Medical Electronics

[littletransistor]

Your circuit is missing the 100uF capacitor from pin 5 to pin 4. The inputs are not balanced without it and then they pickup and amplify a lot of mains hum.

Also, you don't show a wire called Ref from pin 5 that connects to the patient's leg. Without it then the inputs are not biased.

Measure the output DC voltage again. It should be close to half the supply voltage. What is the supply voltage?

Okay - I'll connect the 100uF cap to the 4, and 5, and also Ref to my leg.

Update: Yes - the DC output is already around half of the supply voltage - 3.85 - 3.95V. But still, it's a little bit more or less than the 0.75V difference.

edit: well, I suspect that the computer of mine is playing tricks on me - apparently those line-in inputs should not be directly connected to my ECG machine. I read in another ECG kit that they added the line-in driver or a modulator thingy inside to make it properly work! :eek If so, I might need to go to my school's lab and try it out on an oscilloscope instead.

 

[audioguru]

The computer or oscilloscope is probably earthed so the patient's body must be insulated from earth. Then measure the output with an insulated multimeter set to measure DC. You will see the reading jump with each heartbeat and maybe see muscle contractions in each arm.

 

[littletransistor]

The computer or oscilloscope is probably earthed so the patient's body must be insulated from earth. Then measure the output with an insulated multimeter set to measure DC. You will see the reading jump with each heartbeat and maybe see muscle contractions in each arm.

I hooked it up to a multimeter - strangely the readings shows 7.67V without connection to the body, but when I started connecting it with all the leads, arms and leg, it becomes 0.65V and it did not fluctuate. When I connect the thing to the arms - it becoms almost half of 9V. Something is latching or locking my device... by right it should work even more better than the plain one-741 device too...

 

[audioguru]

The inputs must connect to the Ref through a conducting body or the instrumentation amp doesn't work. Then the resting DC voltage at the output should be near half the supply voltage (the Ref voltage).

 

[littletransistor]

The inputs must connect to the Ref through a conducting body or the instrumentation amp doesn't work. Then the resting DC voltage at the output should be near half the supply voltage (the Ref voltage).

Okay, I did that already. For a much better accuracy, I used a voltage follower (another op-amp) to drive the Ref point. Still, the DC output shows around 7 Volts.

Could my instrumentation amp broken in the process? Must I do this, like getting the gain for the i-amp 100, then another op-amp to multiply the gain for the another time, like those multistage ones?

And can I make my low-pass filter using a variable resistor and a 0.1 uf cap?

edit: Also, I found this **broken link removed** and the author stated that "...This means the output signal varies from something like -8 to +8 volts...".

So my DC output is around +7v, but that one is a split-supply EKG. Any opinions?

 

[audioguru]

If both inputs of your instrumentation amp are connected to half the supply voltage and its gain is set to 1500 or less, then its DC output should be near half the supply voltage.

Are you using shielded audio cables connecting the probes to the circuit? The shield should connect to Ref at the circuit and not connected tto anything at the probes end. Then they aren't antennas.

The circuits with a dual supply polarity are the same as this one. The Ref is at half the supply voltage.

 

[littletransistor]

If both inputs of your instrumentation amp are connected to half the supply voltage and its gain is set to 1500 or less, then its DC output should be near half the supply voltage.

Are you using shielded audio cables connecting the probes to the circuit? The shield should connect to Ref at the circuit and not connected tto anything at the probes end. Then they aren't antennas.

The circuits with a dual supply polarity are the same as this one. The Ref is at half the supply voltage.

I'll connect both inputs to the ref and the 1/2 supply voltage at pin 5. I pretty guess that it is really half of the supply.

Well, I'm not using shielded cables - perhaps I'll try to salvage an old audio cable or a phone cable for these.

 

[audioguru]

I'll connect both inputs to the ref and the 1/2 supply voltage at pin 5. I pretty guess that it is really half of the supply.

Don't guess if the voltage is at half the supply voltage or not. Measure it.

Well, I'm not using shielded cables - perhaps I'll try to salvage an old audio cable or a phone cable for these.

Phones don't use shielded cables. Audio equipment comes with them (in North America they have RCA plugs on them), or buy cheap ones at The Dollar Store.

 

[littletransistor]

Don't guess if the voltage is at half the supply voltage or not. Measure it.

Sorry for replying too quick - I measured it just now. Well, I brought both inputs to pin 5 and using the voltmeter, in respect to pin 4 and 6, there I got a 3.95V. Strange, it should be at least 4.13V (the battery's voltage left is 8.26V exactly when I measured it).

Phones don't use shielded cables. Audio equipment comes with them (in North America they have RCA plugs on them), or buy cheap ones at The Dollar Store.

Hmm... maybe I'll try to search in my storeroom for a spare old audio cable. I have tons of them.

 

[audioguru]

3.96V is only 0.18V from 4.13V and is perfect. The input offset voltage of the instrumentation amp could make it plus or minus 0.75V from half the supply voltage.

 

[littletransistor]

3.96V is only 0.18V from 4.13V and is perfect. The input offset voltage of the instrumentation amp could make it plus or minus 0.75V from half the supply voltage.

That'll be great news then!

So, should the Vcc/2 also be feeded into those both inputs? From the datasheet, it seems that they allowed to do so, but I 'm actually hoping I don't end up frying the chip! Is it possible that when those inputs are feeded with Vcc/2 and also the small signals from the body as well?

Will do some more testing later - gotta sleep now, it's 1.10 am in SE Asia.

 

[audioguru]

So, should the Vcc/2 also be feeded into those both inputs?

The patient's body conducts the Ref to both inputs. Connect a 1M resistor from each input to Ref so that the instrumentation amp doesn't go crazy when the inputs are disconnected from a patient.

 

[littletransistor]

The patient's body conducts the Ref to both inputs. Connect a 1M resistor from each input to Ref so that the instrumentation amp doesn't go crazy when the inputs are disconnected from a patient.

I see. Here's my *almost* final configuration, I can get at least a Vcc/2 on the DC output, but I still couldn't get it to squeeze out an EKG signal. Maybe this thing hasn't finished grounded yet.

Btw I left out the low pass filter because I haven't got the EKG to show some output yet. If I get the PQRST waves, then I'll add the filter.

edit: I forgot to add a plus sign in the capacitor between pin 4 and 5. The + is on the pin 5.

 

[audioguru]

I think the arms of a patient are pretty far away from the heart. Try the sensors near the heart.

 

[littletransistor]

I think the arms of a patient are pretty far away from the heart. Try the sensors near the heart.

Haha... yeah - the patient is me.

I'll try to get some real ECG electrodes btw - and also i'll position these near the heart.

edit: Yes - I attempted to improve my homemade electrodes, I solder the one cent coin (i guess it's copper??) to the wires and then those homemade electrodes to the both inputs. I placed 'em near the heart - where one the ECG electrodes placement is possible. Well, it squeezes out the waves finally, but this thing isn't that consistent. Sometimes I got the signals, and most of the times this thing couldn't still feel the electrical signals I have.

Here's the results in the attachment.(thanks to Jason Nguyen for the program!)

The red box is the exact signal that the thing once picked up, but when I moved a little of the electrodes, everything becomes out and flat!! I got these signals upon repositioning the thing so many times, but I suspect my breadboard is now playing tricks on me. (i got this for almost thrice!)

 

[audioguru]

Good. You got a heatbeat.
Now you need proper self-adhesive sensors and the conductive grease that is used with them.
Then you can add a good active lowpass filter.

Solder a circuit together on stripboard. Then it won't be intermittent like breadboards.

 

[littletransistor]

Good. You got a heatbeat.
Now you need proper self-adhesive sensors and the conductive grease that is used with them.
Then you can add a good active lowpass filter.

Solder a circuit together on stripboard. Then it won't be intermittent like breadboards.

Well, yes - I'm going to do the stripboard thing too!

and finally, I read in another datasheet, this time is AD620ANZ - and the main instrumentation amplifier has a small gain of a 7 only - then it's amplified again with another external amplifier.

Should I need another op-amp to amplify the output, but this time the instrumentation part's gain is smaller?

 

[audioguru]

The instrumentation amplifier you have is fine for a gain of 1500. The other opamps are needed as a good lowpass filter and as an active Ref to the patient.

 

[kchriste]

little T,
I was looking for something else and found this:
**broken link removed**

 

[littletransistor]

little T,
I was looking for something else and found this:
**broken link removed**

Yeah, I've seen it too. Thanks for the link anyway.

Also, I guess I need a leg driver for it to fully function. Let me search around for that right leg driver thingy...

edit: The soundcard's input impedance is 47K ohms - could it be a problem receiving the ECG signals?