0
\$\begingroup\$

I have designed a PCB to read a strain gauge, in which I've used an INA317IDGKT instrumentation amplifier with a gain set to 200.

The strain gauge i'm using is the ATO-LCC-DYMH-103 100kg, supplied with 10V. According to the datasheet, I've connected EXC+ to +10V, EXC- to GND, SIG+ to IN+ (of the instrumentation amplifier) and SIG- to IN-.

First, I've supplied the gauge with 10V and measured with a multimeter between SIG+ and SIG-, and verified that the voltage difference reaches more than 11mV when applying force (which seems correct as the maximum difference is 15mV at 100kg and 10V supply).

However, when connecting the gauge to the amplifier, the voltage difference between SIG+ and SIG- stays fixed to 0.6mV and doesn't change when applying force. Because of this, the output of the amplifier stays at ~0.12V.

This is the schematic of the circuit:

Circuit schematic

In the inputs I put an optional Wheatstone bridge in case I wanted to use other types of differential sensors but I haven't soldered any resistor. To set a 200 gain value, I soldered a 500R resistor between Rg1 and Rg2.

I think the problem is in the amplifier, as I've tried desoldering it and the gauge starts working correctly again, but when I resoldered the amplifier it stopped.

Any ideas of why it's happening?

\$\endgroup\$
1
  • \$\begingroup\$ What voltage do meditere at the input pins of INA317? \$\endgroup\$ Commented Jun 20 at 17:56

2 Answers 2

Reset to default
1
\$\begingroup\$

Your circuit is not suitable for the task, as you are exceeding the common-mode input range for your inamp.

Leaving aside and difference in your positive and negative outputs of the load cell, both sides (SIG+ and SIG-) will be hovering around 5 volts, and you're powering your inamp with 3.3 volts. The allowable common mode input range is from (V–) + 0.1 to (V+) – 0.1 (from page 6 of your data sheet.

I can't say for sure why you're seeing the input difference not moving around, but I assume some transistor on your input stage is saturated.

As a debugging step, that inamp can take up to 7V supply. Try that, and see if it changes or resolves your issue.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Welp I didn't take common mode into account. Not such a bad error as I can change the gauge supply voltage. Thank you very much! \$\endgroup\$ Commented Jun 20 at 18:29
1
\$\begingroup\$

Connect the sensor excitation to +3.3V rather than +10V.

That will put the inputs in the optimal range for a 3.3VDC supply.

enter image description here

(3.3V will be similar just scaled, see the 1.8V curve).

It's possible the in-amp has been damaged by the high voltage applied to the inputs (Absolute Maximum is Vdd + 0.3V).

\$\endgroup\$
5
  • 1
    \$\begingroup\$ What could be the effect of reducing the voltage under the manufacturer's recommendation? The datasheet says that the excitation voltage should be over 5V \$\endgroup\$ Commented Jun 20 at 19:05
  • 1
    \$\begingroup\$ The effect is you will see increased offset errors (as % FS) due to amplifier offset voltage and thermal EMFs. Those particular errors will decrease as percentage of full scale as the excitation voltage increases. On the other hand, self-heating goes way down. Yours is a 'zero drift' amplifier so the effect of Vos is not very important. \$\endgroup\$ Commented Jun 20 at 19:08
  • \$\begingroup\$ P.S. noise will also be higher as a % FS, all other things being equal (they seldom are, and your 'zero-drift' op-amp has no 1/f noise). It's rare to find a zero-drift op-amp or instrumentation amplifier that can handle high voltage power supplies, but there are a few (and they are generally not cheap). \$\endgroup\$ Commented Jun 20 at 21:27
  • \$\begingroup\$ And, of course a decrease in sensitivity. \$\endgroup\$ Commented Jun 20 at 22:16
  • \$\begingroup\$ @ScottSeidman Yes. Kind of two sides of the same coin. If the gain is increased to get the same FS output voltage (to fill the codes of an ADC) then the sensitivity in terms of bits per milligram is unchanged. \$\endgroup\$ Commented Jun 20 at 22:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.