ESP8266 Community Forum

- Tue Mar 08, 2016 3:13 am #42670 Hi all,

I am building a meat thermometer using an existing probe from a shop bought device. (Ie one of those ones where you put the probe into the meat and monitor the internal temp using a display outside the oven) . The probe is terminated with a standard headphone type jack plug.

Before wiring it up, I check the resistance at room temperature and it was about 110k (fluctuating about 1k either side). I then wired it up as follows

Firstly to switch power on to it at will and reduce to 1.1v

GPIO4 ----- 440ohm ---- A -----220ohm---- GND

And then the probe connected at A

A ---- 50k ohm ------ B ------- probe(100k) --- GND

B is then connected to the onboard ADC and the GNDs are common.

This worked a treat for a while (an hour or so) with an ADC value of 724 +/- 1. All of a sudden I started to get more varying and random figures. Generally +/- 20 (sometimes much greater and the odd very wrong). Checking the voltage across the probe it shows a constant 0.704 v. After a bit more troubleshooting I discovered that my probe is now should varied resistance values for room temp (100k to 111k).

It is almost as if the thermistor is broken somehow? The other odd thing is that plugging it back into the device it came with gives correct temperatures. All very odd.

Anyone have any idea what is going on?

Thanks

Lee.

- Tue Mar 08, 2016 7:20 am #42679 Lee,

You're almost certainly trying to use the thermistor outside of it's recommended voltage and current range (you really need to check that with the manufacturer's data sheet). However, your main problem is probably RF pick-up on the leads due to the high(ish) impedance of the sensor circuit. Your quickest fix would probably be a 0.1u cap between the ADC pin and ground.

You should probably consider changing your circuit a little. Connect your 50k directly to GPIO4 to provide a higher voltage to the thermistor and to get rid of your constant 5ma drop across the 440/220 dropper. Use a dropper pair between the sensor and the ADC pin with a 0.1u as close to the pin as you can get it to reduce RF pick-up. You're going to have to check the data sheet to work out what your dropper values need to be (your thermistor resistance is going to be radically different at normal oven temperatures).

Let us know how it goes,

-John-

- Tue Mar 08, 2016 9:22 am #42683 Hi John,

Thanks very much for your reply. I will try the cap fix you suggest and post my results.

If I did use the wrong voltage/ current, is it possible that the thermistor has "broken" and now in this inaccurate state (the resistance fluctuations I mention below are seen by directly connecting the probe to my multimeter - ie not connected to the circuit at all)?

Originally, I had a 1 step voltage divider in place but because of the high resistance of the thermistor, the other resistor had to be more than twice the size in order to reduce the voltage from 3.3.v to 1v (the max for the ADC pin). The theory was that this resulted in too low a current for the ADC pin to work correctly. By doing it in a 2 step process, I could halve the resistance in the thermistor divider and thereby double the current. This did seem to give me more stability but can't be sure the issue I was seeing in the old circuit is the same as this one.

thanks

Lee.

- Tue Mar 08, 2016 11:24 am #42691 Ah... sudden light bulb moment based on your thought...

When I did the "stable" testing before it became "broken", I may well have had the probe wire coiled up (it was brand new) so may have therefore had a different impact from RF pollution.

I reread your explanation and have realised that I misunderstood your proposal. So... to be sure, are you suggesting the following....

+3.3v ------- 50k ------ A -------- probe ----- gnd

A ------- 440 ------- B ------ 220 ------ gnd

B --------- 0.1uf cap--ADC

Looking at the probe curve, at room temp it is around 100k and at around 5deg it is (from memory) around 200k. The second dropper should make sure it is well below 1v for the ADC pin... Should I adjust them to use more of the 0 to 1v spectrum instead?

thanks once again

Lee.