# How to debug a precision temperature controller?

I have an temperature controller which is not functioning as I would expect, as seen in the attached pictures. In addition to the pictures, there is a platinum thermometer attached to the thermometer input. I want to repair and find the issue with this controller, so any advice as to the correct approach is highly appreciated. As a fairly clueless mechanical engineer, I will try to explain my current approach.

To test the functionality of this temperature controller, I have plugged a little electric fan into the output of the controller such that at full power, the fan should turn. Meanwhile, I have plugged the controller into the wall socket. I have observed the following:

1. At room temperature, neither the fan nor the display of the controller respond until I elevate the temperature set point to nearly 90 C. At this point, the fan may start to buzz, but it doesnt turn. Meanwhile, the heater display turns to nearly 80.
2. No significant changes in behaviour are observed when placing the temperature probe under cold water.
3. When heating the probe, the exact temperature at which the temperature set point triggers a response changes a little, maybe 3 or 4 degrees lower. Still, the device is unresponsive at temperatures anywhere near the room temperature.

My thinking is that the temperature probe is the most vulnerable part of the device and is likely the part that has gone bad. So, from what I gather, if I short the thermometer input, then I can at least try to isolate the problem.

I assume that if I short the thermometer input, the device would simply output full power at all times.

I had assumed before looking at the circuitry that there would just be two wires across the temperature input, and I could sodder them. However, there are four as seen in the attached picture. Can I simply sodder them together to achieve the same result? Additionally, if the thermometer probe is not the actual issue, what should my approach be from there?

Any other advice and discussion of how this controller works would be greatly appreciated.

Attached images:

• An incandescent light bulb would be a much better test load than a fan for something designed to run a heater. I also agree with and upvoted Sphero. ...and it's spelled solder (not that you should.) – Ecnerwal Jun 16 '17 at 2:34

What you have there is fairly ancient, I would guess 1970s, so around 40 years old, give or take. Maybe 50.

The temperature probe can easily be replaced by a resistor for testing. You will need to figure out the connections. Assuming it's a platinum RTD it probably has a Kelvin connection meaning that two wires run to each side of the element. The resistance is probably 100$\Omega$ at 0°C but the coefficient could be 38.5 or 39.2 (older US-made instrumentation was sometimes the latter). Chances are the sensor is okay unless it is physically damaged.

simulate this circuit – Schematic created using CircuitLab

Something that old, and from your 'buzzing' comment, my first thought would be that the electrolytic capacitors (particularly the power supply capacitors) have dried out. There are only 5 of them (the green cylinders) and they can be replaced with new ones for a few dollars. Polarity is important, as is capacitance and voltage rating. Take detailed photos and get someone else familiar with electronics to do it (check the parts, do the de-soldering and the soldering) preferably, because if this is screwed up I guarantee nobody is going to want to put much effort into fixing someone else's mess. Just the way it is, sorry.

The other parts are probably about as good as new. It looks like wirewound custom resistors on the PCB- maybe not but hard to see. I think I see a dual-secondary gate-drive pulse transformer (the red object with 6 leads) which probably drives two SCRs on that homemade heatsink.

Functionally this controller probably controls a single heater with the 15A output- and probably via phase control. It's not clear why it would have both a gain and a proportional bandwidth adjustment as normally one is considered reciprocal to the other.

• Thanks so much for your comment. I'll try this during the weekend. – User2341 Jun 16 '17 at 12:37
• I suspect that "proportional bandwidth" is simply the loop bandwidth, which is independent of loop gain. – WhatRoughBeast Jun 16 '17 at 13:27

The controller will not control the speed of a fan. You should connect a voltmeter instead and monitor the output voltage as a function of temperature.