Heart rate/ECG amplifier and "Patient protection" requirements

[crunx]

Hi,

I try to make a battery-operated ECG device based on a SparFun project, https://www.sparkfun.com/products/12650 and https://cdn.sparkfun.com/datasheets/Sensors/Biometric/AD8232_Heart_Rate_Monitor_v10.pdf

That board is based on AD8232 bio-potential amplifier chip. The schematics seems to be pretty straight-forward, and I feel confident it should be quite OK to base on. However, I am worried about safety, as I have learned, that any device connected to the human body has very strict requirements regarding any injected electrical current, and also that a single component failure should also be always safe / withing limits of the regulations.

The design has two resistors of 180k, R3 and R4, which are in series with the main electrodes, and a third, larger resistor in the reference electrode wire. On that SparkFun board the resistors seem to be quite ordinary chips.

Can somebody advise if that kind of standard resistors satisfy the requirements? After all, as the device would be battery-operated and can't carry fault voltages from mains, there should be very little risks. But I don't want to get in any mess with responsibilities and breaking the safety rules!

[Codeman]

A place to start...

http://electronicdesign.com/power/choose-right-resistors-medical-applications

[solutions]

While the board might be battery powered, the next board or PC likely isn't. There should be an isolator on THIS board.

If it were me, I'd also be putting two of the probe's current limiter resistors in series in the rare event one fails short.

IMO, and from what I see, it's a crap design and opens itself up to lawsuits (there are lots of those in the USA - developing world, $300 to the family and you're good to go again).

[crunx]

Thank you for your relevant comments.

I agree, the SparkFun design is not good enough as-it-is. But it seems to be based on the evaluation board schematics, and functionality is thus likely pretty close to the intended.

Regarding isolation against the "next device in line": I plan to have a radio module in this device for carrying the signal to a receiver connected to a PC. That way there will be no fault currents via any cable such as USB. So if the device is in well-isolating and closed plastics casing, there should be no other user-accessible conductive elements but the electrodes.

I agree also about using two resistors in series, as then a single fault will not drop the protective impedance to "zero". I have seen somewhere the basic rule that no foreseeable single fault must make the device dangerous.

To Codeman: Thank you for the link. I have already seen it. However, unfortunately, it describes only pretty generic stuff.

There will not be higher voltages than 3.3VDC (or maybe 4.2V if I put in a Lithium battery) in the system, and this device will not be used with for example defibrillators shocking the EKG device with high voltages. Therefore I hope "ordinary" resistors would be good enough. But the electro-medical world is pretty new to me, so therefore I am a bit worried!

[solutions]

Don't forget that it's the current that kills you, not the voltage

[bigtoy]

If you're planning on interfacing to a PC via USB, consider using a USB isolator. For example a Linear Technology LTM2884, which isolates both the data lines as well as power. (Which is neat - it can provide you with an isolated +5V for powering your circuit).

[vern]

You can't kill anyone with a battery < 5V, the body offers to much resistance, there will never be any current that can kill anyone.
Maybe you can irritate a pacemaker, but I doubt that since they are build to withstand small DC voltages. Otherwise lots of people would die from handling batteries.
There are of course big problems if you have a device that is somehow connected to the mains power.
Don't even try that, there are very high hurdles with isolation, leakage currents and so on....
Also you should not apply a low voltage that is somehow modulated because it may disturb your heart rhythm.
If you are just measuring ECG voltages and transmit them with a low power RF device like bluetooth there should be no problems at all.
But to be absolutely sure you should use the 180kohm resistors in series with the electrodes.
Best practice would be to integrate them in the wires that lead to the electrodes, that way if something breaks inside your box it cannot short any of the resistors.

[solutions]

Sorry Vern - I disagree

The body doesn't offer the resistance, the skin does. Break the skin and you have a nice salty conductor. In some individuals, a simple DC step function in current can ctrl-alt-del them. You keep mentioning voltage - it's current that matters, not voltage.

By your arguments, we can save money by totally eliminating those 180K resistors (in the wire is one of the worst places to put them, IMO - you don't seem to understand that resistors can fail short, not something breaking in the box).

You also cannot make a broadbrush statement that the circuits are DC just because they are on a battery. Any high gain amplifier can easily become an AC source, albeit generally a high frequency one relative to inducing fibrillation.  The old saying - if you want an oscillator, design an amplifier.

I like Bigtoy's suggestion.  It's practical, low complexity, and is interference immune.

[vern]

The skin offers more resistance, but the body also offers resistance.
To get a mild jolt you need about 10mA, > 100mA can be deadly.
To have 10mA from a 5V source the resistance has to be 500 Ohms, I don't see that.
Test it with an Ohmmeter from one side of the tongue to the other (very wet and salty) = 60kOhm with my tongue.
And the current has to go through the heart, which is not very likely since the body is conducting as a whole, there is no low resistance path right through the heart.
The Separated or safety extra-low voltage (SELV) regulation requires Voltages to be  <60V DC or <25V AC RMS.
If you have a device which is (low voltage) battery powered and you can not connect it to the mains (for example with a wall cube or a battery charger) than it is only considered dangerous if you throw it at someone.

About the resistors:
A resistor consists of 2 electrodes with a wire, a carbon or metal film between them.
There is no way such a device can short, it can open (infinite resistance) or change its resistance a certain level.
If you put a lot of power on it the resisting element can melt or burn and create a somewhat low resistant element, maybe even short.
To do that you need a lot of energy. In this case with 180kOhm you would need 300 Volt to generate 0,5W  which is the typical rating for a standard resistor.
To destroy it you would need a lot more energy over several seconds, no way with a small battery.
The only way to short such a resistor is with an external conducting object that will short the electrodes.
That is why these resistors should not be in the box, where something can break and short the electrodes.

With the oscillating amplifier: You are right, it can happen, but it would not be dangerous, only if you exceed the 25VRMS (see SELV).
Even if the person has a pacemaker it is not considered dangerous, since the oscillating voltage has to be differentially applied at certain points, which you can not do without putting electrodes directly into the heart. Pacemakers generate about 1.5 to 6V pulses, even if you manage to get a differential voltage to the electrodes it could not be destroyed.

[optikon]

Thank you for your relevant comments.

I agree, the SparkFun design is not good enough as-it-is. But it seems to be based on the evaluation board schematics, and functionality is thus likely pretty close to the intended.

Regarding isolation against the "next device in line": I plan to have a radio module in this device for carrying the signal to a receiver connected to a PC. That way there will be no fault currents via any cable such as USB. So if the device is in well-isolating and closed plastics casing, there should be no other user-accessible conductive elements but the electrodes.

I agree also about using two resistors in series, as then a single fault will not drop the protective impedance to "zero". I have seen somewhere the basic rule that no foreseeable single fault must make the device dangerous.

To Codeman: Thank you for the link. I have already seen it. However, unfortunately, it describes only pretty generic stuff.

There will not be higher voltages than 3.3VDC (or maybe 4.2V if I put in a Lithium battery) in the system, and this device will not be used with for example defibrillators shocking the EKG device with high voltages. Therefore I hope "ordinary" resistors would be good enough. But the electro-medical world is pretty new to me, so therefore I am a bit worried!

Crunx, in light of the resistor discussions and skin resistance, you can probably sleep at night. You have a battery powered thing with a remote transmitter and fault protection, I find it highly unlikely this could ever harm anybody. Resistors are so stinking cheap, put a bunch in series for double fault protection. I think there is a US medical standard (I think) that allows for a "leakage" current of something like 100uA or so... you ought to look that up..

However, like cracks, use for educational purposes only and never on nice primates.. well perhaps it could be used on smoking chimps since clearly, safety is of no concern there.

Find the standards, I dont think anybody in the thread is saying they know how to make compliant medical devices (I'm certainly not) so the advice is worth what you paid for it.

[vern]

I agree with optikon, the device is totally safe as long as you don't connect it to a power outlet.
I would use it, no problem.
With the smoking monkey I don't know......he might throw the thing at you!

[cadence]

@crunx,

Hi, your original question was only related to the resistors, however there have been a couple of ECG projects printed in Elektor magazine over the last few years that you may be interested in if you haven't seen them already.

A pretty basic Wireless ECG monitor using ZigBee published in the January 2011 issue, and a far more comprehensive project - an Android Cardioscope using Bluetooth that was published over three issues in 2013: July/August + September + October.

Combined with the advice and opinon offered by other members I hope this helps.