annala@neuro.usc.edu (A J Annala) (01/30/89)
I am a neurobiologist with a theoretical exercise in electronic design that might prove to be very useful in reducing the cost of recording equipment used by the entire neuroscience community. I have a HP3964A four channel reel to reel tape deck. Two of the four channels are claimed to have a passband of 70-64,000 Hz (plus or minus 3 db). The problem is that I am recording data with from an electrode, amplifier, filter combination with a passband of 300-3,000 Hz. I want to be able to multiplex the recording of as many of these 300-3,000 Hz channels onto the 70-64,000 channel as possible. In the old days (with much smaller electrode frequency ranges [0-300 Hz we were able to encode three data channels on three separate FM carriers) as the input to one channel on a music quality stereo tape deck. I suspect the same kind of frequency division multiplexing (if that is the correct term) should be possible here. But I haven't a clue how to go about the parameters, design decisions, or implementation details for carrying out such an encoding scheme. I would appreciate any posted comments, email to me (annala%neuro.usc.edu@oberon.usc.edu), and/or US mail of details for how to build an encoder for my tape deck. Thanks, AJ Annala Psychology Department SGM-1011B / Mail Code 1061 Univ. of Southern California University Park Los Angeles, CA 90089-1061
hobbit@topaz.rutgers.edu ($ *Hobbit*) (01/30/89)
Funny that, I was just thinking about how I could record very *low*- frequency signals on my regular old cassette deck and not have the result sag all to hell because everything in the deck is AC coupled with a pitifully small time constant. I haven't started prototyping it yet, but I propose to use a PLL [XR2212 looks like a nice chip] with, oh, 6 to 8 KHz center freq for the VCO. The phase detector loop would be opened up for recording, and I'd just have the incoming signal sweep the VCO, and record the resulting FM on the tape. For playback, re-connect the feedback loop and the error voltage would be my demodulated output. The VCO would need some kind of trimmer to zero out any DC offset from a different tape deck with a different tape speed. I realize that this doesn't help answer the medical probe problem, but if some PLL wizard out there is going to address either one of these msgs, what Huge Caveats should the builder be aware of?? _H*
jbn@glacier.STANFORD.EDU (John B. Nagle) (01/31/89)
If you're grimly determined to do this by analog means, you might
check out some of the newer parts intended for compatibility with older
frequency-division multiplexing telephone carrier systems. These systems
are designed to compress voice channels, which are about the bandwidth
you need, into a bigger chunk of bandwidth. Motorola and Rockwell have
telecom products catalogs that might be helpful.
Then, of course, there's the approach of digitizing everything and
recording the result in a computer. There are data acquisition boards
available for this sort of thing from many sources. The VMEbus Buyer's
Guide shows various suitable boards with up to 64 input channels.
You could probably put together a working system just by buying and
plugging in VMEbus boards. Assuming that the goal is to get the data
into a computer for reduction, rather than generating miles of chart
recorder paper, this is probably the way to go.
It is possible to hook a data acquisition board to a PC, and you
could certaintly record a few channels that way, but you might have trouble
getting the bandwidth for a few dozen channels. But it's worth looking
into.
John Nagle
annala@neuro.usc.edu (A J Annala) (02/03/89)
I just learned from HP that they no longer manufacture the HP3964A instrumentation tape recorder mentioned in my original note. They called it "obsolete equipment" ... and the notes I received by the electronic mail suggest that almost no one records analog signals anymore ... instead current practice seems to be to get ahold of a VME bus (or faster) machine, do a quick analog to digital convert, and store the data to disk immediately. Unfortunately, I cannot afford to buy a VME (or faster) processor system ... therefore, I am stuck recording rapid signals on tape, reproducing those signals at slow speed, and converting them through my low bandwidth A-to-D converter board in my IBM PC. However, I just don't have enough channels on my 4 track HP3964A, and no one seems very knowledgable about multiplexing several signals onto one of those four channels, so ... I am curious about whether anyone out there in net land may have an HP3968A (or other relatively high speed recent vintage 8 channel instrumentation recorder) they might be willing to donate to USC for use in my project. If these things are truely antiques, and yet they are adequate for my use, perhaps such an arrangement might benefit both USC and the donating party. In any case, thanks for all your advice ... and, if anyone thinks about donating an HP3968A (or equivalent) to USC, thanks for your consideration. Thanks again, AJ Annala, USC Neuroscience Program (213) 743-3278.