Thermocouple Readings Sky High


 
M

Mark Swift

TVWBB Member
Hi All,

So I've finally got around to testing my newly built Heatermeter. Sadly I'm experiencing an issue with the Thermocouple readings being way too high.

I've resoldered the amplifier a couple of times and still no luck - I guess I'll need to resolder / replace the caps next to ensure they're correct. Continuity looks okay in accordance with the thermocouple testing page.

In essence it looks like the readings are basically double what they should be, it's current sat at 137 degrees fahrenheit. If it helps, when the probe is first connected it starts counting down slowly from 700-800 before settling.

I've ordered another amplifier to check if the current one is at fault.

Thank you in advance for any help as I'm pulling my hair out trying to get this working!

Mark
 
Just checking, but you don't have the 10k 1% through-hole resistor soldered too, do you? Did you also test the voltage with nothing plugged in (~3.1V-3.2V) then with a wire shorting the two TC+ and TC- connections together (~0.18V)?
 
Bryan Mayland said:
Just checking, but you don't have the 10k 1% through-hole resistor soldered too, do you? Did you also test the voltage with nothing plugged in (~3.1V-3.2V) then with a wire shorting the two TC+ and TC- connections together (~0.18V)?
Click to expand...

Howdy Bryan,

I'm pretty sure I only have the one resistor, and left the 10k out...

Without being home, If I recall the voltage was 3.3v, I'll grab the other readings once I'm home... Anything else you need? I'm really hoping this can be fixed as it's the final piece of the puzzle.
 
Check your grounding and your continuity between the Atmega and the amp. This is exactly what happens when there is not a good connection between them or you have a bad solder joint

Post a picture of the board
 
Bryan Mayland said:
Just checking, but you don't have the 10k 1% through-hole resistor soldered too, do you? Did you also test the voltage with nothing plugged in (~3.1V-3.2V) then with a wire shorting the two TC+ and TC- connections together (~0.18V)?
Click to expand...

The reading without is 3.25v, the reading with starts at 1.4v and heads downwards from there pretty rapidly...

John, can you give me some pointers how to test that?

Thanks guys, I'm lost with what's up with it :-(
 
Two of the center pins on one side of the TC amp chip are ground and REF pins.... you can check continuity between BOTH of those pins and a good ground from elsewhere on the board. On the HM board those pins should be shorted together (by a trace, so those pins can be solder bridged together without issue). If the gnd or REF pin does not have a solid connection to gnd the TC reading may float....

You'll have to look at the schematic and the dot on the TC amp IC to determine which side to look at for the grounds (or test continuity with POWER OFF), I don't recall off hand which side it is....
 
OK. see your pics, you obviously know which (2 sets) of pins can be bridged on the TC amp....
The 1K resistor is what looks most out of place, though looks like it should be good enough really....
I just took a look at the board and it seems the trace that carries the output signal from the TC amp goes to the back terminal for the standard probe, through it the signal jumps to the other side of the board and a trace carries it to the cap and resistor, from there to goes to the ATMega. I would check to make sure there is good continuity through that standard probe contact to the other side of the board, never heard of a problem there but it is something to verify in a situation like this.
 
From doing a quick look at pictures, your soldering looks a little lacking on the smd parts around the amp itself. But, I dont have more time to look, as im standing in line at post office.
 
RalphTrimble said:
OK. see your pics, you obviously know which (2 sets) of pins can be bridged on the TC amp....
The 1K resistor is what looks most out of place, though looks like it should be good enough really....
I just took a look at the board and it seems the trace that carries the output signal from the TC amp goes to the back terminal for the standard probe, through it the signal jumps to the other side of the board and a trace carries it to the cap and resistor, from there to goes to the ATMega. I would check to make sure there is good continuity through that standard probe contact to the other side of the board, never heard of a problem there but it is something to verify in a situation like this.
Click to expand...

I checked the ground, it looks like it's the top 2 centre pins (Board orientated with the LCD at the top). The grounds tested fine to both the test point near the probes, and the one connected to the ethernet plug. I also checked continuity along the paths shown on the TC schematic, they seem to check out and everything seems to be connected, at least to my knowledge... I presume that means I've either got a faulty Amp, or one of the caps has been sent in an incorrect package?

:-(
 
Add solder to your caps(1n looks like they are missing solder) and check all the connections with a hot soldering iron. Make sure in the settings you have the Heatermeter set to thermocouple. I will sometimes have the heatermeter on, while making solder corrections. If you try that, be very careful. I do it, so that I have a good feedback as to when I find the fault with a solder connection.

I really doubt you have a bad amp. But, I guess its not unheard of.
 
I had some initial TC amp problems. Everything looked good from a mechanical & electrical point of view, but the temps were way off as well. I ended up scrubbing down the board with a toothbrush and some 99% isopropyl alcohol a couple of times and my problems went away.
 
did you change the settings to a thermocouple probe?

Mark Swift said:
Hi All,

So I've finally got around to testing my newly built Heatermeter. Sadly I'm experiencing an issue with the Thermocouple readings being way too high.

I've resoldered the amplifier a couple of times and still no luck - I guess I'll need to resolder / replace the caps next to ensure they're correct. Continuity looks okay in accordance with the thermocouple testing page.

In essence it looks like the readings are basically double what they should be, it's current sat at 137 degrees fahrenheit. If it helps, when the probe is first connected it starts counting down slowly from 700-800 before settling.

I've ordered another amplifier to check if the current one is at fault.

Thank you in advance for any help as I'm pulling my hair out trying to get this working!

Mark
Click to expand...


I had the same problem today. My Pit Probe read 630 degrees. I did a continuity test of everything around the thermocouple and the ATMEGA. I sweated the SMD components with a reflow gun. I scrubbed the board with isopropyl and a toothbrush. I tried a different (known working) thermocouple probe. I verified the thermocouple voltages were 3.2 volts and 0.17 volts, and that i'd left the 10k resistor out. I was most of the way through building another heater meter board up to test it. Finally, I connected to the "heatermeter" wifi with my laptop and visited http://192.168.201.1 in a web browser, clicked on "Configuration", and set Probe 0 to be a thermocouple, and clicked "Save & Apply Changes" - magically it was reading 73 degrees.
 
Last edited:
Well, I'm going to try and reflow the joints later and see if that helps... I did already try that and had no luck, I might also swap out the components with new ones as I have a couple of spares.

Question, is it normal for the Thermocouple to start high and settle? Randomly right now it's going much lower, still starting from 400 but within a minute it's settles at 93 degrees (Still to high but much lower than 300!). Normal behaviour, how long do they take to settle if this is in fact normal behaviour?

Many thanks.
 
Okay, I've resoldered a new Amp, and either by coincidence or pure luck it's settled at 79 degrees. It did take 5 minutes to get there with a wire shorted?
 
Something still isn't right, because the 1% settling time of the voltage is 40 microseconds. The low-pass filter on the ADC input smooths that to somewhere on the order of 50 milliseconds, but the largest lag you should ever see in the HeaterMeter reading is 1 second due to the sample rate.

I can't say what's wrong though. Maybe try removing the 100k resistor by the output which will effectively disconnect it from the microcontroller and see if that makes it react properly.
 
Bryan Mayland said:
Something still isn't right, because the 1% settling time of the voltage is 40 microseconds. The low-pass filter on the ADC input smooths that to somewhere on the order of 50 milliseconds, but the largest lag you should ever see in the HeaterMeter reading is 1 second due to the sample rate.

I can't say what's wrong though. Maybe try removing the 100k resistor by the output which will effectively disconnect it from the microcontroller and see if that makes it react properly.
Click to expand...

Hi Bryan,

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.

It's now steady at EXACTLY the right temperature though and seems to be responding perfectly to further changes in temperature?

See: http://postimg.org/image/r9v7d4ygz/
 
THe thing that troubles me is your talk of the TC reading starting out high and settling down to the actual temperature.... I've never had a TC behave like that before that was working properly....
If you disconnect the output of the amp from the ATMega the TC will appear to register and the temp will float upward even when no TC is connected but the second you connect the TC amp to the ATMega it should show "No Pit Probe" or register the temperature of the TC that is plugged in, but the change should be very rapid, not a slow sweep down from the falsely high temp...
 
RalphTrimble said:
THe thing that troubles me is your talk of the TC reading starting out high and settling down to the actual temperature.... I've never had a TC behave like that before that was working properly....
If you disconnect the output of the amp from the ATMega the TC will appear to register and the temp will float upward even when no TC is connected but the second you connect the TC amp to the ATMega it should show "No Pit Probe" or register the temperature of the TC that is plugged in, but the change should be very rapid, not a slow sweep down from the falsely high temp...
Click to expand...

Hi Ralph,

I agree, but to be honest I've exhausted my experience on this one... I think it's time to give up.

I tried it tonight and it seemed accurate for a while against a proven thermometer, but is now showing an error of 100 degrees.

Oh well, I guess I can always resort to a standard probe or perhaps give up entirely! A potential waste of 280 bucks.
 
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.
 
You must log in or register to reply here.

 

Back
Top