Thermocouple troubleshooting

[Steve_M]

I now have 2 HM v4.2.4 boards where my thermocouple circuits don't appear to be working.

The 2nd board board was acting up, I scrubbed it with isopropyl alcohol and that seemed to have fixed it, but in reality, it didn't. The TC circuit was reading something like 450F with nothing plugged into it shortly before I went to use it the other night. A really odd discovery was that when I got the board all snugged up in the case, the TC circuit would start to behave properly, as if the pressure of the case against the PCB was helping it in some way. Not willing to trust that it would remain acting properly while in the case, I disabled the TC probe and used the beta firmware to allow me to use Probe 1 as the Pit Probe, which worked great.

If I remove the the TC amp and associated SMD resistors and capacitors, are there tests I can do on the HM board to verify that everything else is in good working order before re-adding the TC components back on? I'd really like to be able to have a TC pit probe!

 

[Bryan Mayland]

That is strange. There's not much to check apart from the connections themselves, it's just PCB traces it isn't like there are more components involved beyond those you're inspecting.

You could try the noise firmware from the online repository so you can see the raw ADC coming in. The ~450F is really close to the dynamic range switchpoint so I'm wondering if there's something funky preventing a switch there.

 

[Steve_M]

Based on the image below, should I be seeing 10K resistance between 1 and 2 with everything connected? Right now I have the (1) pin of the TC amp lifted off the HM board, and am using a wire to jumper it over to 10K resistor on the TC (-) post.

 

[Bryan Mayland]

I just measured one and I see 11k between 1 and 2. There's the 10k resistor and the 1k on the output in series which should give you ~11k.

 

[Steve_M]

OK. So that's good then.

I've gone and replaced the TC amp, but I'm still seeing out of whack temps. ie: I'm getting a ~155F for ambient room temp.

Really not sure what else I can test.

I've got the noise firmware loaded and the system is running here https://bbq.converged.ca/

 

[Joel Peterson]

I'm so sorry. It looks as if everyone in Toronto is dead.

 

[Steve_M]

Yep. We all died as soon as I unplugged my HM.

Here's a closeup of my TC circuit

 

[RalphTrimble]

Though a few solder joints look a bit jagged and could use a little touch up, nothing seems to be shorting out or placed improperly, and everything seems to have solder in the right places...

Pin 2 is the Offset Reference pin on the TC amp, on the HM board Pin 2 & 3 (gnd) are shorted together tieing the Ref pin to gnd... I would reflow the solder on Pin 2 & 3 (center two pins on side near 10n capacitor) to make sure your TC amp and Ref pin are getting a solid ground... Keep in mind after soldering on the TC amp the readings will be way off until the TC amp cools down to room temp...

 

[Steve_M]

The TC probe is reading about 83F right now, which is still about 13F too hot, as the room is at 21C right now.

I did re-touch a bunch of the SMD solder joints after seeing how ugly they were with the iPhone + 24x magnifying glass ghetto macro lens

Next thing to do will be replace the 2x 1n and 1x 0.1u caps to see if that helps.

 

[John Bostwick]

did you get it corrected? your down to 82 now


I guess I posted a tad quick lol

 

[Bryan Mayland]

28C in Toronto on Dec 27th? Impossible!

Well the ADC to temperature checks out, so there's got to be some reason the hardware isn't working right. I see also that you have virtually no noise on the line either, so it cant be that. I am at a complete loss as to what could cause that.

 

[Steve_M]

So I gave the TC section of the board another few seriously hard scrub sessions and it looks like things are pretty happy now. I'm going to run it through some actual outdoor cooking tests to see if it will hold up as I plan to use it for real on Friday morning.

Bryan, how do I interpret the noise graph at the top?

It's still live at https://bbq.converged.ca/

 

[Bryan Mayland]

The noise graph is one period full of ADC measurements on one channel. Ideally it should be one consistent value or bouncing between two values. There should not be a large range or a waveform. A min-max of greater than 5 and you get the yellow noise icon, more than twenty you get the red noise icon on the webui.