If powering through audio jack is difficult, can use a small lithium polymer battery (this is what the Vestibular Diagnostics folks are planning on for their glasses). Should order some of these.

Need to order the micro-transformer:

Ordered 4-line audio jack splitter into a 3-line mic jack and a 3-line headphone jack

This is the precedent: http://energy.eecs.umich.edu/wiki/doku.php?id=proj:ekgsense:home

Need to figure out how my version is better: one thing is just that the full plans are not online, and also this can be much simpler for a specific application like EKG

HiJack itself can be bought for $79 here

http://www.seeedstudio.com/depot/hijack-development-pack-p-865.html

and I should think about whether I want to just do that.

"15.8 mW per channel from the iPhoneÕs headset port" -- need an ideally matched load which constrains the output audio frequency

Components needed to make HiJack's power-harvesting circuit: see image

But I don't really want to duplicate something that is commercially available. So the question becomes, can I buy and use a HiJack system to interface with the Fab-ECG and Fab-Sampler and just build a nice housing for this.

Sub questions: 

1) How much power do the amplifier board and the FabSampler board require? Can test this by powering it with the bench supply and measuring current output. If too large for HiJack, how to reduce the power consumption?

2) Can I get the HiJack board in time.

On the other hand, the seeedstudio people only have 12 of these and I'd really want to make this a single board. 

So perhaps I do want to just duplicate their energy harvester design.

They only got 15 mW out.

The tiny44 consumes

"Active Mode (1 MHz System Clock): 300 µA @ 1.8V"

I think a better option for this project might just be to incorporate a battery. Because otherwise low power electronics is going to be involved.

Instead going to just do the glasses with Charles: see images

for the pulse oximetry piece:

http://www.oximetry.org/pulseox/principles.htm (easy pulse)
http://embedded-lab.com/blog/?p=5508

FabEEG on one side, FabPulse on the other with connectors as vias from one side to the other to connect power and ground

Leaving out the potentiometer from FabPulse. Also keeping the enable line always-on and thus replacing that transistor with a wire.

Simplified FabPulse has only 1st amplification stage and no voltage follower(used to lower the output impedance) and no LED: but that's still 100x gain with frequency filtering --- should end up with ~ 1V signals.

FabPulse_v1_simplified component value replacements:

R1: 47k -> 49.9k
R2: 6.8k -> 4.99k
C1: 4.7 uF -> 3.3 uF + 2.2 uF stacked
R3: 680k -> 500k
R5: 150 Ohm -> 100 Ohm

Signal is small and I'm not getting appreciable amplification gain.

Better pulse ox setup: http://blog.energymicro.com/2012/11/21/create-a-simple-pulse-oximeter-with-tiny-gecko/

Got gain stage working for first gain + filtering stage of easypulse. Very unreliable though.

Tomorrow: make another Fab-Sampler, think about new pulse designs, work on the frame / integration