Medical Electronics

[littletransistor]

Hey there,

Any begineer home-made Medical Electronic project that I could start with?

I love to do the ECG/EKG machine because it contains some of the components that I have, like those uA741 and some caps and some resistors.

So what safety precautions should I take? I heard that you cannot have a current leakage of more than a few microamps to the body.

 

[ssylee]

Safety precautions that you should take should be protection diode to have leakage current going through the diode instead of through your body when you connect the terminals from your body into an instrumentation amplifier for ecg/ekg.

 

[audioguru]

The 741 opamp is too old to be used in an ECG or EKG circuit. They use instrumentation amps.

 

[ssylee]

I currently use LT1167 at work, so that's one part number that would be recommended.

 

[Dean Huster]

Professional medical monitoring equipment has isolation out the kazoo between patient and equipment. Not only is the isolation necessary for the obvious -- patient safety -- but is also necessary so that unwanted signals from the mains does not enter the input circuitry and swamp the tiny electrical signals generated by the body.

Some manufacturers have developed clever and/or unique methods to achieve isolation while maintaining a lot of versatility by using "mainframes" into which varied "plug-ins" can be inserted to customize the monitor for a specific application or applications. Even 20 years ago, Siemens had plug-ins with absolutely no electrical connections whatsoever. Power to the plugin was magnetically-coupled; the primary and half of the core of a small power transformer was inside the "mainframe" and the secondary and the other half of the core was in the plug-in, both electrically insulated from each other, yet in intimate-enough proximity to effectively couple all the power needed to the plug-in. Data was fed into and out of the plug-in via infrared analog or digital coupling rather than using connectors. Isolation was just one advantage. With no connectors, you didn't have to worry about contact wear, intermittent contacts, dirt, oxidation and all the other problems that plague connectors.

But for you, the number one concern is obviously isolation and a total lack of leakage. Don't try to get close to some "magic" "few microamps" of current. Keep it in the nanoamp range or less. The use of optoisolators and high-grade power transformers will be a must. Even though they made what was considered one of industry's best neonatal monitors, Tektronix got out of the medical monitor business (sold off the entire business division who got to keep the Tektronix name) because of the tremendous liability costs inherent in such items. Ergo, be really, really careful. As a teacher, I'd NEVER allow a student to design, make or experiment with any electronic device, line- or battery-powered, that was directly connected to a person's body. It's just not worth the risk.

Dean

 

[littletransistor]

Professional medical monitoring equipment has isolation out the kazoo between patient and equipment. Not only is the isolation necessary for the obvious -- patient safety -- but is also necessary so that unwanted signals from the mains does not enter the input circuitry and swamp the tiny electrical signals generated by the body.

Some manufacturers have developed clever and/or unique methods to achieve isolation while maintaining a lot of versatility by using "mainframes" into which varied "plug-ins" can be inserted to customize the monitor for a specific application or applications. Even 20 years ago, Siemens had plug-ins with absolutely no electrical connections whatsoever. Power to the plugin was magnetically-coupled; the primary and half of the core of a small power transformer was inside the "mainframe" and the secondary and the other half of the core was in the plug-in, both electrically insulated from each other, yet in intimate-enough proximity to effectively couple all the power needed to the plug-in. Data was fed into and out of the plug-in via infrared analog or digital coupling rather than using connectors. Isolation was just one advantage. With no connectors, you didn't have to worry about contact wear, intermittent contacts, dirt, oxidation and all the other problems that plague connectors.

But for you, the number one concern is obviously isolation and a total lack of leakage. Don't try to get close to some "magic" "few microamps" of current. Keep it in the nanoamp range or less. The use of optoisolators and high-grade power transformers will be a must. Even though they made what was considered one of industry's best neonatal monitors, Tektronix got out of the medical monitor business (sold off the entire business division who got to keep the Tektronix name) because of the tremendous liability costs inherent in such items. Ergo, be really, really careful. As a teacher, I'd NEVER allow a student to design, make or experiment with any electronic device, line- or battery-powered, that was directly connected to a person's body. It's just not worth the risk.

Dean

Thanks for the advice. I'll do some more research on these stuff.

Okay, so the nearest instrumentation amplifier is the LT1167, if I want to construct my little ECG machine (using Jason Nguyen's example - he has made one too but with a few opamps together).

Is one instrumentation amplifier represents the whole thing, like **broken link removed**

or must I have to combine it one-by-one? (that means 3 instrumentations amp together)

 

[audioguru]

An instrumentation amp has the opamps and perfectly matched resistors inside.
An ECG circuit feeds back to the patient some of its zero signal instead of having a ground.

 

[littletransistor]

An instrumentation amp has the opamps and perfectly matched resistors inside.
An ECG circuit feeds back to the patient some of its zero signal instead of having a ground.

Thanks. Looks like I'm purchasing the simple versions of the LT1167s first.

Jason Nguyen's Homemade ECG

**broken link removed**

^ Here's the link, and his circuit, he has constructed the whole amplification circuit that is similar to that of an instrumentation amplifier. Can I just substitute the whole 3-opamp area into one LT1167?

 

[audioguru]

You can substitute anything you want if you understand the circuit.

 

[littletransistor]

You can substitute anything you want if you understand the circuit.

Hey there,


sorry for not replying earlier - i'm in the midst of developing a simple EKG system. So far it has been good, but there's still noise picked up. I've already 'installed' a low-pass filter to get rid of 60Hz hum, but there's still a small waveform 'riding' on my ECG signals. I guess I need a hi-pass filter.

Some ECG kits actually did their hi-passes using op-amps, but can I just do the conventional RC way?

 

[audioguru]

An instrumentation amplifier IC has matched resistors inside for extremely good common-mode rejection. 60 cycle hun is a common-mode signal that is rejected quite well.

You don't want a highpass filter because the signal you want is at an extremely low frequency.

Many ECG circuits use an active 0V feed from the output of an opamp to the patient. This also reduces hum pickup.

 

[littletransistor]

An instrumentation amplifier IC has matched resistors inside for extremely good common-mode rejection. 60 cycle hun is a common-mode signal that is rejected quite well.

You don't want a highpass filter because the signal you want is at an extremely low frequency.

Many ECG circuits use an active 0V feed from the output of an opamp to the patient. This also reduces hum pickup.

I see. Hmm... I have made up my own instrumentation amplifier, but the problem lies in the weak amplification. One side of the buffer can accept the signals, but another of the side doesn't at all. I'm pretty sure the amplified voltages can be more than 4V...

 

[audioguru]

Instrumentation amp ICs have a gain up to thousands, and the gain from each input is exactly the same.

 

[littletransistor]

Instrumentation amp ICs have a gain up to thousands, and the gain from each input is exactly the same.

Hmm... maybe I should get a cheap Burr-Brown INA126PA or something equivalent.

And also, I'm powering the EKG circuit only using one (1) 9V battery - the opamp must be operating at split rails (positive and negative voltage), but could there be a single 9V battery to operate an op amp?

Currently I connect those opamps using this way:

positive terminal of opamps to positive terminal of battery, and negative to negative. Is it incorrect way?

 

[audioguru]

Any opamp will work from a single power supply voltage if its inputs are biased to a voltage near the middle. Then the input, output and feedback need to be capacitor-coupled.

Some 741 opamps don't work when the total supply voltage is 9V (they are spec'd with 30V) or when a 9V battery drops to 6V.

 

[littletransistor]

Any opamp will work from a single power supply voltage if its inputs are biased to a voltage near the middle. Then the input, output and feedback need to be capacitor-coupled.

Some 741 opamps don't work when the total supply voltage is 9V (they are spec'd with 30V) or when a 9V battery drops to 6V.

Hey there. I just reconstructed my homebrew instrumentation amplifier, but with using single supply. This time it failed to work. Here's the schematic... what could be wrong in the circuit? Btw, the two inputs are for the left and right arm together.

 

[audioguru]

You didn't bias the inputs so the input opamps didn't work.

An instrumentation amplifier has resistors that are matched, and has additional resistors so that the input opamps also have gain like this:

 

[littletransistor]

You dodn't bis the inputs so the input opamps didn't work.

An instrumentation amplifier has resistors that are matched, and has additional resistors so that the input opamps also have gain like this:

Hey, the thing you suggested amplifies, but there's another problem - I connected both electrodes (right and left arm) and the noise is attentuated, but I still couldn't get any signals. The strange thing is, when I keep touching the tip of the electrodes, the signal jumps and jumps.

I'm pretty sure the homebrew amp is working well, and I only could be missing one last thing!

What could be the problem?

Here's the attachment - the waveform spikes are caused by the repeatedly touching the tip of the electrodes. The straight line (polluted by noises) are when I connected the electrodes to both arms.

 

[audioguru]

I think Vcc/2 also must be connected to the patient.
The gain of the instrumentation amp must be 500 to 1000.

Attach your schematic with its parts values.

 

[littletransistor]

I think Vcc/2 also must be connected to the patient.
The gain of the instrumentation amp must be 500 to 1000.

Attach your schematic with its parts values.

Okay, this is the modified, simplified homebrew instrumentation amp I got from T.I's Single Supply Guide Book and the resistors values are stated already. Yet I couldn't get it working - any testing guides? I read that all of the opamp's output must be equal to the Vcc/2 if I'm not mistaken for the testings.

The gain is around 1785 - is it permissable for this operation?

Also, must I ground the output for this battery-powered system?