expect it wouldn’t work.

The type I have in reach is the GX-F12A

The datasheet provides a stable sensing range of 3.3 mm which looks good on first sight. With aluminum, being a non-ferrous metal, the sensing range has to be multiplied with a correction coefficient of 0.53, which results in 1,749 mm.

Taking into account the variation at the maximum operation of ±8 % and the variation of the temperature characteristics with ±8 % (will definitely happen at 70°C) the sensing range has to be corrected to 1,221 mm.

That isn’t feasible for my setup: My bed consists of 6mm aluminum plate covered with 0,5mm ‘Pertinax’ (plus glue) which reduces the headroom to appx. 0,7 mm and the Z-offset theoretical to the half of that value.

I can’t make sure to mount the nozzle-tip within a range of 0,35 mm. Sensor and nozzle are mounted non adjustable at fixed positions - simply changing the nozzle can result in a larger deviation than that.

So, i’ll close the chapter of inductive probes for now and regretfully get back to my CR-Touch…

Does that apply to your aluminum bed? I just ask because the correction coefficient for inductive probes and aluminum applies in general.

As far as I can read it wants to have a ferro-magnetic flexplate (steel plate).
I do not know whether it reliably works with the aluminum plates too.

That was the reason why I got rid of my Pertinax sheet due to the former inductive Z probe.

…so you have a steel sheet now?

Hmm, my Pertinax is glued quite tight to my aluminum plate and I love it too much to peel it off. The adhesion is insane when hot and cooled down I just pick up the printed parts. I have never needed to use a spatula or felt the need to bend the bed (which is impossible, b.t.w).

Yeah I loved it as well and the reason for swapping it to flex plate and inductive probe was somehow more related to a design issue I later discovered.
But now with the new bed that can be removed in quite no time this would be a good option again… maybe a clean steel plate and glue 0.5 mm Pertinax onto it again.

@LifeOfBrian

:rofl:

But beware of warping if the steel plate isn’t massive enough. My pertinax shrinked about 2 mm on a bed width of 400 mm.

Maybe I’ll give a try to that later.

…done.

Just some numbers:

probe accuracy results: maximum 0.920156, minimum 0.913672, range 0.006484, average 0.917221, median 0.917422, standard deviation 0.001208
PROBE_ACCURACY at X:170.000 Y:185.000 Z:5.000 (samples=100 retract=1.000 speed=5.0 lift_speed=25.0)

accuracy seems to be three times better than my inductive probe and Z_TILT is dead on with the first iteration. I have to admit, that I use another CR-Touch than before, since I’ve grilled the last one … other story, but expecting Creality to use common colors for wiring ( red = + / black or blue = - ) is a really bad idea.

But 15 microns is 10% or less of typical layer height, so it’s totally fine.

Realistically and practically.

If you want the absolute best for personal enjoyment then indeed you have to look somewhere else.

Which probes work well in a fully heated chamber near a hot bed? Because the bltouch will get slowly demagnetised due to the chamber heat + bed heat (near the probe the air is affected mostly by the bed).

Leave it near a 120°C bed without lifting the nozzle/lowering the bed and in a matter of weeks it won’t detect anymore.

Sorry, but no. with the three times higher accuracy I get a Z_TILT in at least three iterations, compared to minimum three rounds of 5 iterations each until Z_TILT has matched - that matters to me and isn’t the hunt for the absolute best, but realistic and practical.

Please explain how that works. I had a look into my CR-Touch and it is an optical sensor - not magnetic. The only magnetic thing is the stowing mechanism, which has no effect to the accuracy.

Are you sure? https://www.youtube.com/watch?v=3c1L-E-Mvqk
For me, that thing is strictly magnetic! Of course, it has an optical sensor, but you may have the same problems in a heated chamber with a higher ambient temperature!

edit

I am. All this coil and magnetic stuff is only for the stowing mechanism. Mesurement is optical only.

Try this in your printer.cfg and listen:

[bltouch]
…
stow_on_each_sample: false
…

I don’t know the cr-touch, but the bltouch is (for sure) based on a Hall sensor which detects displacement of a magnet placed at the top of the pin. I dismantled one, I saw it. It is not only for stowing, it’s the basis for detection.
There is absolutely no optical detection.

That magnet is not designed with a very high Curie temperature so it will irreversibly lose magnetic strength above 70 °C or so (my guess).

BL touches are known to fail too soon if used in hot chambers (60°C).

Confirmation about bltouch: https://www.antclabs.com/bltouch-v3

And it seems that cr-touch is optical based, so we are both right :slight_smile: https://all3dp.com/2/cr-touch-vs-bltouch-differences/

… but that changes the accuracy! The solenoid changes if it heats up, but the solenoid triggers your optical sensor!

This might be promising The BDsensorM is coming soon. | Details | Hackaday.io if you’re able to keep the electronics out of your heat chamber! …but 1,5m sounds very good!

edit:
It’s available https://www.pandapi3d.com/product-page/bed-distance-sensorm