Hakko 936 clone to run from DC Switched Mode Power Supply?

[Tek Tech]

I'm getting some parts for the 936 clone for a few dollars from ebay: the iron, the controller PCB w. knob, and a black plastic enclosure. It requires a regular 24V 50VA transformer which I want to substitute with a switched mode 24 VDC.

I would replace the triac (for AC) with a MOSFET (for switching DC) and the triac controlling optocoupler with a transistor output one. It seems very straight forward project. Did anyone make a simple and working AC to DC conversion like that in this particular station? What would you suggest, caution against, or recommend (other than buying the 50VA transformer)?

There is another 2-year old thread on the same topic, but without any resolution.
https://www.eevblog.com/forum/projects/hakko-936-clone-run-on-dc/

4 reasons I want to try this:
- Challenge (if it hasn't been done before),
- Saving money, space, and weight,
- Create something new to share with others,
- Why not?

This is the schematic:

[amyk]

It's been done commercially already --- look for the "9936" model of clone. I suspect the reason it's not very popular is because an SMPS is less reliable and tolerant of overloads, and the weight of the transformer helps to keep the base in place on the desk.

[magic]

I have done it with a custom circuit made of a single comparator, MOSFET and some resistors soldered to a SOIC breakout board. It sorta worked.

Things to watch out for:

A1321 type heater (with ceramic heating element and thermistor rather than thermocouple) is piezoelectric, sharp voltage steps will cause it to "tick" audibly when it turns on/off. Can be annoying and perhaps harmful to the heater Not a problem with the A1322 and A1323 parts with thermocouples and nichrome wire heaters.

At turn-off, inductance of the cables will cause voltage spikes at the transistor. Add a freewheeling diode or observe RBSOA, avalanche ratings etc of the transistor.

The sharp turn-on and turn-off transients will couple capacitively and inductively into the ground wire. There will be voltage across the wire itself and even if its end is tied to ground you will see voltage spikes at the soldering tip. Possibly a problem for sensitive low voltage CMOS parts.
I reduced them a bit by adding a gate resistor at the MOSFET and hence slowing down the output edges. Perhaps a better solution is to add a choke in series with the heater and a freewheeling diode covering the whole heater-choke circuit.

[Tek Tech]

Magic, thanks for the reply and a ton of good info. [Dziękuję!]

I was not aware of the variety of different options. I ordered my parts from different sources on ebay and just got the controller board for the thermistor type handle (A1321) but got the handle with a thermocouple, so it's for a slightly different controller (as the one in the schematic in the 1st post). The description on ebay was misleading, as they often are. No big deal, I will just experiment some more with what I have.

I've got a Weller WESD51. This fake Hakko 936 is just a fun project.

[magic]

My driver handled both Hakko-style A1321 heaters and thermocouple knockoffs.

Passing 147µA through A1321's thermistor causes it to develop voltage output roughly close to that of a type K thermocouple and has (almost) no effect on a thermocouple sensor. It's not exactly the same voltage and slope of the curve is slightly different, but close enough that K type could be adjusted within 200~375°C and A1321 within 150~450°C with the same potentiometer. Or something like that, I don't remember exact numbers because I don't use that driver either. Temperature measured with an external thermocouple at the tip.