No "issue" only a feedback...

[ulrichloose]

A wonderful circuit that is still very relevant... I modified the circuit for a different type of display and it works just great.

However, I changed two things:

I think the thermocouple amplifier is suboptimally designed. In my opinion, the 1uF capacitor at P+ is rather unnecessary and ensures that the entire board "chirps" with the PCM frequency. The filter effect with an 8 ohm heating element is marginal. I also think the 2n2 capacitor is placed suboptimally. Its effect is many times higher when used in parallel with R8 200K and significantly increases stability.

Since the 490 Hz PCM frequency "annoys" me because of the noise, I changed the frequency via

TCCR1B = TCCR1B & B11111000 | B00000100; // for PWM frequency of 122.55 Hz

which just minimizes the "noise nuisance" and has no impact on the function.

A simple single OP LMV321RILT is used as the OP and it really works without any problems.

I think the removal of the switching power supply with the replacement 78M05 a pity... At least with my display/processor, almost 2 watts are "wasted", which increases the temperature significantly. Since I wired the whole thing as a prototype on a stripboard, I installed the 7805, but later connected a typical China stepdown module to reduce the 24 V to 8 V. Since the converter clocks at 1 MHz, this has no measurable impact on the stability of the thermocouple amplifier.

But this is just a suggestion for own experiments/changes...

---

But also a question on this occasion:

Why is the thermocouple amplified 200 times and the ADC operated at 5 volts. Wouldn't the internal reference of 1.1 volts and significantly lower gain make more sense?

[wagiminator]

Hi, thank you very much for the many tips. Yes, if I were to develop the soldering station again with my current knowledge (it was one of my very first projects back then - and I'm not a trained electronics engineer!), I would do a lot of things differently. However, after many hundreds of hours of soldering, which this station has done faithfully and reliably without any complaints, I, on the other hand, tell myself that I should let things rest for a while and not continue to optimize them to death. But I'm very happy when others take up the design and improve it.
Regarding your additional question: I seem to remember that the internal voltage reference of the ATmega has a lower accuracy than the 5V voltage regulator. But maybe I'm wrong.