I’m looking to use a pwm thermostat to drive a water pump and fan. I found the board to do what I need, it uses a 10k b3950 25/85 sensor. I need a sensor with a long lead and threaded with a probe for the cooling system. The sensors I’m finding are close but not exactly 3950. Id this going to make the readout and actual temp it is trying to maintain incorrect and is there a way to calculate by how much? I see a couple that are 3892 and 3977. It doesn’t really need to be dead nuts but I need to know where it is.
Here’s the boards sensor description.
Product Name: PWM 4-Wire Fan Temperature Controller
Work Voltage: DC 8V~60V
Control Board Working Current: 35mA(12V),22mA (24V)
Control Output Range: 10%~99.99%
Temperature Probe Specification: NTC 10KB=3950
Temperature Measurement: -9.9?~99.9? Acceleration
Temperature: 5?~94?,
Full Speed Temperature: 10?~99?
Speed Measurement: 10rpm~9990rpm
Stall Alarm Speed: <375rpm
Work Temperature: -25?~85?
Work Humidity: 5%~95%RH
Size: 64.564.515.5mm
I don’t know a ton besides what I’ve learned tonight and found this forum searching for an identical parameter threaded sensor.
Greetings,
The Wikipedia article on thermistors explains the associated mathematics nicely.
Running those equations for the three different beta values mentioned results in plots that are nearly overlapping, suggesting that you’ll probably see as much or more error attributable to initial tolerance than from minor beta variations. Calibration would be required in any case if measurement accuracy is of interest.
A product such as the PANW103395-395 would seem quite close to what’s described.
That’s great. Thanks so much. The other thing I was concerned about is adding lead length and its effects. I know that with a thermocouple there are very specific extensions.
Most folks don’t worry too much about lead extensions on thermistors.
Phenomena occurring at junctions of dissimilar metals are the entire Big Idea behind thermocouple measurement, so to randomly go about inserting chunks of coat hanger or whatnot is pretty much messing up the whole program there.
It’s a thing to be aware of with thermistor measurement if one’s going into the deep weeds trying to guarantee absolute accuracy to a fraction of a degree, but there are other potential error sources that a person needs to look at first in such cases.
Great. I kinda of assumed that but my whole reasoning behind this thread was to find out how much resistance affected this type of sensor/readout. It’s going to be used to control pwm brushless water pump and fan in a race car, so within 5° or something is more than enough accuracy for me. I currently use a k type thermo couple reader and EGT alarm box to trigger relays for my 2 brushed fans stepped a few degrees apart but I am building a new car with a fully electric brushless based cooling system. Being able to ramp them up and down with cooling requirements will make everything live longer and happier. Thanks for the input. If you can recommend anything more appropriate for my control board that would be great too. I would really like to be able to add a throttle position sensor to give full water pump power preemptively when past a certain throttle position before heat buildup begins too. I talked to one guy I know about possibly using a raspberry pi or audrino or something along those lines.
The wire resistance will have little effect on the accuracy up to about 500’ of 2 conductor 22 AWG (~ 17.5 ohms). However noise pickup on long wires can be a problem with such a high impedance sensor. Using shielded cable often helps on very long cable runs.
1 Like
