Thermocouple issues

[godnemoz]

Hi
I have a duet wifi with a thermocouple in my heated bed it keep jumping up to 2000c even when only the 5v power supply is going so I was wondering if there was a filter in the temp so it did not register just the 2000c would that work?

[dc42]

Themocouples only output about 40 microvolts per degC, so the wiring between the thermocouple and the daughter board picks up interference easily, especially by induction. Use twisted pair wiring all the way back to daughter board, and keep it away from other wiring as far as possible. If you are using a bare tip thermocouple, make sure that the thermocouple tip is insulated from the printer metalwork. If you are using a cartridge thermocouple, grounding the cartridge housing may help.

[godnemoz]

Its a cartidge thermocoupling ill try earthing the heated bed to see that makes a difference
It insulated wire within a braded metal outer
The bed has a AC heating pad on witch I havent been able to try yet beause of this issue

[Shotfire]

I was able to fix that problem by taking a piece of RG-59 coax (Same stuff used for cable tv) and removing the inside conductor and white insulation but leaving the wire shield in there. I helps to warm the cable up in the oven at a low temperature before pulling the the center conductor out, it helps to have two people, you have to work it out a little at a time. Then I ran the thermocouple wires through the cable, there should be plenty of room, on my dual head I ran two thermocouple's through it. Then on the Duet side solder a short wire (the shorter the better) onto the shield wires in the RG-59 coax, put some heat shrink over the connection, then connect the other end of the short wire to a ground on the Duet board, any ground should work. Problem solved. David, you may want to put this fix in the Wiki since this seems to be a re-occurring problem.

[dc42]

Thanks, I've updated https://duet3d.com/wiki/Connecting_thermocouples#Troubleshooting.

[godnemoz]

@dc42:

Themocouples only output about 40 microvolts per degC, so the wiring between the thermocouple and the daughter board picks up interference easily, especially by induction. Use twisted pair wiring all the way back to daughter board, and keep it away from other wiring as far as possible. If you are using a bare tip thermocouple, make sure that the thermocouple tip is insulated from the printer metalwork. If you are using a cartridge thermocouple, grounding the cartridge housing may help.

Is there on way to filter the noise (2000c) out by firmware so that it ignores the 2000c reading

I tryed to strip some coax cable but the sheild wire keeped coming with it then i tryed just earthing the sheild thats already on the thermocoupling and that just gave a reading of 2000c then I connected it to a rex_c100 and that read the thermocoupling with out no nosie

[dc42]

The firmware already does some filtering.

What is a rex_c100 ?

[godnemoz]

@dc42:

The firmware already does some filtering.

What is a rex_c100 ?

rex-c100 is a Digital PID Temperature Controller

[dintid]

@dc42:

The firmware already does some filtering.

What is a rex_c100 ?

I'll pipe in here, as I have the same problems, only with Thermocouples on hotend.

I have tried 2 different original daughterboards, so I think we can say the fault isn't here.
I've tried both channels on both boards.

Issue:
it shows temperature fine when using short wire at 10cm. - when I put by hand around it, it starts heating up just fine.
When using long wire (up to hotend) the temprature is mostly fine when thermocoupler isn't touching anything - as soon as I touch it, it jumps to 2000c
When installing in hotend it's just at 2000c constantly.

I've run a wire away from the machine, away from any potential interference, but as soon as I touch the thermocouples the temperature jumps to 2000c.

I've tried at least 5 different thermocouplers… they all work fine on other controllers/firmware combos (marlin, smoothie etc)...

[dc42]

Did you read the wiki page on connecting thermocouples?

[dc42]

I've done some more research and experiments. I can reproduce the 2000C temperature spikes (lots of them!) if I wrap the thermocouple cable around a stepper motor cable that is carrying current.

There is a mod you can do to the board to make it less susceptible to interference. The mod consists of adding a 0.01uF ceramic capacitor between each thermocouple terminal and ground. So you need four 0.01uF capacitors per thermocouple board, You can use either SMD or small wire leaded capacitors. Here is the procedure:

1. Remove the nylon pillar by squeezing the end with long-nosed pliers.

2. On the underside of the board, scrape away the solder resist to expose the copper ground plane where you want to connect the ground side of the capacitors.

3. Solder the capacitors between the terminal block spills and the ground plane you exposed.

4. Replace the nylon pillar.

Here's one I did earlier, using wire-ended ceramic disc capacitors on one channel and 0603 SMD capacitors on the other.

We will implement this modification in the next PCB revision of the thermocouple board.

[Jan]

Thanks!
I was facing severe issues with this problem as well.
Actually I am using the ardafruit break out board with the duet WiFi.
I applied this mod and added two 250ohm series resistors as well (up to 1000 ohm shall be ok according to the maxim faq https://www.maximintegrated.com/en/support/faqs/max31855-faq.html))
I bought some ferrite beads as well but didn't put them so far.

So far the result is very promising and I ordered my machine a half day print for test, and will report later …

[dc42]

I've modded some more boards and done some more tests. I tried a 100 ohm resistor in series with each terminal and no extra capacitors, but that didn't help. Adding the capacitors solved the problem, whether the resistors were included or not. So my recommendation is to add the capacitors first, and only add the resistors if the capacitors alone don't solve the problem completely.

I found it easiest to use 0805 SMD capacitors. If you have a no-clean flux pen, apply flux to the terminals spills and the bared copper. Heat one of the terminal spills with the soldering iron and use tweezers to push one end of a capacitor into the molten solder joint. Repeat for the other terminal. Then solder the other ends of the capacitors to the copper of the PCB that you bared already.

We're working on implementing this as standard in our thermocouple daughter boards.

[Jan]

Ok, from here I can confirm it works after the endurance run.
I finished a 14h print by now without temp faults (which was completely blocking since my thermistor hot end failed and I switched to TCs, at least as soon as I switched the motors on).

I now only have to decide if I take out the series resistors or leave then in.
But as long as it works I think I consider it as a running system

Thanks again for proposing this mod!

[Jackal]

David, thanks for the testing and results, I will do the mod when I got some free time.
On a side note, the thermocouples I got have crimped Y-connector, it's a bit tricky to have all 4 wires connected to the board and I ended up taping some of the prongs with insulation tape.
Is is ok to trim those connectors off and connect the bare wire to the board? Will it affect the reading?

[dc42]

Yes you can connect the bare thermocouple wires to the terminal block. Keep them twisted around each other right up to the terminal block, to minimise interference pickup.

[dintid]

@dc42:

Did you read the wiki page on connecting thermocouples?

Yes.. obviously. Since I've been trying to write "how to's for you"… Many times

But what kind of info do you especially refer to in there, as it seems you hint that there should be a solution? (apart from the soldering of small SMDs).

[dintid]

Just found that an old bead-type Thermocoupler works fine.

Unfortunately ALL the thermocouplers in m3 cartridges (fitting for newest E3D and UM2+) is doing the 2000c act. Yes, also if the thermocoupler board and thermocoupler is the only thing attached.