Measuring 10mV analogue signal

polus

Sep 16, 2004
20
Joined
Sep 16, 2004
Messages
20
I have a varying analogue signal of about 10mV. Does anyone have any clever ways of the most accurate way of measuring this (It has to be extremly accurate)? I have experience in programming PIC

 

audioguru2

Apr 6, 2004
12,026
Joined
Apr 6, 2004
Messages
12,026
What frequency range?
What has to be accurate? Voltage level? Low distortion? Low noise?

 
Last edited by a moderator:

polus

Sep 16, 2004
20
Joined
Sep 16, 2004
Messages
20
Frequency range... hmmm I really bad with analogue electronics, I dont really know.

The voltage level has to be accurate. I guess if there was noise before the conversion then that would alter the accuracy of the digital output?

Sorry for my lame-ness.

 

MP1

Dec 7, 2003
3,399
Joined
Dec 7, 2003
Messages
3,399
You could make your own ADC with comparators. The more you string in line, the more resolution you will get. Well, up to a limit, of course.
But then, since you are familiar with micros, why not just choose one which has a built-in ADC? 10 bit ADC is pretty common.

MP

 

Kevin Weddle

Feb 23, 2004
1,620
Joined
Feb 23, 2004
Messages
1,620
A 10mV signal is small, but not the smallest. After you amplify, you must know that what you are getting is multiplied representation. The gain can be difficult to accurately guess. So when you finally divide the result to find the actual value you will have errors in your findings. This is the problem with amplification, you can't be exactly sure of the original.

 

MP1

Dec 7, 2003
3,399
Joined
Dec 7, 2003
Messages
3,399
Kevin, if he uses a micro with a 10 bit ADC, he will not have to divide the signal back down again. For example, he could amplify the signal to 1 volt, feed it into the ADC of the micro, and the numeric representation would be what he would use. This could then be formatted and sent to a display as xx.x volts, etc.

MP

 

polus

Sep 16, 2004
20
Joined
Sep 16, 2004
Messages
20
I was thinking of using a 16-bit ADC on an analogue devices 8051 (I have a couple of samples... :) )

Anyway for the application accurately amplifying this sort of signal what kind of amplifier would you guys recommend?

To be honest one of the reason Im doing this is I really want to know more about analogue electronics and expand my digital knowledge because Im a software nut and I enjoy messing around with electronics. Ive done all sorts of high level programming in C++ and some embedded in C but my knowledge really falls short at the electronic design level (something Im hoping to change).

Thanks for your ideas, even the simplest of advice helps.

 
Last edited by a moderator:

gg4rest

Feb 11, 2005
46
Joined
Feb 11, 2005
Messages
46
The first question is what is the frequency of the signal you are measuring? How fast is it typically changing? You need to know this because you will have to handle it with filtering otherwise you will get aliasing when you sample it. The second question is how fast do you want to sample it? Do you want a plot of the signal or simply a digital readout of the signal that changes every second or so? Is your input signal bipolar (positive and negative)?

To amplify the signal you should use an instrumentation amplifier. You can get them contained in an IC or you can make one yourself out of operational amplifiers (opamps). It consists of a standard difference amplifier buffered at the front end by two more opamps. You might want to make your own if you are looking to learn analog electronics. Make sure the bandwidth of the opamps are sufficiently high.

You should select the gain of the amplifier to take advantage of the full swing of your ADC. For example, if the maximum input signal is 10 mV and your ADC range is from 0 to 10 V, you want a gain of around 1000. That way you won't have very large quantization noise.

Once you have the amplification, you need to address the questions I brought up in paragraph 1. I'll be glad to help you answer them once I know more detail of the input signal.

 

polus

Sep 16, 2004
20
Joined
Sep 16, 2004
Messages
20
gg4rest, thanks for the advice. Im just going through everything and then will get back to you.

(Didnt want to just leave this thread hanging)

 
Top