micro Volts/Ohms to a Micro or other ADC

[tequiler]

Hello All!

I'm trying to make my own soft sensor system, utilizing a silicone tube with an Eustatic Gallium-Indium alloy. My end goal is to be able to measure the variations of microVolts and/or microOhms when applying force to or when the sensor is deformed with an ADC on a microcontroller.

I'm having trouble with the circuit I've designed. I went with: a current source based on a voltage regulator to force a voltage drop over the element, and then use a 3 Op-Amp "instrumentation" Amplifier to amplify the voltage difference between the two terminals of the element. (Partially inspired by the image attached)
I've also attached my circuit diagram and prototype for reference.

My circuit, despite being configured to have a gain of ~200 v/v, doesn't seem vary it's output, just floats around 1.83 V, even when changing the load to something a bit higher.

I've been thinking about switching out my 3 Op-Amp In-Amp for an integrated one, but with shipping time and cost... I'm a bit hesitant, unless this switch would help?

Am I forgetting something else or overlooking something obvious?
Is there an IC, breakout, Microcontroller that already does this?
Any help would be appreciated!

[Kleinstein]

With the 3 OP instrumentation amplifier one has to keep an eye on the common mode ranges. Especially precision OPs are often not rail to rail.

The breadboard circuit does not have decoupling caps - depending on the OPs used it may oscillate. What OPs are used ?

When using a relatively high gain the advantage of integrated INAs is not that large - the main need for precision matched resistors is with lower gain.

The LM317 is specified for 5 mA minimum current, so quite a bit more than the 500 µA indicated in the plan.

With the sensor in the current form the sensitivity to small changes may not be very large: when one displaces metal from one spot it will add to the rest and compensator part of the increasing resistance. Ideally the excess metal should go to an area outside the sensor part or at least to an intentionally thicker region.  As shown the characteristic may be more quadratic and thus low sensitivity to small changes.

[tequiler]

Thanks for the reply Kleinstein!

I'm using 3 UA741Cs, supposedly they should work at 0v and 5 V supply, no?

The current I'm sourcing from the voltage regulator current source is a bit higher, as to increase the voltage over the sensor, at least that was my logic.

I hadn't thought about your comment about the sensor's design, it makes perfect sense. For now though, if I was to apply enough pressure to decrease the cross-section small enough I should see a decrease in resistance, shouldn't I? (At least something measurable, right?)

Best Regards!

[Kleinstein]

The sensor would show a measurable effect for stron pressure. So for a first test it is OK.

The UA741 is not at all suitable: it does not work correct with 5 V - I would consider it more like 8 V minimum. For some reason not all DS show a min supply.  The common mode  range is also a problem : the guaranteed limit is some 3 V from the rails - kind of impossible to meet with a 5 V supply  .

To at least make a work a little, a LM358 would be the better choice. Precision is still poor.
Noise and drift wise the 741 is also not that great for a small signal.

[ejeffrey]

Your schematic has at least two errors, you should check them against your circuit.

First, the lower input opamp is reversed.  The lower input signal should got to the + input, the - input should go to the feedback resistors.

Second, the differential amplifier stage is inverting. The way you have connected it the output voltage will be negative and not suitable for single supply operation.

[tequiler]

Thanks all for the great advice and help Everyone!

For future reference, I managed to get it working by opting for a single In-Amp IC (AD620) instead of the 3 Op-Amp In-Amp (having tried it with UA741, OP07 and LM358).

[radiolistener]

I think you're needs to put your reference Voltage for sensor as a reference for ADC, to use relative measurements and avoid measurement errors due to reference source drift. Also you will need to measure temperature of your ADC and op-amps and use temperature calibration to compensate temperature drift.

I'm not sure but op-amps looks here as a most weak components. I would remove them at all and use 24-bit delta-sigma ADC directly. With 5V VREF it will give you LSB weight about 0.298 uV. If you want to get better resolution, you can apply LPF filter.

I have some experience with precise pressure measurement and can say that the most complicated part will be to eliminate temperature drift of all components. In my case it was even more complicated because working temperature range of the device was -60..+80 °C and temperature drift was very siginficant.

[Zero999]

For some reason not all DS show a min supply.

Because it's application specific. The minimum operating voltage is the point when the common mode range, and/or output swing become too narrow to be of any use.