Gas station without pumps

2012 August 20

Better electrode placement for EKG blinky

Filed under: Circuits course,Data acquisition — gasstationwithoutpumps @ 17:00
Tags: , , , , ,

After talking with an EKG expert, I changed the placement of the electrodes to correspond more to the conventional limb electrodes (LA=left arm, RA=right arm, LL=left leg).  I didn’t want to use limb electrodes, because there are then more muscles that can cause EMG interference, so I put the electrodes on the torso about where the limbs connect.  He suggested that Lead II (LL-RA) would give the most canonical signal for recognizing the parts of the EKG waveform.

Connection of electrodes to get “Lead I” which is LA-RA.  The top two electrodes should probably be placed more symmetrically—I think that the right-arm (–) electrode may be better placed than the left-arm (+) electrode.
A portion of the Lead I signal. The PQRST are all quite clear. The voltage scale is approximate, based on a computed gain for the EKG blinky amplifier. The zero-line is set by averaging the readings for about 10 seconds of recording.  The sampling period was 3 msec.
The Lead II electrodes (LL-RA). This is the conformation that the EKG expert said would most likely give the canonical waveform for a normal, healthy individual.
The waveform for Lead II is indeed quite strong, with very clear P and T waves. Note: despite the same time scale, the Lead I, Lead II, and Lead III recordings were all done separately, with the wires moved between recordings.  They do not correspond to the same heartbeats.
The arrangement of wires for Lead III (LL-LA).
The R wave is very hard to see in Lead III, probably because the dipole is oriented almost horizontally, so the vertical lead placement sees little difference in the field.  The R wave was not strong enough to flash the LED.

I wonder if I should add a ‘calibrate’ button to the blinky EKG to put out a 1mV pulse for a fixed duration so that the gain does not have to be estimated. That may be overkill for the blinky, but would make sense if I decide to make a 2-channel EKG. Note that 2 channels (say Lead II and Lead III) are really all the data in the first 6 “leads” of a 12-lead EKG. All the rest (Lead I, aVR, aVL, aVF) can be computed from those two.

It looks like a gain of about 2000 is needed to turn the 0.7mV R waves into 1.4V pulses to light up the LED.  Gains over 3500 would clip the tops of the R waves.

2012 August 15

Measuring Ag/AgCl electrodes

Filed under: Circuits course — gasstationwithoutpumps @ 13:46
Tags: , , , , , ,

I had a better idea for measuring Ag/AgCl electrode characteristics than I mentioned in yesterday’s post Better measurement of conductivity of saline solution. Rather than buying fine silver wire and soaking it in bleach (though that may still be worth doing), I just took two of the Ag/AgCl gel electrodes I had for the EKG lab and stuck them face-to-face.  The adhesive disks do a pretty good job of sealing in the electrode gel.

I expected this electrode sandwich to have very low resistance, both because of the good properties of the Ag/AgCl electrodes and because of the short distance between the metal electrodes.  I was not sure what to expect for capacitance, as the non-polarizable nature of Ag/AgCl would lead me to expect little change in resistance as a function of frequency.

Trying to measure the resistance of the electrode sandwich with an ohmmeter is still misleading though, as it gradually creeps up as the meter charges a capacitance, going from 40Ω to 80Ω in about a minute.  Reversing the connections drops the resistance reading to 20Ω, from which it gradually increases as the capacitance is charged the other way.

Plot of resistance of Ag/AgCl sandwich made from two Vermed SilveRest EKG electrodes as a function of frequency. Note that a power-law fit seems much better than an R+(C||R) model.

I’m getting a power-law fit with a small exponent, rather than either a constant resistance or something that fits an R+(C||R) model. Looking at I vs. V plots on the oscilloscope, it looks like a pure resistance at high frequency (no phase changes), but there is some phase change at lower frequency.  The small change in resistance with frequency is not easily modeled with an RC circuit.

For the EKG frequency range of 0.01Hz to 150Hz, the resistance of the electrode pair (using the power-law fit) varies from 48Ω down to 8Ω.  Given the large resistance of the skin, this electrode resistance is so small as to be negligible. Putting small Ag/AgCl electrodes into a saline solution, however, might have quite a different effect, where the resistance of the electrodes plays a more substantial role in the electronic behavior of the system. So maybe I do still need to order some silver wire and make my own Ag/AgCl electrodes (either by bleaching or by electroplating).


« Previous Page

%d bloggers like this: