Equipment for testing microvolt Instrumentation Amplifier

Thread Starter

elkyn.h.sanabria

Joined Dec 22, 2017
5
Hello all,

I'm just beginning my formation as analog microelectronics designer. I've just finished my first relevant IC project: An integrated CMOS Instrumentation Amplifier for biomedical applications (now in fabrication). There are some postlayout simulated specifications for the designed prototype:

Differential gain: 25 or 50 V/V (fixed)
CMRR: 100dB
Differential voltage input range: 80uV to 7mV (rms)
Bandwidth: 2MHz
Noise referred to input (for BW=2MHz): 16uV (rms)

Now I'm trying to write the test plan for the circuit. The test PCB and the tests to do on the circuit include measuring gain, CMRR, PSRR, BW, noise, THD, slew rate, etc. are in planning process, but my problem is about the equipment required to perform this tests (waveform generator, signal and noise measurement), since typical function generators and oscilloscopes are not fit to this purpose, due to the tiny amplitude of the voltage input range and noise. One can think that input signals can be generated by a typical function generator followed by a resistive divider, but this may be not a good idea due to the thermal noise contribution of resistors, which would "bury" the input signal into noise.

So, if any of you can share your knowledge about the test and measurement equipment suited for this tests (brands and references) I will be very grateful. I have no experience with special testing equipment, so the internet research I've made so far left me quite confused... :(

Thank you for your help!
 

GopherT

Joined Nov 23, 2012
8,012
Hello all,

I'm just beginning my formation as analog microelectronics designer. I've just finished my first relevant IC project: An integrated CMOS Instrumentation Amplifier for biomedical applications (now in fabrication). There are some postlayout simulated specifications for the designed prototype:

Differential gain: 25 or 50 V/V (fixed)
CMRR: 100dB
Differential voltage input range: 80uV to 7mV (rms)
Bandwidth: 2MHz
Noise referred to input (for BW=2MHz): 16uV (rms)

Now I'm trying to write the test plan for the circuit. The test PCB and the tests to do on the circuit include measuring gain, CMRR, PSRR, BW, noise, THD, slew rate, etc. are in planning process, but my problem is about the equipment required to perform this tests (waveform generator, signal and noise measurement), since typical function generators and oscilloscopes are not fit to this purpose, due to the tiny amplitude of the voltage input range and noise. One can think that input signals can be generated by a typical function generator followed by a resistive divider, but this may be not a good idea due to the thermal noise contribution of resistors, which would "bury" the input signal into noise.

So, if any of you can share your knowledge about the test and measurement equipment suited for this tests (brands and references) I will be very grateful. I have no experience with special testing equipment, so the internet research I've made so far left me quite confused... :(

Thank you for your help!

That is like saying, I just built a bench for my garden and now I am building a bridge for the motorway - can someone please tell me how I can test my bridge?

What level of testing do you want to do? Do you want to test your amplifier circuit with a simple capacitively coupled signal or do you want to test it in the biomedical environment you designed it for? If so, what kind of electrodes? What frequency (application) do you have in mind?
 

Thread Starter

elkyn.h.sanabria

Joined Dec 22, 2017
5
That is like saying, I just built a bench for my garden and now I am building a bridge for the motorway - can someone please tell me how I can test my bridge?

What level of testing do you want to do? Do you want to test your amplifier circuit with a simple capacitively coupled signal or do you want to test it in the biomedical environment you designed it for? If so, what kind of electrodes? What frequency (application) do you have in mind?
 

Thread Starter

elkyn.h.sanabria

Joined Dec 22, 2017
5
That is like saying, I just built a bench for my garden and now I am building a bridge for the motorway - can someone please tell me how I can test my bridge?

What level of testing do you want to do? Do you want to test your amplifier circuit with a simple capacitively coupled signal or do you want to test it in the biomedical environment you designed it for? If so, what kind of electrodes? What frequency (application) do you have in mind?
Thank you for your response.

Ok, let me explain a little further. This IA was built to amplify bioimpedance signals coming from in vivo cervical tissue (yeah, now you know what it implies) so is impossible for now to get the signal from its application environment. I've already planned how to perform the gain, CMRR, BW and the other tests, but I have no knowledge about the hardware I need to generate the input signals and to measure the output. Basically I just need a sinusoidal signal from 100 Hz to 2 GHz and the amplitude range I mentioned above, that -yeah- can be feed to the input capacitively. About the output, I must say that its intended load is just the gate of a transistor of a subsequent stage, but for this testing purposes it is gonna be buffered to a low impedance connector.

Lastly, I must say sorry if I was not quite clear stating my problem... English is not my first language. Hope you understand.
 

GopherT

Joined Nov 23, 2012
8,012
I am guessing that you have arbitrarily decided you WANT to monitor up to 2GHz but that decision just made everything enormously difficult and unnecessary. I don't know if any biological processes that can generate GHz (or even MHz) signals.
 

Thread Starter

elkyn.h.sanabria

Joined Dec 22, 2017
5
I am guessing that you have arbitrarily decided you WANT to monitor up to 2GHz but that decision just made everything enormously difficult and unnecessary. I don't know if any biological processes that can generate GHz (or even MHz) signals.
Sorry! I mean 2 MHz...
Yeah man... Is absolutely necessary to reach at least 1 MHz to see the bioimpedance contrast between a healthy and a displasic cervical tissue. A paper about my research has just been published...

http://ieeexplore.ieee.org/document/8229506/

Thank you again for your interest.
 

Thread Starter

elkyn.h.sanabria

Joined Dec 22, 2017
5
Tough one. really tough problem.

My initial thought is a variable attenuator such as: http://weinschel-catalog.apitech.com/item/continuously-variable-attenuators/nuously-variable-attenuators-n-sma-dc-2-4-ghz-5-w/940-60-34

That one assumes 50 ohms. Food for though, for now.
Thank you man!

I think this device definitely could help to attenuate the signal generated by a typical waveform generator in the bandwidth of interest, without growing much noise.
 
The attenuator likely assumes a transmission line. i.e. The system is driven by 50 ohms impedance and it expects to be terminated in 50 ohms.

Higher frequency scopes may be 50 and 1 M switchable. RF Signal generators will likely have a 50 ohm output. Audio would likely be able to drive 600 ohms.

I said your not out of the woods, but it becomes a place to start.
 

Janis59

Joined Aug 21, 2017
1,187
Sometimes I have a deal around such `tablets` like your`s. The best method to understand do it works is to bring one AAA battery and construct a voltage divider around it. This will be the floating DC microvolt generator. Other similar may be applied with a larger, ie milivolt scale `troublemaker` floating source. Then use a laptop based digital oscilloscope what will tell what there happened while You shorted one or other source.
 
Top