At 02:09 PM 6/24/99 -0700, you wrote:
Electrodes are the real problem... most contain an adhesive of some sort...
could come off, 
likely will have an allergenic reaction over 6 to 9 hours of sleep and most
certainly will over the long run- 1 to 2 weeks,  Don Lancaster had some
articles in the late 60's to mid 70's as I remember about "lie detectors,
GSR devices. There has been a plethora of articles about this subject...
maybe Circuit Cellar, Likely Nuts and Volts. The easiest electrode is an
anti-static grounding strap, Has a snap on connector, this can be sensed
and alarm if it becomes disconnected might be possible to use a conductive
screen under the sheet, place the strap on an ankle, possibly the greatest
degree of freedom, Finger straps were the primary electrodes for GSR
devices. Some random thoughts... sweating usually involves a change in
temperature... very easy to sense, use an LM50 from National Semiconductor
get it from Digikey. use an LM339 as a window detector (2 sections) use the
remaining 2 comparators 
to drive a piezoelectric sounder. if some rudimentary attempts at low power
operation are attempted (a max 667 5V regulator has a 15 ľa quiesent
current) the device could last several weeks from a 9V battery. If
temperature sensing is not usable then configure the lm339 as a
multivibrator using the skin resistance as one variable... feed the output
to a high pass filter (as the skin sweats its resistance gets lower AND
since T=R*C the frequency will go up as gsr goes down and rectify the
filter output connect a section of the LM339 as a comparator and the
remaining section(s) as some type of output device. The '339 is capable of
operation up to a mhz or more so a broadcast band radio with a severely
crippled antenna could be used as a wireless detector. with a megohm of
series resistance 5 ľa of skin current will flow this is WELL below the
threshold required feel anything... I don't know what your electronics
expertise is, I could send you some partial circuits and I would be very
happy to do so, however I do not have the time to develop any kind of
finished device without some type of renumeration. This having been said if
you take a look in some of the alternative medicine magazines you just
might find what you are looking for. I can think of several ways a PIC
processor (even a 'c54) could be used to do all of the above including send
a signal to a radio (estimated parts cost of 6 to 10 dollars)


        Robert K. Johnson
        rkj1@ix.netcom.com

>Many magazines have had reasonable quality schematics for "Stress
>detectors" etc., IIRC Popular Electronics had a schematic for one such a
>while back.
>
>  Lots of issues to be aware of here:
>
>  This is a Life-Critical type system - you want to make sure this thing
>works & works well & reliably & doesn't have lots of failure modes.
>
>  For one:  You want either a WELL-isolated power supply, if you're
>going to plug this unit into AC power (And I don't mean a typical "Wall
>Wart", I mean something that's certified for medical use!) - OR, you
>want to use a deep-cycle lead-acid battery (probably a better choice!)
>like a golf cart or marine trolling battery as a power source, due to
>the possible danger of electric shock.  (Some power supplies have been
>known to have a catastrophic failure mode where when they fail, they
>output 117VAC to the Vcc and ground output leads - The Dauphin DTR-1
>laptop is one such - you don't want a family member wired up to this
>with a taped-on electrode, then shocked while asleep, rather
>obviously!)  I'd lean towards a nice deep cycle lead-acid battery, with
>a few days extra capacity if possible (should be!) for probably much
>less than an isolated supply.  And charge it every other day or
>something, but only when not attached to the family member.  Safe
>isolated power units may be cheaper than last I worked with Spacelabs,
>though.
>
>  You'll need to find a good source of the electrodes (The only source I
>know of, is those off of one web page that I cannot find right now -
>it's about using muscle firing rates to input signals to your computer
>instead of a keyboard, I cannot remember the acronym they were using
>right now, they use some reasonably cheap electrodes, I do remember.)
>It'll be in Alternate input methods off http://wearables.org/ someplace,
>bug me if you cannot find it.  Aaah.  SMUT, that's the acronym.
>Something or other, Motor Unit Training.  That's about watching motor
>neurons cycle rates etc., not skin galvanic response, but I suspect the
>electrodes are transferable.
>
>  How do you run wires from the family member to the device?
>Alternately, could use wireless & strap a small unit to the family
>member, battery powered, but then you want some sort of regular "ping"
>from the unit to the receiver (you want to know that the sender's quit
>working!), could be done with IR or RF.  Either way has problems &
>advantages.
>
>  What sort of alarm do you want?  How do you calibrate for summer
>versus winter? (Want to prevent false triggers from too much blankets,
>or those really hot nights.)  Might use 2 sets of electrodes or
>something?
>
>  I'm a little distracted (getting set to do job search here), so I
>may've missed any number of things.  In a time rush, want to try to
>catch some people about a job <G>
>
>  Mark
>
>Peter Krey & Family wrote:
>>
>> I am emailing the Pic Microcontroller Discussion list to
>> request help in locating galvanic skin response (GSR)
>> circuit designs.
>>
>> My motivation here is to help a family member with a
>> rare, and life threatening health disorder, whose process
>> begins with increasing perspiration while sleeping.
>>
>> The end result of one of these episodes is similar to a
>> very severe epileptic seizure.  Needless to say that this
>> family member's health severely deteriorates with each
>> occurrence, including some close brushes with death.
>>
>> So identifying GSR circuit designs is key to developing
>> a microcontroller based device to "alert" family members
>> of this condition, and take actions, such as give medication,
>> ..., etc.
>>
>> Herein lies the reason for this email.  While various web
>> search engine searches have uncovered some "hobbyist"
>> level GSR schematics, I cannot judge the quality, or
>> reliability of the circuitry.
>>
>> So here are a few ideas on how PICList readers could be
>> of assistance:
>>
>>  - Can you think of any useful & reliable GSR designs you may
>>    have seen in the past ??
>> -  Note that since GSR's are one of the key sensors used in
>>    polygraphs or lie detectors, maybe you have seen something
>>    as part of projects under those names.
>> - Are you aware of any university or research work in this
>>    area ?
>> - Can you think of any useful reference design sources ??
>> - Who are some of the more knowledgeable circuit designers
>>   you have come across ??  Where can you refer me ??
>>
>>  Thanks in advance for your feedback and suggestions.
>>
>> Sincerely,
>>   Peter Krey