Hi People, quick question: could I use small Op-Amps to boost the Voltage output of my electrodes for my EEG setup? By that I mean, could I attach a small IC Op-Amp to an Electrode and have that boost the voltage instead of using some fancy, expensive multi-channeled amplifier? So, I would have an individual IC Op-Amp per electrode. Thanks!

 

Take a look at Intersil's AN1298: http://www.intersil.com/content/dam/Intersil/documents/an12/an1298.pdf
(free Adobe Reader required)
In a nutshell: no. You won't be able to attain anywhere near the performance of a real instrumentation amplifier when trying to make your own.

 

Take a look at Intersil's AN1298: http://www.intersil.com/content/dam/Intersil/documents/an12/an1298.pdf
(free Adobe Reader required)
In a nutshell: no. You won't be able to attain anywhere near the performance of a real instrumentation amplifier when trying to make your own.

Ok thanks! Can you get cheap-ish IC instrumentation amplifiers?

 

Just visit any major electronics distributor's site and search for 'instrumentation amplifier'. You will find lots to choose from (depending on what you consider 'cheap' ).

 

Just visit any major electronics distributor's site and search for 'instrumentation amplifier'. You will find lots to choose from (depending on what you consider 'cheap' ).

Great thanks! What sort of specs should I be looking for? Considering I need as accurate EEG data as possible. Thanks!

 

Just a guess from me, as I come from a software background ->
its the really the 'microphone' which is the only important bit to the hardware, if your starting from the same raw signal as them, digital filters (even tho making it a fully brittle data process -especially if you used a FFT, thats the brittlest you can get!!!) should be able to outperform analogue filters, and do the whole thing with programming - if you wanted to compete as a performance product, and the cheapness that goes with software.

I saw an eeg setup and it was just a simple band pass filter, what if theres something better you can do than that easily...

 

I recall the INA126 is used for EEG aquisition.

Don't know whether that helps you?

 

Just a guess from me, as I come from a software background ->
its the really the 'microphone' which is the only important bit to the hardware, if your starting from the same raw signal as them, digital filters (even tho making it a fully brittle data process -especially if you used a FFT, thats the brittlest you can get!!!) should be able to outperform analogue filters, and do the whole thing with programming - if you wanted to compete as a performance product, and the cheapness that goes with software.

I saw an eeg setup and it was just a simple band pass filter, what if theres something better you can do than that easily...

What do you mean?

 

I recall the INA126 is used for EEG aquisition.

Don't know whether that helps you?

That's pretty helpful thanks! Do you know if the 2-channeled version (INA2126) has any differences other than the fact it's 2-channeled? Also, I like your ECG profile picture

 

As a general rule when working with sensitive signals I stay away from multiple amps in a package, at least at the first amplification stage.

 

I need as accurate EEG data as possible.

That doesn't equate to 'cheap' . I'd be looking for an amp with minimum input offset and temperature-drift (but I've not worked with instrumentation amps; other members here may know of more important features?).

 

The most important spec for an EEG instrument amp would be the CMRR, Common Mode Rejection Ratio.
You have two long leads coming from the patient to the inputs of the in-amp. There will be a large common mode signal which you really want to reject.

 

they put homomorphic filtering on xrays and it brings them out better, theres got to be some "contrasting noise reduction" equivilent for a 1d signal.

 

The most important spec for an EEG instrument amp would be the CMRR, Common Mode Rejection Ratio.
You have two long leads coming from the patient to the inputs of the in-amp. There will be a large common mode signal which you really want to reject.

So the longer the wire from the electrode the more noise created by it?

 

Yes, long wires act as noise receiving antennae, but it is nothing that hasn't been working for a few decades. Two things about EEG signal processing. 1. It's been around for almost 100 years, although the first 50 really didn't count, so the requirements are very well known; there are no magical secrets to success. 2nd, while it is very well known, that does not equate to cheap. The signals are not very fast, but they are tiny and need careful attention. To get signals as reliable as those in pro equipment, you have to do what the pros do: a high-quality instrumentation amp with some form of DC stabilization, an adjustable bandpass filter, low noise power supplies, above-average power supply decoupling at each chip, and careful attention to pc board layout techniques.

Patient-contact electronic devices have very large liability insurance costs that contribute to their high price, so making EEG amps is much cheaper than buying them as long as you accept the risks. But there is more to it than a few opamps on some perf board. What bandwidth are you interested in?

ak

 

Yes, long wires act as noise receiving antennae, but it is nothing that hasn't been working for a few decades. Two things about EEG signal processing. 1. It's been around for almost 100 years, although the first 50 really didn't count, so the requirements are very well known; there are no magical secrets to success. 2nd, while it is very well known, that does not equate to cheap. The signals are not very fast, but they are tiny and need careful attention. To get signals as reliable as those in pro equipment, you have to do what the pros do: a high-quality instrumentation amp with some form of DC stabilization, an adjustable bandpass filter, low noise power supplies, above-average power supply decoupling at each chip, and careful attention to pc board layout techniques.

Patient-contact electronic devices have very large liability insurance costs that contribute to their high price, so making EEG amps is much cheaper than buying them as long as you accept the risks. But there is more to it than a few opamps on some perf board. What bandwidth are you interested in?

ak

Mainly Alpha thanks. Also, can you recommend any amps to purchase or any circuit designs for amps?

 

I put "'eeg preamp schematic' and 'eeg amplifier circuit' (without quotes) into Yahoo and got a bunch.

Let's be clear here - patient contact electronic circuits are dangerous, but as long as they're not in a car...

ak

 

I always having dreams about doing armed robbery eeg'ing to your robot body, and making the next generations internet cafes.