Figure 44 of the AD8235 Data sheet shows the device paired with AD8609 in a hi-pass configuration with break at 0.3 Hz. The AD8609 has very high voltage noise at this frequency. Fundamental of Heart Beat is about 1 Hz with content to about 60 Hz.
As I am in the task of designing a mobile low power, single cell (Li-Ion), and hopefully very quiet EEG and EKG device, can anyone tell me why the AD8609 was selected rather than a very quiet amp like the AD4528?
The AD8235 and the AD8609 were both released in 2009. The ADA4528-1 was released in 2011.
Note in figure 44 that one op amp is used as a comparator, one is an integrator, one is a gain of
one Sallen-Key filter, and one with a noise gain of 403. I would not use a quad in this application.
Also note that there is a +2.5V supply and a -2.5V supply. There are some pros and cons to using
split supplies vs. single supply. I don't know who did the application, but my guess would be that
it is meant to be a conceptional design, rather than a production-worthy design to be copied.
We do see a lot proprietary customer schematics in applications, so we can't really show the
"best" solution, but rather concepts.
To get the low flatband noise and bandwidth, the ADA4528-1 burns 1.8 mA max at 25 C, so if you
were building a battery powered instrument, this may not be a good choice. Again, if battery
operation is required, the major choice would be split supplies or single supply, which would
determine if rail to rail inputs and outputs are required, which have their own warts.
Bipolar op amps have the lowest 1/f corner for voltage noise, so you may want to look at the
AD8622 or ADA4096-2, which are bipolars.
I think the AD8609 was chosen mostly because of the 50uA supply current/amp.
But in addition to Harry's comments, I would like to point out that the AD8235 is in a gain of 5 in this circuit. AD8609 guarantees 3.5uV p-p from 0.1 to 10Hz. Referred to input, that's only 0.7uV p-p, which is not significant compared to the 4uV p-p noise of the AD8235.