I'm designing active electrodes for surface EMG, active electrodes being electrodes directly connected to a pre-amp and filtering stage on the measurement site. Im working with the AD8221 IN-AMP to pick up the differential signal and I am trying to eliminate the need for a reference electrode by driving the reference pin of the IN-AMP via a voltage divider and buffer to bring the AD8221 to midsupply level. However with my simulations I have failed to receive the desired midsupply voltage level at the reference pin and infact only get a few hundred millivolts. Can someone please explain to me the theory of operation with the reference pin? and whether this method will effectively eliminate the need for the reference electrode?

Circuit specs; +/- 5V power supply for the entire circuit. Gain = 100. (Rg = 499 Ohms)

I have read through the other threads related to my topic, but none have encountered the problem I have. I am running a simulation on MultiSim v.13

Update: the current schematic

enter image description here

  • \$\begingroup\$ Even using this method, done with a proper low impedance vref, you will probably still need a reference electrode to keep down common mode noise \$\endgroup\$ – Scott Seidman Mar 21 '15 at 19:53
  • \$\begingroup\$ Hi Scott, thanks for the reply. Your help is much appreciated. Could you please tell me how the reference electrode would help keep the common noise down? I was hoping to use a tri-polar arrangement, so 3 electrodes, and then double differencing. Wouldnt that significantly reduce common noise? \$\endgroup\$ – RKTim Mar 21 '15 at 20:29
  • \$\begingroup\$ Three electrodes are fine, two (which I thought you were trying) are not \$\endgroup\$ – Scott Seidman Mar 21 '15 at 20:40
  • \$\begingroup\$ Scot, thanks. May I also ask if it is possible to use the same reference common to 3 inamps? \$\endgroup\$ – RKTim Mar 21 '15 at 20:48
  • \$\begingroup\$ Sure. Just make sure it's low impedance. I just used amN LM285-2.5 for an ad623 and 3 op amps, no prob \$\endgroup\$ – Scott Seidman Mar 21 '15 at 20:52

The reference pin has a 20k ohm input impedance and your potential divider may be formed from large resistors i.e. values in excess of 10k - this will cause a problem. Try driving the ref pin with a voltage source. Also note that there is a leakage current from this pin of about 50uA. This will produce a 0.5 volt error when fed from a voltage source with a 10k impedance.

  • \$\begingroup\$ Thanks a lot for the reply. I shall try what you have suggested and update. I was using 100k resistors. \$\endgroup\$ – RKTim Mar 21 '15 at 20:26
  • \$\begingroup\$ May I also ask, is it possible for this same reference to be common for 3 INAMP's? I am trying to use the double differential method to eliminate Common mode noise. \$\endgroup\$ – RKTim Mar 21 '15 at 20:27
  • \$\begingroup\$ You can use the same reference (low impedance output) for multiple AD8221s \$\endgroup\$ – Andy aka Mar 21 '15 at 21:33

Looking in detail, now, the ad8221 will not work single sided. If you are planning on rail spliting, this might be Ok, but be careful of your common mode input range. You can use the ad623 single supply. The gain of 100 is probably way too high. You should assume there can be 150 mvolts of difference just because of electrode junction potentials, and you'll saturate without a big supply range, especially if working single sided. Recommend modest gain, like 5 or so, then high pass and low pass filtering, then an op amp gain stage. Overall gain of 500 may be more comfortable than 100.

  • \$\begingroup\$ Ahh, I see. Okay, I am planning on using a DC-converter (TC7660) to pump a negative supply for the chips. I am using Sallen-key 2nd order filters (high pass and low pass) pass band = 20-500Hz. I am however having problems simulating the power supply for the amplifiers, I am connecting a DC source across the power inputs of the amps. I shall reduce the gain and update. Many thanks Scott! \$\endgroup\$ – RKTim Mar 21 '15 at 21:27
  • \$\begingroup\$ @RKTim you might peek at the ad623 or ad8237. Your design gets a whole bunch easier while remaining robust \$\endgroup\$ – Scott Seidman Mar 21 '15 at 21:37
  • \$\begingroup\$ Reading the data sheet for the AD8237, it seems like a much better choice. \$\endgroup\$ – RKTim Mar 21 '15 at 21:57
  • \$\begingroup\$ Ad627 is another \$\endgroup\$ – Scott Seidman Mar 21 '15 at 22:06
  • 1
    \$\begingroup\$ I have changed to the AD8237 and set up like below, \$\endgroup\$ – RKTim Mar 21 '15 at 23:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.