Subject: Recent Reliability Studies
From: Kevan Hashemi <hashemi@opensourceinstruments.com>
Date: Thu, 04 Jun 2015 11:58:59 -0400

Dear Collaborators,

We have been making test batches of transmitters and running them in water at 60C. According to the literature on corrosion in electronic circuits, failure takes place 10 times faster at 60C than at 37C. We use 37C as an upper limit for the operating temperature of an implanted transmitter.

Of nine rat transmitters, the first failed in 5 weeks at 60C, which implies 50 weeks at 37C.

Of nine mouse transmitters, the first failed in 9 days at 60C, which implies 90 days at 37C. One of the mouse transmitters is still running after 25 days, while the remaining 8 failed between 9 and 12 days.

By tearing apart circuits as soon as they fail, we identified the 10-uF capacitors as the source of intermittent leaks that drain the battery.

There are five 10-uF 10-V rated capacitors on each single-channel transmitter. These capacitors are the highest-density capacitors available in the world. The higher the capacitor density, the thinner the dielectric insulation and the more vulnerable it is to corrosion.

Our hypothesis is the roughly 1 in 5 of these capacitors has a microscopic crack created during capacitor construction (as opposed to during circuit assembly, which was our problem last September). We suspect that the capacitors on our rat circuits have the same problem, but because the rat battery has five times the capacity, the leaks cure themselves before they can drain the battery, and so we have a longer time to failure.

We are equipping 5 rat circuits with mouse batteries and 1-uF 35-V rated capacitors in place of 10 uF 10-V rated components. Despite my calculations, it turns out that the circuit works fine with the lower capacitance, except that the low-frequency cut-off goes up to 3 Hz from 0.3 Hz, which I'll worry about later. We will poach these transmitters and see if we get longer life. We are also poaching four unmodified rat circuits with mouse batteries as a control.

It looks like we no longer have the severely-cracked capacitors we encountered last year. And we have solved the square-wave and low-gain problems with our improved epoxy encapsulation. What we think we are seeing now is a reliability problem that has existed for ten years, but we were never able to observe until we started our 60C poaching. Now that we can observe these failures, our hope is to break through to a new level of reliability in the next six months.

Until then, your cooperatin and patience will be much appreciated.

Yours, Kevan

-- 
Kevan Hashemi, President
Open Source Instruments Inc.
www.opensourceinstruments.com