I don't know why, but FWIW, I've seen the same sort of accordion pattern on mine. I saw the same thing when I was using an inductive probe, before I got fed up with it and bought a Smart Effector, too, so I don't think it's specific to the Smart Effector. The actual first-layer height seems to come out OK, so investigating this has been a low priority for me.

Chances are your delta arm length setting is wrong – it's just a little bit too long, if I recall the relationship correctly. It's my understanding that auto-configuration systems can't detect this error, so it's important that you get it right manually. This is set via the "L" option to the "M665" command. If you're using a Duet controller, you'd set this in config.g (although it can be overridden in config-override.g, if your config.g is set up to load it, so you might need to set it in both files to be sure). In some other firmware systems, you'd edit the firmware configuration file, rebuild the firmware binary, and re-upload it.

@Nuvag:

Nice designs, i would like to use a 50mm radial fan with a circular duct and may create a remix of srs5694's design for that.

I toyed with doing that. In fact, my design is based on an earlier effector-and-fan design that I did, which used a 50mm radial fan, before I got fed up with the inductive probe I was using and decided the Smart Effector looked like the best way to do bed probing on a delta printer. Although I've abandoned that effector design for myself, I'm planning to publish it, probably this evening (US Eastern time). You might be able to merge bits from the two designs once I publish the effector.

That said, one reason I didn't pursue the radial fan idea is that placement of the fan seemed awkward. It would either need to lie on its side or be placed upright but extend out quite a way from the effector, with a rectangular duct to link it and the circular duct. It didn't seem worth the effort to me personally to design it – but of course if that's the design you prefer, and if my design helps you get there, then that's great!

Any idea how far away the fan duct should be from the heat block depending on the heat resistance of the material? Is the silicon sock enough to shield a pla/petg/nylon duct when getting very close to the sock?

I don't know for sure, but I can say that I printed my duct in PETG, and it seems to be holding up fine to printing PETG. (I've got an E3D V6 sock on my heat block.) My other printer (a Robo 3D R1+) has a PLA fan duct that's about as close to its nozzle as my fan duct design design gets to its nozzle, and even the Robo's fan duct seems OK. (It looks like it's softened a little, but not much.) I print PETG on that printer, too, so the heat block gets up to 250C or thereabouts.

Mine is about 5mm above the tip of the nozzle. (This is using a genuine E3D V6 but with a third-party stainless steel nozzle.) This height is dictated largely by the bracket and fan, and should be similar to the height of dc42's fan duct. Note that the airflow is aimed to converge at the center of the torus at 8mm below the duct – in other words, 3mm below the print surface, in a theoretical sense. In practice, of course, the air stream is not a mathematical line; it has a diameter, there's turbulence, etc. I'm not an expert on airflow by any means, so I was just guesstimating at something that would produce good airflow in the area immediately surrounding the tip of the nozzle, without directing much airflow on the nozzle or heater block – I didn't want to cause problems keeping the hot end up to temperature.

If you care to experiment, of course, you can adjust the targetZ variable in EffectorCircularFanDuct.scad file, or adjust EffectorFanBracket.scad to lower the fan and duct. Either will require you use OpenSCAD and build a new .stl file. You could also look at https://www.thingiverse.com/thing:2573606, which is a modified bracket for the E3D Volcano. Somebody's already tried that combination and said it works, but I don't have an E3D Volcano for testing, myself. My guess is it would lower the fan duct too low on an E3D V6 or most clones.

OK, I've now tried the 1.21RC3 firmware and "M558… B1". This works, but it has a big drawback: I get heating faults during probing. Specifically, when I tried probing after both the bed and the hot end had reached their target temperatures (60C and 210C, respectively), the hot end produced a heater fault with the message "temperature excursion exceeded 15.0C." (I could see the hot end temperature plunging throughout the probing process. The bed temperature dropped, too, but not as dramatically.) I also tried probing while both the bed and hot end were still being heated. In this case, both of them produced heating faults, with the message "temperature rising much more slowly than the expected 0.1C/sec" (or 0.0C/sec for the hot end).

I looked for 1.21RC3 firmware and found this:

https://github.com/dc42/RepRapFirmware/tree/dev/EdgeRelease/1.21RC3

There's no DuetWiFiFirmware.bin file in that directory, though, which I believe is the file I'd need to update the firmware.

FWIW, I did try going to bang-bang mode on the bed heater (by commenting out the "M307 H0…" lines in config.g and config-override.g). This greatly improves matters -- the Z probe doesn't produce any false alarm except for a brief blip, apparently when the heater turns off. (Just after the bed reaches its target temperature; I can also hear the power supply's fan drop in speed.)

Mine is 12v, with an aluminum plate above the heat bed. I've just done some more experiments. First, I tried adjusting sensitivity with the M672 command, as noted in the documentation. This didn't help; I got false alarms with sensitivities ranging from 1 to 255.

Next, I next brought the hotend down to 5mm above the bed, with the bed cool, then turned on the bed heater (leaving the hotend heater off). The Z sensor triggered when the heat bed reached 57C (the target was 60C). I then took the temperature down in 5-degree increments. The Z sensor would turn off whenever I set a lower temperature (presumably because the heated bed was no longer receiving power), but would trigger again when the heat bed more-or-less reached the target temperature. This was true down to 40C, at which point I stopped lowering the temperature and began raising it in 5C increments. Again, the Z probe would trigger whenever the bed's temperature approached the target temperature. I wonder if it might have something to do with the way the Duet controls the approach to the target…? (I have done PID tuning on the bed heater.) Does it modulate the bed heater current in some way that might induce a magnetic field with some heating elements? If so, is there some way to adjust that -- say, by tweaking PID values?

I then removed the aluminum plate, and that had no effect; the Z probe would trigger with or without it.

I took the bed heater out of the frame, but left the heater plugged in. (It's got a fairly long cable.) The Z probe stopped triggering. I then moved the bed heater around and discovered that the Z probe would trigger whenever the bed heater got too close to the effector. What "too close" is varied with orientation; edge-on, I needed to get to within 3 or 4 cm of the effector. When I tilted the heater on its edge, so that a camera where the effector was would "see" it as a circular object, the effector would trigger when the bed was 10cm or more away -- about the same distance as triggers false alarms when everything's properly in place. Thus, I'm pretty sure it's the bed heater itself, not the way the cables are laid out.

I also tried rotating the bed heater in the frame (mounted normally), to no good effect; false alarms would occur no matter the orientation.

Based on this thread, I tried unplugging both my hotend fan and my print-cooling fan, reasoning that the combination of their magnetic fields and whatever the heated bed was producing might be causing problems, but unplugging the fans didn't help, so it seems to be the bed heater alone that's the culprit in my case.

@Threephourths:

Im using this on both of my smart effectors now and I really like it. Well done!

Thanks! I'm glad you like it!

@dc42:

I've not heard of anyone having that problem before. What type of bed heater is it?

It's a Chinese knockoff of a SeeMeCNC Onyx (rev. 6) heater.

I haven't yet looked into it in detail, but mine has the very odd problem that it signals contact (green LED) when the heated bed is warm (over about 45 or 50C) and the head is closer than about 100mm to the bed. Turning off the bed heater makes it work as expected. The wires for the heated bed and the effector don't cross, although my Duet and all the wires are located directly under the heated bed, so it's conceivable there's some interference that might be hard to eliminate, short of relocating the Duet.

I intend to try adjusting the sensitivity to work around this problem, but I haven't yet gotten around to that. For the moment, I can do my G32 probing cold and G29 after raising the bed temperature but after turning off the bed heater, to get a decent height map file.