From: sleeper75se_at_yahoo.se
Date: 2001-11-28 16:25:27
Hi Moritz,
my computer doesn't have an IrDA port. What do I do? :-) But the
Microchip device looks perfect for the job if you want wireless
communication. It certainly feels a lot safer than a opto-coupler.
But I think IrDA is for the advanced systems and optical fiber for
the simple ones!
I've done a lot of thinking about this (I'm home on sick-leave today)
so please let me tell you what I came up with:
The MCP2150 costs $5.70 in quantities of 100 or more, at Arrow
electronics you pay $6 for single items. I just found another device
that might work though, HSDL 7000. It is not as capable as the
Microchip IC, but it is much cheaper, you can get it for about $2.90.
It is specified for use in medical applications and can be connected
directly to an UART, though it needs a baud-rate clock (baud-rate x
16) to work. To that, you must add the cost of the tranceiver
LED/photo-diode. HSDL 1001 might be a good choice for $4 at Arrow
electronics.
Datasheets can be found here
http://literature.agilent.com/litweb/pdf/5965-5363E.pdf
and here http://literature.agilent.com/litweb/pdf/5964-9278E.pdf
So the IrDA solution is $6.90 with Agilent and $9.70 with Microchip
IC's, not counting the cost for adding IrDA support if your PC does
not have it (another $7-$10 if you use the same setup on the PC-side).
One note: A HSDL1001 eats 100mA of current, so it may not be the
optimal solution for a battery operated device.
Now, does communication with the EEG-device really need to be
bidirectional?
You mentioned calibration as the main reason.
I'm assuming here that we want to estimate the frequency response by
feeding a signal of known amplitude but with selectable frequency
into the system, measuring what's coming out at the other end?
What do you think about using a switch that sets the acquisition unit
in calibration mode?
The software would simply ask the user to turn it on at regular
intervals (perhaps once a week or every 10 sessions). When the switch
is turned on, the microcontroller would start the calibration signal
and repetedly let it sweep over 0.1 to 100Hz or so, until the PC
says it is done measuring. That way, it would be possible to drop one
data path.
Looking back at the fiber-solution, the Toslink transmitter is just a
LED with a driver circuit. For the speeds we are looking at
(115200bps or less) you can replace it with a simple flat-top LED and
the cost for the transmitter drops to 20 cents. Put together with the
transistors and resistors (about $.20 each, total of $1.5) the photo
diode ($.50) DSUB connector ($1.50) and nylon fiber ($2.50/meter) you
get $3.70 or more, depending on the length of the cable.
I know you guys have been working on this for a long time, so I'm
definitely not the one to tell you what to do or not. I must admit
the difference is not as large as I thought it would be, and in a
$100 project, $3 may not matter much. On the other hand, it is
interesting to discuss different solutions. :-)
Be well,
Andreas
This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:32 BST