Need a solution for connecting multiple microphones to PC

[repoman0]

For my electrical engineering capstone project. We're testing an ultrasound positioning system using six cheap electret microphones and a simple biasing/power circuit - the whole system will eventually be implemented on some sort of microcontroller.

But, in the meantime, we need some way to have at least four, preferably six analog audio inputs to a PC for testing our algorithms in MATLAB. 192kHz sampling of all inputs would be best (for good ultrasound reproduction). Does anyone have a solution? We can spend up to $300ish, less is better.

We will not be using XLR cables but regular 3.5mm jacks.

Thanks!

 

[videogames101]

http://www.amazon.com/M-Audio-Delta-.../dp/B000085ZKX

you might need some adapters, but 10 inputs will be hard to find in one solution anywhere else.

edit:although it seems it's limited to 96khz recording, not sure if thats a dealbreaker

 

[repoman0]

Certainly looks good - 96kHz will probably limit us to under 40kHz when the final result should be narrowband at 45kHz but looks fine for testing. Thanks, I'll update if we get it and test!

 

[corkyg]

Sounds like a mixer board for the mikes - and one connection from it to the PC.

 

[Modelworks]

You will need to amp those electret mics for most cards to be able to accept the signal. Not many cards have 6 mic inputs, usually they are line level.

I would put a jfet opamp like the lm358 connected to the mics then output that to line level on the sound cards. I also use this setup for input to a dsp chip for sampling so it is something you can use on the final project too.

Circuit I use for the mics
http://circuitsfun.blogspot.com/2011/08/simple-lm358-mic-preamplifier.html

 

[repoman0]

Right, I realize that the power circuit I built already will need to be revised so that it can bring the little mics up to line level. The Delta 1010 linked above has exactly 6 RCA line inputs.. perfect! I'd found this earlier - http://www.epanorama.net/circuits/micamp.html - not sure if the LM358 that you linked or the BJT amp will have better quality. We'll have to try both!

Mixer was my first thought as well but they are very expensive even for 2 inputs, and use XLR / 48V inputs which we don't want to deal with.. line in is way easier, and better transfers to DSP and microcontroller later.

 

[vbuggy]

RME Octamic II will do what you need. A reasonably priced multi-mic 192Khz ADC.

EDIT: Oops, scratch the reco - I didn't see the $300 limit.

 

[Modelworks]

I'd found this earlier - http://www.epanorama.net/circuits/micamp.html - not sure if the LM358 that you linked or the BJT amp will have better quality.

Use the LM358 or even better the same circuit as the LM358 but use a TL064 opamp . It has quad inputs and very low noise. Both are Jfet based so quality is good. Using BJT the problem you will have is the noise floor is much higher. I can connect a single mic to the opamp and change out the pot for a 1.2M resistor, giving it 100,000 gain. The mic is so sensitive that I can hear a whisper at 10 feet from the mic. Paper tearing sounds like sand paper at the same distance. Not sure you want that high a gain but opamp is the way to go.
Price is also cheap, just 68 cents
http://search.digikey.com/us/en/products/TL064CN/296-1773-5-ND/277419

I feed that into a dsp chip and process it with filters to lock onto voices in the distance.

 

[repoman0]

Awesome, I really appreciate the suggestions. Will definitely buy some of the TL064, I'm about to place an order with digikey for the parts to build a bunch of those amplifier circuits that you linked.

We don't need a super high gain; the ultrasound noise will come from PWM power regulation of BLDC motors using switching freq above 20kHz and it seems pretty "loud" .. not that we can hear it!

 

[repoman0]

All electronics and the Delta1010 are in, now to get to work on real time hardware access with matlab