Dear SCT Users,

I have one new mouse transmitter here that has the dreaded Square Wave 
Problem. It generates a full-scale square wave when its leads are open 
circuit. This transmitter has no breaches in the silicone or epoxy, and 
its antenna is completely sealed.

Our hypothesis that the square wave was caused by water leaking into the 
tip of the antenna is disproved.

It turns out that silicone and epoxy are water-proof, but they are not 
humidity-proof. Most silicone vulcanizes by penetration of water vapor. 
Epoxy absorbs water vapor.

It also turns out that conducting films form on circuit boards when 
there are three ingredients in place.

(1) An electric field, such as between two terminals at different 
voltages. The square wave problem arises only in transmitters that have 
been running for at least an hour, not turned off.

(2) Moisture. When the transmitter is implanted or sitting in water, 
moisture will penetrate our encapsulation, so that humidity inside will 
be 100%. If the temperature drops, there will be condensation upon the 
circuit board. The square wave problem arises only for transmitters that 
have been in water for several days.

(3) Some other active chemical like acid flux residue or partly-cured 
epoxy. The square wave failure has occurred only when we either use acid 
flux on the board OR we have improperly mixed epoxy.

It turns out that the amplified X signal on the A3019 is close to the X 
input pad. The X input impedance is 10 MOhm, and the amplifier gain is 
40 so a conducting film with impedance less than 250 MOhm would cause 
the amplifier to oscillate. The amplified X would add to the X input. 
Indeed, we observe the X input of an oscillating transmitter to have a 
large signal on it, even when we apply no signal.

The problem does not arise when the source impedance of the signal 
connected to X input is less than 1 MOhm. Thus, when the electrodes are 
connected to the brain, this oscillation should be stopped by the 
10-kOhm source impedance of EEG. When the electrodes come loose, the 
square wave can start.

And so we arrive at the latest hypothesis. There is no way to stop 
moisture getting in, and we have a large number of ready-made circuit 
boards. We will stop using acid flux on the boards AND make sure the 
epoxy is mixed properly. These precautions may stop the problem. But in 
any case, so long as your electrodes are well-connected to the brain, 
the oscillation should not arise.

I appreciate that this hypothesis will most likely be disproved soon. 
But until then, at least I have hope.

Any time you find an oscillating transmitter, please retrieve from the 
animal, and try holding one lead in each hand, with wet fingers, to see 
if the oscillation persists. It should not, and indeed does not in the 
examples we have here. The impedance between your fingers is about 100 
kOhm. I believe Dennis and Rob have a few such transmitters. Perhaps you 
could test them for me.

Yours, Kevan

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