Thermistor Tester Circuit for US Sensor thermistor cables

uxorious

We are starting a project at work that will consist of adding mono headphone connectors onto the ends of thermistor wires. I would like to create a tester to verify the connections by reading the resistance and verifying it against a known table from the thermistor manufacturer and current temperature. Here's my idea:

1) Load the temperature/resistance table into eeprom
2) Connect a temperature sensor to the stamp (SHT11, I have an extra one)
3) Connect the thermistor to the Stamp (RCTime circuit?), push "go" button
4) Check current temperature on SHT11
5) Lookup current temp in table to get expected resistance
6) Measure resistance of thermistor
7) Compare values are within a range of each other
8) Light a green or red led for pass/fail

Anyone see any problem with this?

My other idea (so that I can use my last spare stamp for other stuff) is to make the tester just work of the resistance of the two thermistors and if the difference is greater than X ohms a circuit is allowed to drive an LED. So a known good thermistor is plugged in and say the room temp is 25C; the resistance is 10k per mfr spec. The thermistor under test should also measure 10k. I would think a simple circuit with 2n2222's and resistors could be made that would light an LED if the two "resistors" are the same value???? The room temp range of resistance for the sensor is 9k-14k, see attached resistance curve for the sensor. We are using USP7714 and USP8116 from US Sensor.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
~~ dRu ~~

Post Edited (dRu-dRu) : 11/21/2007 5:15:05 PM GMT

kjennejohn

Try a window comparator. See the diagram attached. Let's say the Vcc is 12V. With Rset being 14K, and the thermistor under test being 9 k to 14K, Vtest will be 4.7V to 6V. Thus, the values of the three resistors (Rlo, Rmid and Rup) for the two comparison voltages, VloLim (4.7V)and VhiLim (6V), are for you to determine, but should be under 100K so as supply sufficient current. If Vtest falls between 4.7V and 6V, then the lower comparator will produce a high out and the upper comparator will put out a low, thus lighting the green LED, indicating a pass condition. If below 4.7V, then the lower comparator puts out a low voltage, lighting the lower red LED, indicating the thermistor is below 9K. If higher than 14K, the upper comparator puts out a high, lighting the upper red LED. This is if you use comparators with direct TTL drive. If you use units with open collector outputs, then you'll need a pullup of about 470 - 560 Ohms at each comparator's output.
Of course, if your Vcc is other than 12V, these values will change. In that case, remember that the high limit voltage will be half of Vcc and the lower limit 40%.
Hope this helps,
kenjj

kjennejohn

Hi. I decided to expand on this design. I redid the diagram to show pullup resistors, pin numbers and their values. This is using the good ole LM339 comparator IC. This is available at RadioShack, Catalog #: 276-1712, $1.49, or DigiKey, P/N 497-1587-5-ND, 55 cents in single quantity. One data sheet for this is at:

www.st.com/stonline/products/literature/ds/2159/lm239.pdf

There are several reasons to use this particular chip. It's readily available, inexpensive, stupidly easy to use, operates with split voltage supplies (+/-1V to +/-18V) as well as single voltage supplies (+36V) and has an open collector output. The last two features make it suitable for use with several logic types, including ECL. The main advantage I see is this allows you to use this as a voltage level translator. That is, you can have sensors operating off 12V to 24V (typical) and present 0 to 3V/5V to your downstream logic or microprocessor. It all depends on what you connect your pullups to. There are scads of sites discussing the many uses for this chip.

That should do it. Happy Thanksgiving!
kenjj

uxorious

I think that will work perfectly. I have a Digikey order just about ready to submit, so I will add some of these parts. I'll post my results.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
~~ dRu ~~