2

I have a strain gauge with maximum output of 5mV. I want to amplify this signal 1000 times in order to use my DAQ system to read it (the analog input range is 0-5V for DAQ connector).

I have purchased a MAX9613EVKIT+ (schematic below). The default gain factor is 10, but I want to change the R5 and RC3 from 10k ohm to 1 Mohm to get a GF of 1000. The thing that I noticed is the input impedance for the analog input channels is between 500k and 1M. If I use a 1M feedback resistor, and assume 1M for connector, will I have 50% error on my gain? Or is only the R6 important for DAQ system?

MAX9613EVKIT+ schematic

In other words, is it OK to use R5=RC3=1 M ohm? What is going to be the error in the gain?

SamGibson
  • 17,231
  • 5
  • 37
  • 58
Arash
  • 23
  • 4
  • The feedback resistor, R5, will have no meaningful effect on the output impedance of the op amp. R6 will set the effective output impedance. – lakeweb Feb 21 '18 at 00:50
  • Welcome to EE.SE. This maybe a low quality op-amp, which loads down the feedback resistors. For 1 meg ohm resistors you need a JFET or CMOS op-amp. Try a CA3140T if you can find one. –  Feb 21 '18 at 00:54
  • Hi @Sparky256, the [MAX9615](https://datasheets.maximintegrated.com/en/ds/MAX9613-MAX9615.pdf) biases at like 1.5 pA. I think he will be ok. – lakeweb Feb 21 '18 at 00:56
  • @lakeweb. Ok fine. The CA3140T bias is .5 pA, so little difference. –  Feb 21 '18 at 02:49

1 Answers1

2

I have a strain gauge with maximum output of 5mV

The MAX9613 is a poor choice for amplifying a strain gauge (short story is use an instrumentation amplifier) because it will place a resistive load on the output of the strain gauge (a resistor) and dramatically alter its output.

Irrespective of the feedback resistor, the dynamic impedance looking into R1 (which connects to a virtual earth point) is purely R1. This will produce bad results when you are trying to measure a signal that has an output impedance in the same realm as R1 (1 kohm). Additionally, the MAX9613 has an input offset voltage of 100 uV and so this represents an error of +/- 2%.

You will also get mismatch errors because of tolerances of all your resistors. Try setting up a simulation and alter one value by 1% to see what I mean.

Use an instrumentation amplifier is my advice.

Andy aka
  • 434,556
  • 28
  • 351
  • 777
  • @ Andy I don't have enough space to be able to use a signal conditioner like "Dmd4059-Dc " for strain gauge, so I am looking for a small circuit board that I can use. I googled and found "instrumentation amplifier AD620 " [link](https://www.aliexpress.com/item/instrumentation-amplifier-AD620-Voltage-Differential-MV-microvolt-signal/32763760871.html) Do you think it would work properly for my case? – Arash Feb 21 '18 at 17:27
  • @Arash what power supplies are you using/have available and what is the configuration you have for the strain gauge i.e. bridge, single element with resistor feed or maybe a constant current feed? – Andy aka Feb 21 '18 at 17:41
  • @ Andy I am using a battery which can provide +-12 and 5 V. I am using half bridge completion module with two strain gauges. – Arash Feb 21 '18 at 17:55
  • @Arash I think the 620 would be much better but the devil is in the details so read the DS however, a quick look tells me that the input error voltage is 3 times smaller typically. There are more accurate ones like the AD8221 (25 uV max input offset error) and I can recommend that on the basis of using it many, many times. – Andy aka Feb 21 '18 at 18:01
  • Yeah good choice hehe – Andy aka Feb 21 '18 at 19:52
  • @ Andy I am trying to make a PCB using AD8221. In the data sheet [here](http://www.analog.com/media/en/technical-documentation/data-sheets/AD8221.pdf) in Figure 50, you can see the diagram. This is for quarter bridge, but if I want to use 2 strain gauges, do I still need to connect the Ref on op-amp to 2.5 V battery? Just want to make sure that everything is correct. Thank you! – Arash Feb 21 '18 at 22:35
  • You connect ref to a voltage that represents the “neutral position” i.e. the voltage you want to be on the output when zero strain (or neutral strain) is applied. If that is 2.5 volts then so be it. Btw you can up vote answers and, if you get an answer that completely answers your question you can formally accept that answer. – Andy aka Feb 22 '18 at 00:13
  • @ Andy The last question: The maximum output that I'll get from strain gauge is 5mV. I was wondering to have a gain of 1000 to bring it to 5V. Based on what you said about zero strain (voltage), the gain of 1000 is not going to work (because I will shift it too) and I need to have a gain of 500 (so I have 2.5V, plus the 2.5 that I'll shift). Right? – Arash Feb 22 '18 at 17:26
  • Maybe you can draw a pic of what you mean else just try it and see. – Andy aka Feb 22 '18 at 17:33