Other Parts Discussed in Thread: ADS1294
We are implementing ADS1294R in our product for high-fidelity EEG monitoring applications. However, we are using 4 active dry electrodes to create 2 fully differential channels which include buffers and amplifiers between the ADS1294 and the electrodes. The ADS1294 channels are facing outputs of instrumentation amplifier and buffers, so it won't be able to inject the AC lead off currents. In that regard, we would like to receive guidance on how we can still implement real-time impedance measurement (very critical for dry electrodes) with our current HW architecture. Below are a few options we are considering and any feedback would be greatly appreciated:
1. Use an analog CMOS SPDT switch to bypass ADS1294 to electrode during impedance measurements. However there are 3 issues with this: a) Previous part we used TMUX1136DQAR and it seemed to induce a few mV of noise in EEG. Is there a better switch we could possibly use such as ones you are using inside ADS129x chips? Can you provide some best SPDT switch options for such sensitive application? Also, in this method, we cannot do simultaneous impedance and EEG? We cannot use physical switches because the process needs to be automated for our application. We cannot ask the user to move a slide switch whenever signal gets bad as it will deteriorate our user experience.
2. The other method we are considering is designing our own external current source directly on active electrode PCB (controlled by uC GPIO) as suggested in this link (https://www.ti.com/lit/ug/slau507/slau507.pdf). That way we can inject the current directly bypassing the active electrode circuitry. On the way back, the return currents will return through the active electrode circuitry. Does this method sound feasible? Also, do you have any reference circuit design to generate a very low AC constant current source (at 30Hz etc)? What type of switch can we control using a uC to generate a sine wave current source of 6nA? If we are out of DAC pins, we plan to generate a square wave and filter it into a sine wave. Also, would we need a current source for all electrodes or only one in each differential pair? Though theoretically, it might work, however we are concerned with the tolerance in measurement (resistor values, offsets etc.). I read somewhere that the tolerance of your internal current source is also 20%, is that correct? Do you think implementing an external current source (to each active electrode) could obtain in par performance to your internal AC lead off detection current source?
3. Another option is to use an external impedance IC like MAX30002.Do you have any alternative or better solutions for this like a simple 6nA sine wave current source? We ideally don't want digital communication close to our preamps so this option may not be feasible.
3. Another option is to remove active electrode circuitry and connect ADS1294 with dry electrodes directly. In this case, we are afraid that the poor contact or not-so-good reliability of dry electrodes may cause an issue with your performance, but atleast we would know when that happens by doing continuous impedance measurements?
Lastly I had some questions, in improving our active electrodes:
1. We are currently using this high input impedance buffer (AD8244ARMZ) with 4 channels and would like a similar alternative (with very high input impedance) with only 2 channels (to save space and cost). Can you recommend an alternate part?
2. We are using passive filters for active electrodes and would like to know if there is any suggestion on using active filters for fully differential signals? How can we resolve tolerances in opamp offsets and passives tolerances etc.?