- Ask a related questionWhat is a related question?A related question is a question created from another question. When the related question is created, it will be automatically linked to the original question.
Hello,
We are working in DC coupled Biopotential recordings. We are currently comparing two different electrode types that are exposed to the same environment at the same time, and should theoretically be giving the same recorded signals back. We are using Instrumentation amplifiers to record these differential signals.
Our circuit design works wonderfully when we are working with only one pair of electrodes and only one INA. Problems arise when the second pair of electrodes is added in here.
In the above circuit, we have two INA amplifiers which are powered by an isolated DC/DC converter, both relative to the isolated AGND, where the original power (relative to digital GND) comes from an external power supply. The INA outputs head to an ADC on the same isolated power supply. To bias the amplifiers to common mode input, the 10M ohm resistors are added onto the electrode inputs. An asterik in the diagram represents that this component is not necessarily added in. Again, this circuit measures well when only one amplifier is in the solution. As soon as the second is added, we are still able to record signals, but much the inherent value of the signals is compromised due to the 20M ohm resistance connecting WE to RE through the AGND connections. This takes the INA leakage current from the pA scale down to the 100nA scale. The reason for this is that the different metal electrode types each have different Nernst potentials in a biological fluid. The WE electrodes could have a 500-1000 mV DC bias compared to the RE electrodes, thus creating a current flow from WE to RE.
The question is, is it possible to still record with both electrode types and maintain the necessary common mode biasing of the amplifiers? Are there ways to circumvent this?
Hey Patrick,
I see about 50 nA when I have 1V on the WE side. The high gain is causing the output of the INA to rail out. There are two possibilities to fix this. Lower gain can be used for the WE circuit or AC coupling on the input can be used. It sounds like AC coupling isn't possible due to the nature of the input. Have you tried lowering the gain for the WE circuit?
Best Regards,
Chris Featherstone
Hi Patrick,
Our circuit design works wonderfully when we are working with only one pair of electrodes and only one INA. Problems arise when the second pair of electrodes is added in here.
Are there any reasons that you have to measure the both electrodes in the same beaker or electrolyte? If you compare each pair of electrode in a separate electrolyte, they your problem is not there.
We are currently comparing two different electrode types that are exposed to the same environment at the same time, and should theoretically be giving the same recorded signals back. We are using Instrumentation amplifiers to record these differential signals.
The difference in WE and REF electrodes are only in mV or <1mV, so you would need to have gain resistors in INA121 in order to record it properly. Yes, if you leave pin1 and 8 open, Gain = 1 V/V.
I have 'fixed' the issue by removing all 10M ohm resistors to ground on the amplifier inputs and instead directly connected the Vref with AGND and attaching this through a fifth external physical electrode connection to the solution.
Chris's suggestion is correct. You should remove 10 Mohm resistor from the measurement. I don't think that is needed, since the electrolyte's ionic conductive is high, it should not have input bias current return issues in INA121.
In addition, I think that you may not need isolated power supply and all the measured nodes should be reference to AGND. Of course, you should not use switching power supply rails for the application and measurements. Battery power or precision DC linear supplies are the best for the application. You may have to place a Pt or stainless steel foil (non-reactive and electrochemical stable) at or around the inner wall of the electrolyte and tie the conductor to AGND to establish the input bias current path to AGND. The impedance between electrodes are equivalent to your "10MΩ" resistors.
however leaves various mysteries in regards to what exactly should bridge this AGND to the conductive solution, as different (small, <1mV) DC biases appear based on the connection material and whether or not these are buffered with resistors.
I do not know what your reference electrode looks like. Are we talking about half cell in 3 electrode system, see the image example of aqueous Ag/AgCl reference electrode below.
There is agarose ionic membrane at the bottom of the Ag/AgCl or other ref electrode constructions. So we are talking about different electrolytes in WE and REF electrodes. In 3-electrode system, REF and WE should be placed next to each other, and CE can be further from WE. The potential differences are always present between WE and CE due to the half cell electrical potential, but current flows between them are negligible.
I am not certain why you are comparing REF1 vs. REF2 and/or WE1 vs. WE2 in this manner. You are only comparing the differences of electrodes in DC bias voltage or the construction difference of electrodes (half cell potential). This is only a single point DC measurement, unless you are doing electro voltammogram type measurement. Are you doing AC measurement in the above configuration?
Please let me know your overall setup details, perhaps I am able to provide you some ideas.
Best,
Raymond
