A few years ago, I needed a teeny, tiny potentiostat for my biosensor research. I found a ton of cool example projects on Hackaday and on HardwareX, but they didn’t quite fulfill exactly what I needed. As any of you would do in this type of situation, I decided to build my own device.

Now, we’ve talked about potentiostats before. These are the same devices used in commercial glucometers, so they are widely applicable to a number of biosensing applications. In my internet perusing, I stumbled upon a cool chip from Texas Instruments called the LMP91000 that initially appeared to do all the hard work for me. Unfortunately, there were a few features of the LMP91000 that were a bit limiting and didn’t quite give me the range of flexibility I required for my research. You see, electrochemistry works by biasing a set of electrodes at a given potential and subsequently driving a chemical reaction. The electron transfer is measured by the sensing electrode and converted to a voltage using a transimpedance amplifier (TIA). Commercial potentiostats can have voltage bias generators with microVolt resolution, but I only needed about ~1 mV or so. The problem was, the LMP91000 has a resolution of ~66 mV on a 3.3 V supply, mandating that I augment the LMP991000 with an external digital-to-analog converter (DAC) as others had done.

However, changing the internal reference of the LMP91000 with the DAC confounded the voltage measurements from the TIA, since the TIA is also referenced to the same internal zero as the voltage bias generator. This seemed like a problem other DIY solutions I came across should have mentioned, but I didn’t quite find any other papers describing this problem. After punching myself a little, I thought that maybe it was a bit more obvious to everyone else except me. It can be like that sometimes. Oh well, it was a somewhat easy fix that ended up making my little potentiostat even more capable than I had originally imagined.

I could have made a complete custom potentiostat circuit like a few other examples I stumbled upon, but the integrated aspect of the LMP91000 was a bit too much to pass up. My design needed to be as small as possible since I would eventually like to integrate the device into a wearable. I was using a SAMD21 microcontroller with a built-in DAC, therefore remedying the problem was a bit more convenient than I originally thought since I didn’t need an additional chip in my design.

I am definitely pretty happy with the results. My potentiostat, called KickStat, is about the size of a US quarter dollar with a ton of empty space that could be easily trimmed on my next board revision. I imagine this could be used as a subsystem in any number of larger designs like a glucometer, cellphone, or maybe even a smartwatch.

Check out all the open-source files on my research lab’s GitHub page. I hope my experience will be of assistance to the hacker community. Definitely a fun build and I hope you all get as much kick out of it as I did.