Help with instumentation amplifier

[jm2t]

Hello edaboard, I'm new here and I'm sorry my first post has to be a question, but I am really desperate with this one.

I've been working on a EKG amplifier using the AD624. You have the pin connection and schematics for a gain = 1000, extracted from the datasheet below.



The big problem is, I connect everthing right, and the V output, when the two inputs are connected together (in theory, Vdif=0) keeps pointiong as -7,6V respect to ground. Not only that, but if I diosconnect the inputs it has little to no effect. It's as if it was saturated to -7,6V regardless of the inputs (i checked and the negative feedback loop from pin 9 to 10 is correct). the only change I can induce is a 0.5V change by connecting a 9V battery between the inputs.

tried with diferent gains and the problem remains the same. i wonder if i am making a big stupid mistake here, but I've been the last 4 days testing every possibility I can think of with no result. the dual supply i'm using are 2 9v batteires. Below i leave you the schematics (the variable resistor sets a 3V offset, but adding that makes no effect, except that when connecting the 9V battery to the inputs, it reads -3.2V instead of -7.2V, without adding the battery though, it stays at -7,6V) and the breadboard altough its not very clear how all is connected.





if you need any more info and have any idea it would really be appreciated.

Thank you very much

 

[IanP]

For a quick test, I’d connect output with the (-) input for a gain (1) amplifier, and try it that way –from -9 to +9 with a potentiometer ..
:wink:
IanP

 

[Audioguru]

All working ECG circuits in Google Images and on instrumentation amplifier datasheets are made neat and tidy on a compact pcb (not on a messy breadboard with wires all over the place). The gain of the instrumentation amplifier is low so it can deal with the high common-mode DC and AC signals picked up by the patient.
The common-mode signals are inverted by an opamp and fed back to the patient's right leg to cancel them. Then an ordinary opamp amplifier filters and provides most of the voltage gain.

The messy long wires on your breadboard and many strips of intermittent contacts are antennas that pickup all kinds of interference that your instrumentation amplifier is trying to amplify 1000 times.

Here is a project using an AD624 operational amplifier with a gain of 1000 like your circuit. It has the patient connected to 0V and uses an opamp for an adjustable offset voltage. It was made on a tidy pcb. http://www.cisl.columbia.edu/kinget_group/student_projects/ECG Report/E6001 ECG final report.htm

 

[jm2t]

For a quick test, I’d connect output with the (-) input for a gain (1) amplifier, and try it that way –from -9 to +9 with a potentiometer ..
:wink:
IanP

Hi, thanks for the answer. I did that and indeed it works as a unity gain amplifier, +8.2 volts that i guess are the saturation for the amp (the seame as if I turn off the V- voltage supply). Weirdly enough, i don't get any effect varying the voltage on the in - input using the potentiometer.

@auidioguru: thanks for the answer too. Yes, I know and I was going to solder into PCB but first I needed to see it was all working properly. i know that ECG model and is the one i am primarily based (note the 3V offset on that schematic wich i also plan on using).

However that schematic does not use the right leg drive (wich is a great way to cancel the common mode interference), and simply grounds the patient to the circuit ground (wich i am also doing). The problem is, even using the 3V offset it seems to do nothing adn stays on -7.6V.

You are right that the interference is high, i was planning to use shielded cables, notch filters, etc to reduce interference. The thing is, having a CMMR of 130Db, shouldn't it at least tay around the 0V mark when the inputs are interconnected instead of goinf to -7.6V? I think I'm making another mistake.

I built a rudimentary 100000 gain differential amplifier with a cheap LM324 on the same breadboard and, altough with LOTS of noise, I showed the QRS and mean volatge when interconnecting the inputs was near 0. That's why I think building it "as is" in a PCB would not solve the problem.


Thank you again for all the answers, hope for your reply

 

[Audioguru]

Maybe you have the pins of the 200k trimpot mixed up. Measure its adjusted output go smoothly from -9V to +9V. Then the output of the AD624 should go smoothly from -7V to +7V.

The inputs of the AD624 MUST have continuity to 0V when you connect them together. The patient's body connects them to 0v when it is used as an ECG measuring circuit.

 

[jm2t]

Maybe you have the pins of the 200k trimpot mixed up. Measure its adjusted output go smoothly from -9V to +9V. Then the output of the AD624 should go smoothly from -7V to +7V.

The inputs of the AD624 MUST have continuity to 0V when you connect them together. The patient's body connects them to 0v when it is used as an ECG measuring circuit.

Hello, that might be possible. i attach a picture below of how I have it connected, I think the trimpot is fine. V+ and V- on either side with out on centre, tuning it goes smoothly from -9V to +9V

Then, connecting it to the LM324 on a voltage follower produces the anticipated +-7,2V smooth transition.

Connecting it to the ad624, no gain settings added, just the loop to in - and out. and the connection of in + to the potentiometer. This does NOT produce a smooth transition, at -2.5V on the potentiometer it jumps to +8V and then it stays there (+/- 0.1V) all the way from -7v to +9V. you have to go down to -7.5V to see it drop back to -7.5V. When you rise again after the -2.5 point, it again shoots to +8V.




I really do not understand this behaviour. the connections are as they are made on the right picture. The center picture is for reference as for the gain = 1000, wich I tested too with the same result.

Do you have any idea what could be wrong?

 

[Audioguru]

I think you should replace the faulty AD624.

 

[ftsolutions]

I am also not seeing ANY bypass/decoupling capacitors on the power pins of the high-gain AD624. It could be a faulty IC, or it could still just be running toward one rail because of noise injection and oscillation. We cannot tell if it is going into HF oscillation by simply looking at a multimeter. I would recommend adding appropriate power rail caps to your breadboard if you want to continue to use that breadboard as a platform for working on this circuity - breadboards are good for somethings, and that usually includes adding noise.
just trying to help..

 

[jm2t]

I am also not seeing ANY bypass/decoupling capacitors on the power pins of the high-gain AD624. It could be a faulty IC, or it could still just be running toward one rail because of noise injection and oscillation. We cannot tell if it is going into HF oscillation by simply looking at a multimeter. I would recommend adding appropriate power rail caps to your breadboard if you want to continue to use that breadboard as a platform for working on this circuity - breadboards are good for somethings, and that usually includes adding noise.
just trying to help..

Hi, thanks for your reply. Tought about that, and I already connected a 10uf cap for both power rails and 0.1uf capacitors near the voltage inputs of the IC's in the circuits. The results are the same.

Do you have any definite test that can confirm me this is a faulty IC? This is the complete diagram for the ad624:

 

[jm2t]

Hello again. The IC was faulty, I replaced it with a new one and it worked...

But now, today, it has stopped working, wherever I connect it to my 2 electrodes on my body it SATURATES to +8V or -8V. When i connect them altogether it behaves as expected (0.3V) and connecting it to a potentiometer in voltage follower mode gives a smooth transition. So this IC is not faulty, I don't understand what is going wrong.

i changed the elctrodes, the cables, I even soldered it in a board in the chance it was being saturated with noise (altough it WORKED before). i really did not change anything from yesterday to this morning except I disconnected and reconnected the battery, yet yesterday it behave normally and today its sturated.

I would really appreciate a bit of help here, I'm closing desperation. thanks

PS: below you have the pics as when it was working yesterday. Now it is just a flat line at 8V.

 

[Audioguru]

Most working ECG circuits use an opamp as an inverter to feed a negative common-mode signal to the patient instead of a ground. The feedback is AC and it is also DC because the patient's body picks up a lot of AC and DC common mode noise. The feedback is frequently fed to the patient's right leg.

 

[jm2t]

I dont think thats the problem. I have temporarily solved it aplying a offset potential on the reference electrode with a pot, but with this I cant use right leg drive (ref must be connected to ground)