In order to create a homeade reflow solder oven, I am using a K-type thermocouple connected through a MAX6675 module and reading it with an ATMEGA8 microcontroller.

When the oven is disconnected from mains, the readings work fine. Through trial and error this lead me to determine that when the thermocouple itself is mounted to the chassis of the oven (which is grounded), the readings are no longer accurate. Obviously I know I can unground the chassis, but I don't really want to do that for safety purposes.

How can I make the thermocouple work while attached to the chassis so I can get proper temperature readings?

I was considering using some RTV silicone to put around the thermocouple intersection point with the frame of the oven, but I wasn't sure this was the best route since I would have to make sure the entire braided metal jacket of the thermocouple doesn't touch the chassis at all.

Also, since I am in the development stage, I haven't tried using the microcontroller without a direct USB serial connection to my computer and as such there are different grounds. Could this be the issue? Will it simply disappear when I move to the stand alone circuit. Thanks!

  • \$\begingroup\$ The probe should not touch any metal \$\endgroup\$
    – PlasmaHH
    Nov 21 '17 at 13:51
  • \$\begingroup\$ Is the thermocouple amp and microcontroller grounded to the chassis of the oven? \$\endgroup\$
    – jms
    Nov 21 '17 at 13:58
  • \$\begingroup\$ Your measurement circuit should only be grounded at one point - so if your thermocouple junction is grounded, make sure that nothing else is. \$\endgroup\$
    – brhans
    Nov 21 '17 at 13:58
  • \$\begingroup\$ Can you devise some way of attaching the thermocouple to the chassis with a mica washer and a bit of thermal grease in between so that it is electrically isolated? I guess that the thermocouple in question is not the one measuring the temperature inside the chamber. \$\endgroup\$ Nov 21 '17 at 15:50
  • \$\begingroup\$ You are correct that grounded thermocouples are much better thermally, but this chip requires a floating thermocouple with respect to ground. A decent commercial temperature controller would not require that. \$\endgroup\$ Nov 21 '17 at 15:58

After removing the duplicate ground from the USB port and powering via a DC 5V adapter (common with the micro) it works. However, I noticed the readings were very slow to respond. I didn't feel like it was worth the effort to try and completely isolate the thermocouple from the grounded chassis, so I purchased a different thermocouple that has a fiberglass jacket and upgraded to the MAX31855.

Readings are much more stable now and faster to react. Thanks again!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.