1
\$\begingroup\$

I am trying to hack a scale so that I can read the data onto my Arduino. There are four piezo sensors, one at each foot, in the scale. Obviously I can't just read straight off of the sensors, so I bought some LM386N 400mW low Audio Voltage Amplifiers.

I spent some time online looking for simple instructions on using, alas I could not find any. Here is a picture of the simple setup:

Setup

EDIT: I just recently read that load cells at NOT piezo electric sensors (I was mistaken that they are). And in this case most home scales use load cells. So. Are the IC components I have correct? or do I need another part?

\$\endgroup\$
  • 1
    \$\begingroup\$ You forgot the power supply connections to your amplifiers ;o) \$\endgroup\$ – jippie Jul 7 '12 at 21:35
  • \$\begingroup\$ Well nothing is connected at all! Haha. \$\endgroup\$ – charles horvath Jul 7 '12 at 21:36
3
\$\begingroup\$

Audio amplifiers in general aren't appropriate for amplifying load cell (strain guage) outputs. This is because audio has a lower frequency limit. Most audio amps will block DC or expect their input to be AC coupled thru a capacitor. This is fine, even desirable, for audio, but not good for the very low frequency signals coming from strain guages.

It is possible that some "audio" amplifiers are DC coupled, especially if they are just chips. However, these may still be difficult to use as general instrumentation amplifiers because the DC offset was not a consideration in the audio application. It is probably fairly large and maybe not even specified because nobody would care when the device was used as intended. There may also be issues of input impedance.

What you want is a instrumentation amplifier or at least a precision opamp. Which one you need depends in part on how the sensor is arranged and what exactly its output signals are. Some are differential, in which case the instrumation amplifier would be best. These amplifiers have high input impedance, known fixed gain, and real differential front ends (meaning good common mode rejection).

We can probably be more specific if you explain more what exactly is coming out of your sensor.

\$\endgroup\$
  • \$\begingroup\$ So what I want is a low noise, low drift Op-Amp? \$\endgroup\$ – charles horvath Jul 7 '12 at 22:06
  • \$\begingroup\$ From what I read, these load cells output very small changes from 2 to 6 milivolts or so. Which the Arduino can not detect. I need it boosted to 1 to 5 volts. \$\endgroup\$ – charles horvath Jul 7 '12 at 22:09
  • \$\begingroup\$ @charles: I believe the 2-6 mV. However, is this ground-referenced? Differential? How is the sensor powered? How many wires go to the sensor in total? \$\endgroup\$ – Olin Lathrop Jul 7 '12 at 22:13
  • \$\begingroup\$ Oh, I linked a picture to the sensor. It has three wires, black, red and white. I assumed the black was GND, the red is 5V white is output. \$\endgroup\$ – charles horvath Jul 7 '12 at 22:14
2
\$\begingroup\$

I would also go for the instrumentation amplifier. The LM386 is only good as audio amplifier, and even then only if you want sound, without any questions about quality. From the 386's schematic I don't believe the DC gain Olin refers to will be the problem, but simply the gain. This is set at 20 and you can increase it to 200. That's it. But for the 200 the datasheet seems to want a capacitor coupling, so that won't work for DC. A gain of 20 is much too low.

Like also Olin says an instrumentation amplifier is the best solution, but a good opamp will do as well. You need one with low bias current, preferably FET inputs, and a low offset voltage. The MAX4208 has excellent figures with 1 pA bias current and 20 µV offset maximum, but isn't cheap. The MCP6061 is an opamp with also a low input bias current, but a bit higher offset voltage, but is much cheaper.

\$\endgroup\$

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.