Etch-resistant thermometer.


I'm trying to automate some of the processes involved in making a PCB, I'm trying to automate the develop/wash/etch process by using a set of servos to lift and transfer a PCB from the developing tank into the wash tank and then the etch tank. The most important part of this is to stop me over-etching my boards when I forget to remove them.

Part of the project involves monitoring the temperature in the developer and etching baths. The problem I'm having is this, if i monitor the temperature of the plastic walls of the tank then I get very inaccurate figures, but if I place the thermometer unit in the tank then it doesn't last very long, however carefully I try to seal it. The best technique so far seems to be seal the temperature unit in adhesive-lined shrink wrap, this lasted several days before failing, the next-best technique was to seal the whole unit in blob of silicon sealant - but this has the effect of thermally insulating the unit giving inaccurate readings.

The problem seems to be that to that if I protect the thermometer module from the etchant, I no longer get accurate readings (due to the thermal inertia or insulating properties of the material), but if I don't protect the module then it doesn't last very long!

Any suggestions on how to accurately read the temperature of a liquid that eats metal?

A bit of background - the three tanks (develop, wash and etch) are homemade plastic tanks, the heaters are 50 watt, 12 volt light bulbs, and the bubble pump is from a fish-tank. A mini-pro arduino controls the temperature (when the probe hasn't been dissolved!) - I've tried several different probes, but to be honest, the probe isn't the question, the queston is how to stop it being eaten.


What is the name of the chemical you are using as an etchant?

name of the chemical

Ferric chloride.

Take a very small test tube (or other glass tube) fill it with etch resistant epoxy or silicone, stuff the thermal sensor in the still soft epoxy, fill up with more epoxy as needed to cover the leads and wire insulation. Drop the entire thing in the tank.

Based on this:

Stick the sensor in a latex glove and stretch it tight around your temperature sensor. Latex does have some insulating properties, but stretch it thin and the effect should be small.

Needless to say, keep the open end of the glove above the solution "waterline".

Get some capilliary tube and seal the end with a blowtorch.
Seal rear with epoxy but do no fully immerse the tube

Take a very small test tube

I've tried this, using the small test tubes that very small drill bits come in. Unfortunately this still results in the temperature changes being detected too slowly.

Remember, the etchant is only heated to a relatively low temperature, so only a small amount of insulation is required to cause the reported temperature to lag the actual temperature by several minutes.

Stick the sensor in a latex glove

I've not tried this, this could work, the latex should be so thin that it shouldn't insulate the probe too much.

I believe you can get thermistors encased in ceramic or epoxy. In my experience, thermistors have a very low thermal mass / are very sensitive. I have one with protected leads making it immersible. I remember it being rather cheap (< $6?) … Maybe something like this…

In your case, the temperature difference is what is important so you even may be able to get good results without the hassle of linearization.

In other words, you may want to look for a more appropriate probe.

Use a piece of heat shrink tubing at least three times as long as the depth of etchant. Place the sensor in the middle and shrink that area down. Bend the tube in two, hold both ends together with some insulating tape, adhesive or a short length of wider heat shrink. Both ends of the heat shrink can now be kept out of the etchant while the centre (sensor) is immersed. You can also seal the open ends to be extra sure no etchant gets in.

Thanks Henry_Best, if the latex glove doesn't work, I'll try that.


Ime glass is the only diy solution that works over time.
I used to use aquarium heaters with bi metallic. Thermostat which worked well but temp range of those is limited now.
The glass needs to be very thin over the thermistor.

Only solution is to buy a commercisl glass or titanium sensor.

I am not sure how accurate a contactless infrared thermometer would be in this application, but it might be worth a go.

gardner that's genius!