[Solved] [Duet3] Incorrect readings from DHT22 after a move

pkos

Hi everyone,

I'm wondering if anyone ever noticed anything like this.
My setup consists of a CoreXY machine running Duet 3 6HC (3.1.1) standalone and 3HC. In general, everything works perfectly, except for the DHT22.

I have it connected to the temp daughterboard pins for vcc, gnd and signal (I tried cs0 and cs1). There is a 10k ohm resistor connected between the vcc and signal legs of the dht22 itself (problem appears without it as well).

Readings come in perfect, until I start moving the motors.
Even the smallest move causes the readings from the DHT22 to go bonkers and start displaying 2k C.

I tried manually setting the location via G92 and even moving back on X by 1mm will cause the reading to go crazy.

What's funny is that as long as I don't touch the motors - data is read correctly. I can heat up the machine, run fans, turn on relays and whatnot - nothing affects the DHT22. I tried switching that one out to a new one too in hopes of a faulty sensor - but even that did not help.

Has anyone else seen anything like that? Any hints on what I should check? It's definitely not the wiring - I've changed that twice already.

Phaedrux

Does the wiring run along with the motors?

pkos

Yes it does. I guess I should shield the DHT ones then, since serial is so susceptible to noise... This will be interesting since the wires run in a cable chain.

dc42

Yes, signals using open drain outputs with pullup resistors (such as I2C and one-wire protocols) pick up noise very easily from stepper motor cables. You need to use shielded cable to connect the DHT22. Stereo microphone cable should work.

pkos

Just as an update - switching over to a shielded cable (I used exactly a stereo mic cable) did not solve the issue. I'll have to rethink and move the DHT22 to a different location and see if that helps.

dc42

Do you have a pullup resistor between the signal line and +3.3V?

A Former User

@pkos
Had something similar with temperature values going nuts every time any motor was "on" -> Had build a printer with a lot of conductive metal and it turned out I had to "isolate" the steppers because they brought so much noise into all the metal parts that the sensitive stuff just couldn´t cope with it...

E.g. if one of your motor sits directly metal to metal on a metal(-plate/-beam/-etc.) on which also your sensor sits, there is a high probability the strong field of your stepper will propagate through it and that´s it then...

To have a cheap solution to isolate your stepper while still having something strong, stable, that can handle high heat and is flame retardant try the base-material of pcb´s, like an "FR4" with "V0"

hope that maybe brings you to new ideas what to try...

I found this post because I also want to attach a DHT22, can you give me a hint how to start?

Phaedrux

@LB Curious if you had your motors and metal work grounded.

A Former User

@Phaedrux
Well in the beginning no, then yes: I made 2 separate shielding-"trees", 1 for the motors directly connected to PE (earth in germany, do not know in UK/USA how to say) and 1 shielding-"tree" for the sensors. That was separate... Even then I had problems but less then before. Only after electrically isolate the motor (we just milled the same plate that was made out of metal out of a piece of sturdy quasi-non-conductive plastic) it was what you could consider "good"/"working"

But my understanding of all of this is far too low/weak to know if it was not just an unlucky combination of all those modern high-frequency-swithing-things, e.g. the cheap switching power supply, your fridge in the same room, some other machines in the room which may have some frequency or voltage or current switching not necisarrily not connected to earth, which all can (not necessarily have to) "pollute" your mains and or earth before you even had your printer and on top of it the crazy high-speed pwm-switching to drive your big stepper motors... will never find out because I have no equipment to be able to and understand to little about it

pkos

Bah I totally forgot this thread for a while.

So to answer @LB - not the case here, motors are mounted onto ABS printed parts and only those hook up to aluminum extrusions.
Also, my printer has the frame and the hotbed grounded.

@dc42 - yes, I do. I tried putting it in multiple places to be sure (I'm not that amazing with electronics) - no change regardless of whether they were next to the DHT22 or closer to the board.

I tried switching to a different spi port, but no difference there. The only other thing I didn't try was to move the sensor to a different place so that the wires don't run close to motor wires, but this will kinda defeat the purpose of what I want to achieve. I want the sensor to keep moving up along with the gantry to know if the temps at that height are also good enough.

dc42

@pkos, what value pullup resistor did you use?

One-wire buses rely on passive pullups, which makes them sensitive to noise (just like I2C). I think a combination of shielded cable (using the shield for the ground connection) and a pullup resistor of around 2k2 ought to work.

A Former User

@pkos said in [Duet3] Incorrect readings from DHT22 after a move:

So to answer @LB - not the case here, motors are mounted onto ABS printed parts and only those hook up to aluminum extrusions.
Also, my printer has the frame and the hotbed grounded.

Just for the sake of completness and to definetly have all possibilities checked you could just unwire/cut/deplug your PE/GROUND (frame+hotbed).

pkos

@dc42 - I used a 4k7 resistor.

@LB - will try. Currently need to figure out why he 3HC stopped waking up on power on after the latest beta firmware upload (probably just a coincidence, but need to figure this out)

A Former User

@pkos said in [Duet3] Incorrect readings from DHT22 after a move:

@dc42 - I used a 4k7 resistor.

@LB - will try. Currently need to figure out why he 3HC stopped waking up on power on after the latest beta firmware upload (probably just a coincidence, but need to figure this out)

I just tried 4 different DHT´s, first the bare-ones (without housing...) now the "ASAIR AM2302" because of their double-house and pullout-safe wire-ends and screwhole on the outer housing for putting it somewhere, on a duet2ethernet1.0.4a & RRF3.1.1.
Looks good so far for me, had no troubles but to sort out how to wire them up (https://forum.duet3d.com/topic/14146/duet3d-dht-22/16?_=1605704384964)... and I like to see the humidity drop in the chamber... hope you get yours running!

pkos

So I finally got round to moving the DHT22 and played with wiring some more... only to find out the DHT22 does not work in 3.2.
I see there's actually quite a couple threads on it

@dc42 shamelessly pinging asking for help

A Former User

For everybody reading this most likely deals with the same/strongly-related issue:

https://forum.duet3d.com/topic/21061/dht22-not-working-after-update-from-3-1-1-to-3-2-0/3

pkos

Update then on the data disappearing after a move issue itself. After moving the DHT to a different place in the enclosure - problem is solved. Readings come in without a hitch now regardless of how I move the gantry. I suspect if I moved the motors anywhere close, the problem would come back, but that is not likely to happen.

Let's consider this problem solved. Now just to figure out why DHT22 doesn't work on 3.2.

A Former User

@pkos

-> We should start a thread how to print enclosures for the motors out of esd-filament

pkos

Totally