MAX6675 Type K Thermocouple Huge Fluctuation


The project is this to use a MAX6675 Thermocoulpe to measure the barrel temperature of a Lambretta scooter. The Thermocouple is read by an Arduino Nano and the temperature displayed on a nokia 5110 screen.

Due to the scooter being without a battery I.e kickstart AC stater plate, I am using an AC/DC rectifier (image attached) its an LM2596 AC/DC off ebay.

Using a desktop AC supply I can get a constant DC 5V that runs the circuit without issue through the rectifier.

The problem occurs when I hook it up to the scooter, the arduino powers up as does the nokia 5110 screen, however the readings from the thermocouple are completely random ranging between 0 up to 600. I can only presume that this is due to some level of interferance / unstable current etc, but I just simply do not have the knowledge to fix it,

I’ve tried to present you with all the necessary information I can, I have attached:

A schematic of my circuit (To the best of my ability)
Photos of the circuit and system
The Code File

If you require any more information please let me know.


thermacouple_project.ino (2.4 KB)

With thermnocouples, the red wire is - not +.

Thank you for your reply, the system works fine whilst its on my desktop power supply, the problems only occur when its connected to the scooter power supply. I've swapped the terminals and have had the same result I'm afraid :/

The LM2596 is NOT an AC rectifier module. If you are applying AC voltage to the input, you will eventually destroy it, and possibly everything else connected to it.

here is the rectifier I bought and am using, it says AC/DC but considering the source I suppose it may be wrong?

Could you please have a look and let me know what you think?


That module looks OK to use.

Thermocouples and amplifiers are extremely sensitive and if the power supply is noisy, you will have problems. First test your circuit using batteries for power, before trusting an AC power supply.

I have found that the MAX TC amplifiers are pretty unstable. You might try putting everything in a metal box. I use either an AD595 or just a TC through an op amp to give me a maximum 5 volt reading, which is good for 1300C.

here is a link to my code and other stuff.

I tried the link function but it didn't work. So just cut and past it into your browset

Hi all

Thanks for the replies, @jremington I have tested the system with both DC batteries and also a stable AC desktop power supply and it functions fine, it is only when we attached it to the scooters AC power supply that we get the issues.

@RPCoyle Thanks for the info, do you think the use of a different op amp will help the stabilise the signal?

I don't think the op amp is the problem. The digital signal produced by the MAX seems to be very prone to being disrupted by any noise on the line powering the chip. I had a problem with power line noise coming clear through the computer power supply and causing problems. If I unplugged the computer from the 120 AC line, the problem went away. I tried a few different filters, but none of them worked.

I suspect you're are right, the AC supply is generated by a stater plate directly off the 2-stroke engine, so I suspect it is a very unstable, noisy supply... I am very stuck at the moment. My skill is hugely limited in this field.

Yes, I busted my head trying to work around the noise problem… The MAX chip seemed the ideal solution. Too bad it didn’t work.

I don’t know what accuracy you need on your read out. With the setup I wrote the code for, I could get readings within about 5 deg C. or less if you adjust the op amp signal gain to be correct right around the temperature you expect to get.

Good enough to a ceramics kiln. Or as they say " good enough for horse shoes or hand grenades".