Biosensing with multiple Instrumentation Amplifiers

[biomurph]

Hi,

I'm designing a circuit with multiple instrumentation amplifiers to measure biopotentials. Let's say for the sake of argument that I'm using 4 InAmps (thinking that this will scale to 8, 16, etc...).
I'm using sense electrodes that are connected to the '+' inputs of the Instrumentation amplifiers, and one reference electrode is connected to all of the '-' inputs. Thereby, I'm measuring 4 sites against the one reference.

The body is also connected to a low impedance connection that is outputting the GND (Common Mode Voltage = Instrumentation Amp input range/2) also known as DRL (Driven Right Leg).

I have also connected 'ground path' resistors from each of the '+' and '-' inputs to CM voltage to allow for the input bias current on the InAmp inputs. This is critically necessary, otherwise the input signals will 'rail'. Now comes the question.

It's super important to make sure that the resistance of each of the 'ground path' resistors are as equal as possible. What happens when I attach multiple '-' inputs together like this? Should they all just get one bias current resistor? How do I calculate that?

If I want to add versatility, and give my user the option to either include the '-' input of one into the reference, or use it as an independent differential input, can I do that easily, or do I need a switch or something. The thing is, the design spec is asking for this versatility of either using a reference electrode for multiple input channels, or each channel being it's own differential input.

Image of the circuit shown for conceptual focus only.

[Gyro]

Hi,

A search on the Analog Devices website for 'ECG' brings up a good amount of biomedical devices - or possibly more importantly for you, app notes....

http://www.analog.com/en/search.html?q=ecg&Site_Search_393ead15-036b-4d0f-983b-abe90609ce16=

[Audioguru]

An INA amplifier uses three opamps and a few matched resistors so that its inputs are truly differential, but your circuit does not use differential signal inputs so why not use ordinary opamps with negative feedback instead? But without having differential input signals then there will be a lot of mains hum and other interference.

The patient's right leg is biased by an opamp so why do you need input resistors when the inputs are connected to the same patient?

The patient's right leg is usually fed an inverted common-mode signal that cancels mains hum, interference and DC from other muscles but yours is simply DC that does not need an opamp. 

[biomurph]

@Gyro, I'm not interested in using a pre-packaged AFE. I'm interested in figuring out how to 'roll my own'.
@Audioguru, The schematic I showed is for discusstion purposes only. Please imagine that they are instrumentation amplifiers, and the gain and reference, etc are all taken care of.

In the bottom of this page, there is a discussion of 'Floating Voltages'
http://www.planetanalog.com/document.asp?doc_id=527518
This is the issue I'm talking about. I see that adding the input bias-current return path resistors does make a big difference when I'm using only one channel.

My question is, how do I figure the value of the resistor on the inverting inputs if they are all connected together to a single electrode?