0
\$\begingroup\$

I'm designing a circuit that requires me to measure three low value (8-60 ohms) resistors with a hard wired common connection on one end. My ADC doesn't have enough inputs to measure all three at once along with everything else it needs to do.

My two questions are:

  1. What's the best method of current measurement? A Wheatstone bridge will draw way too much power, so I'm leaning towards a constant current source, and measuring the (amplified) voltage drop over the resistors. At 20mA (I don't want to go much higher as the resistors may warm up and change resistance) this gives me a voltage drop range of 0.16-1.2V.

  2. How should I switch between resistors? Automatic switching is required as this device needs to be a one-button operation - plug in sample, press button, receive information label. Is there a better option than two SPDT relays?

Thanks!

\$\endgroup\$
1
\$\begingroup\$

Get a dual 3:1 analog multiplexor. (Or 4:1)

Wire the first 4:1 to each of the 4 resistors, with the common node to Current Source.

Wore the 2nd 4:1 also to each of the 4 resistors, with the common node to ADC Vin+.

Tie the ADC Vin- to "GND". I presume you can implement a 10 milliohm Ground connection between the 3 resistors. The value of 10milliOhm * 20mA is an error of 200 microVolts out of 0.16 volt, or error of 0.16/0.0002 = 1/800 or 0.125% error due to GND plane/GND wiring resistance.

| improve this answer | |
\$\endgroup\$
0
\$\begingroup\$

Create a 20 mA current source and set up a mux with 2 control lines. Have your micro loop through the control lines 00 through 11 (for 4 total combinations, one of which you're not using) as often as necessary. Since the micro is giving out the control signals, you know which muxed input you're looking at at any given time. You can therefore read the value with the ADC for each resistor to which the current source is applied at the moment. Use high-precision, low-drift resistors.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ That's pretty much what I've got already. How would you build the mux - relays? \$\endgroup\$ – Joe of Loath Apr 20 '18 at 13:32
  • \$\begingroup\$ Why do you want to use relays? Your mux(es) would be doing all the switching. You wouldn't need any additional switching hardware as far as I can tell. \$\endgroup\$ – schadjo Apr 20 '18 at 13:43
  • \$\begingroup\$ I don't have any muxes, that's why I'm asking. All the muxes I've seen have non-zero internal resistance (mostly 50 ohms and 1K in some cases), which means they're absolutely no good for measuring with a constant current source. \$\endgroup\$ – Joe of Loath Apr 20 '18 at 14:23
  • \$\begingroup\$ "they're absolutely no good for measuring with a constant current source". That's simply not true. Think about it: all you care about is the current passing through the resistors you want to measure, to develop a voltage corresponding to temperature across said resistor. Parts with 0 internal resistance are ideal parts that don't exist. Consider a much smaller current (1-5mA) to generate less heat in your mux(es), and measure the voltage across your resistors. Even if the internal mux resistance is unknown, the constant current will give a repeatable voltage on your measurement resistors. \$\endgroup\$ – schadjo Apr 20 '18 at 15:41
  • \$\begingroup\$ To put @schadjo 's advice into other words, '4-terminal' resistance measurements work great through muxes. \$\endgroup\$ – Neil_UK Apr 20 '18 at 16:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.