Thermocouple Readings Sky High


 
Steve_M said:
Something else I found was that I'd get whacky TC results if my case was too tight, as if the case was adding some extra stress/flex to the PCB. I was able to reproduce it, so it wasn't a one time thing.

Similar to your situation, my temps were sky high and eventually settled down. I would still suggest a very thorough cleaning of the board.
Click to expand...

Hi Steve,

TBH, I've scrubbed the hell out of the board numerous times... I think the joints are mostly very clean and tidy and I see no issues...

Perhaps it's a damaged board? That would be very upsetting as I'm pretty heavily invested in it :-(
 
RalphTrimble said:
I know someone reminded you to set the pit probe to Thermocouple, not sure if there was mention of setting 5 mV/C?
Click to expand...

Hi Ralph,

Sure thing, it's set to the Thermocouple and 5 mV/C and has been all along...

I just want to check, when I ordered the Mouser project some parts were unavailable so I ordered alternatives. Instead of these (http://goo.gl/8ClVKX) I ordered these (http://goo.gl/eDGrwm). Clutching at straws now :-(
 
Last edited:
Mark Swift said:
Hi Steve,

TBH, I've scrubbed the hell out of the board numerous times... I think the joints are mostly very clean and tidy and I see no issues...

Perhaps it's a damaged board? That would be very upsetting as I'm pretty heavily invested in it :-(
Click to expand...

The circuit isn't that complicated really, so maybe take a jumper wire and CAREFULLY make the connections where the traces should be to see if you can locate a problem?

Looking at the images of the board from OSHPark posted below you can follow the traces, save images to zoom in on the area in question

HMV4.2.4Top.png


HMv4.2.4Bottom.png


Notably, the output of the TC AMP (the two pins that are bridged on the end of the chip) goes to the (unused) hole for the standard probe, from there to the other side of the board where a trace goes over to the 100K resistor (as well as to the probe header, so check that area for anything that looks funny too). So, take a jumper wire and go from the end two pins that are bridged and tap the end of the 100K resistor that is furthest from the ATMega.... Does that fix it? Then, from the other end of the 100K resistor it goes over to the corner pin of the ATMega, so move your jumper there, fixed? Try directly the the pin on the ATMega (rather than the socket), fixed? That traces the output of the tc amp, you can do the same for the input if you dont find an issue on the output... You could trace this down with the continuity function of a multimeter rather than jumper wire as well... That would verify the board is good or bad.

Otherwise, you may have a cap in the wrong spot, I can't seem to find any indication of what is what on the caps, so no way to verify visually, so you have to either replace with known values or test with a meter (most likely the cap would need to be out of circuit to test). Did you use lots of flux when you soldered the tc amp? Maybe there is gooped up flux under the chip that kinda shorts things out? Some flux remover spray might help for that, or a repeated flushing with alcohol maybe.
 
Last edited:
that electrolytic cap will be fine....

I wouldnt suspect the large electrolytic caps in the power supply section, I would suspect the tiny SMD caps that you can't read the labels on..... easy to mix up....
 
RalphTrimble said:
The circuit isn't that complicated really, so maybe take a jumper wire and CAREFULLY make the connections where the traces should be to see if you can locate a problem?

Looking at the images of the board from OSHPark posted below you can follow the traces, save images to zoom in on the area in question

HMV4.2.4Top.png


HMv4.2.4Bottom.png


Notably, the output of the TC AMP (the two pins that are bridged on the end of the chip) goes to the (unused) hole for the standard probe, from there to the other side of the board where a trace goes over to the 100K resistor (as well as to the probe header, so check that area for anything that looks funny too). So, take a jumper wire and go from the end two pins that are bridged and tap the end of the 100K resistor that is furthest from the ATMega.... Does that fix it? Then, from the other end of the 100K resistor it goes over to the corner pin of the ATMega, so move your jumper there, fixed? Try directly the the pin on the ATMega (rather than the socket), fixed? That traces the output of the tc amp, you can do the same for the input if you dont find an issue on the output... You could trace this down with the continuity function of a multimeter rather than jumper wire as well... That would verify the board is good or bad.

Otherwise, you may have a cap in the wrong spot, I can't seem to find any indication of what is what on the caps, so no way to verify visually, so you have to either replace with known values or test with a meter (most likely the cap would need to be out of circuit to test). Did you use lots of flux when you soldered the tc amp? Maybe there is gooped up flux under the chip that kinda shorts things out? Some flux remover spray might help for that, or a repeated flushing with alcohol maybe.
Click to expand...

Hi Ralph,

I'm pretty stupid when it comes to PCB's and am having a hard time tracing the routes, I've no idea how you trace it to the 100k resistor?! Are you referring to the one near the 68 / 22k resistors? :confused:

Edit: I presume you're talking about the 100k near the pit jack...
 
Last edited:
well, it's not rocket science... the lighter purple lines on the board are the traces (wires) that connect the circuit. There are traces on both sides of the board, and at some through holes the traces jump from one side of the board to the other...
Here is an image of the tc amp section of the board where I have traced out the major routes for you, I'm sure you can get the idea from looking at this...

HMV4.2.4_TC_Traces_Bottom.jpg


HMV4.2.4_TC_Traces_Top.jpg


The above image shows the trace where it goes to the top side of the board and jumps over to the 100k resistor....
 
Last edited:
I checked all these points with my multimeter and get a solid thumbs up from the buzzer and each point as expected...

Ralph, I'm sorry to trouble you with this and really appreciate your help. Right now I feel like throwing it in the trash!
 
Mark Swift said:
To be clear, I'm seeing updates every second, it's just the initial settling of the temperature once the points were shorted took some time, in the order of a few minutes.
Click to expand...
The updates every second aren't what I was referring to. I mean that if you just plug a probe in (or short the two pins) you might get one update where the probe is halfway between the "unplugged" state and the settled state at the right temperature. It should not take several minutes to settle, it should take milliseconds. I suggested disconnecting the 100K resistor to disconnect the thermocouple amp from the rest of the circuit if you haven't tried that yet. Then you can see if the thermocouple amp is struggling for some reason or if the ADC input section is causing the problem. I'm grasping at straws here though, on account if it being a very unusual issue I've not experienced.
 
Also, if you have solder wick you could go over the solder on the TC amp and remove some of that solder. Perhaps if your issue is under the chip melting and removing some solder will clear it up... It helps to dip the solder wick in some flux first, that really gets the solder flowing...
 
You could always bypass the onboard TC circuit and use a fresh TC circuit on something like the rotodamper.
 
Yes, the RDTC board John Bostwick and I put together has the whole TC amp on it... You could literally lift the corner pin of the ATMega out of the socket and solder the RDTC input there (you would have to run it through a 100K resistor and put a .1uf cap across it to gnd to put the RC filter back in place). So if there is some kind of mysterious short on the board that makes the pit probe act crazy you could go that route as a last resort....
 
Okay, I woke up early this morning and decided to resolder all the caps / resistors / amp. Still no luck, it still does the same thing.

As per the testing procedure the Amp is getting a nice steady 3.25v - shorting the 2 points shown I then see the voltage start high and count down. This particular time it started at 1V and counted down, the strange thing is, if I remove my multimeter, wait a few seconds, and then put it back on, it's still counting down and has continued i.e. it doesn't start again from high. It slows as it gets past 150 ish and then creeps along a decimal point or so every few seconds. I'm baffled, and very very frustrated. As per before all continuity passes.

Note: All this was with the 100k resistor removed as per Bryan's advice.
 
Last edited:
Sorry you're having such a hard time... None of us that have built a bunch of TC amps have experienced this one before so we are all grabbing at straws.... Any chance you got a bum TC or bad connector? The one time I had a mysterious TC issue turned out I had loose wires in my TC connector that were making intermittent connection and/or short...
 
RalphTrimble said:
Sorry you're having such a hard time... None of us that have built a bunch of TC amps have experienced this one before so we are all grabbing at straws.... Any chance you got a bum TC or bad connector? The one time I had a mysterious TC issue turned out I had loose wires in my TC connector that were making intermittent connection and/or short...
Click to expand...

Hi Ralph,

I get these readings when shorting the 2 points, to I presume that removes any concerns about the TC?

Don't be sorry, it's just my luck :-(
 
yah, sorry, I was still thinking back to when you said the tc temp would float from high down to a proper temp over a period of time. You are right, the test procedure would work without the TC itself.

This morning did you just re-touch the resistors/caps or did you replace them, particularly the caps, with components of known proper values? Since you can't tell the caps apart visually if you have the wrong ones in place you would never know...

And on the soldering of the TC amp chip, did you just touch up the solder to reflow it, or did you wick up some of that extra solder as I suggested earlier? Also, did you use flux on the board when you soldered the SMD stuff on?
 
RalphTrimble said:
yah, sorry, I was still thinking back to when you said the tc temp would float from high down to a proper temp over a period of time. You are right, the test procedure would work without the TC itself.

This morning did you just re-touch the resistors/caps or did you replace them, particularly the caps, with components of known proper values? Since you can't tell the caps apart visually if you have the wrong ones in place you would never know...

And on the soldering of the TC amp chip, did you just touch up the solder to reflow it, or did you wick up some of that extra solder as I suggested earlier? Also, did you use flux on the board when you soldered the SMD stuff on?
Click to expand...

Howdy Ralph,

This morning I replaced all the caps and resistors, I'm not sure how to confirm the value of the caps - but I ensured they came out the correctly labeled Mouser bag! I was considering ordering a separate batch of caps just to make sure they're correct. I love the option of the RD board, but being UK based, I'm not sure on the logistics of locating such a board; I've already sunk more £££ than I envisioned into the HM. That doesn't mean the rubber arm cannot be twisted further however ;-)

I'm actually doing my first cook now, as like I've said previously once the temp has come down it seems reasonably accurate (I keep checking it against a known source every now and again to ensure no foul play!).

http://postimg.org/image/4s9z3zkm5/
 
After the temp has come down to normal, if you hold the TC in your hand how quickly does it rise and then when you let go does it drop back down quickly?

On the caps, I guess beyond trusting mouser, and they do make mistakes, you could measure them with a good multimeter that has a cap test function. Most cheap meters don't have the cap test though...
 
RalphTrimble said:
After the temp has come down to normal, if you hold the TC in your hand how quickly does it rise and then when you let go does it drop back down quickly?

On the caps, I guess beyond trusting mouser, and they do make mistakes, you could measure them with a good multimeter that has a cap test function. Most cheap meters don't have the cap test though...
Click to expand...

Yes, the strange thing is, once it's come down, it reacts what I presume is normally... It also drops back down very quickly too (Albeit not past the lowest value if it was still settling. I have a pretty decent multimeter I think (http://goo.gl/yLVZHS), but like most things, it's probably far more than I need, and hence I'm not sure how to fully utilise it! :-/

It seems a very strange issue indeed, I'm just hoping if I can live with it - I thought it might be solvable through electrical genius, but it seems on this occasion possibly not.
 
You must log in or register to reply here.

 

Back
Top