Shielding nessessary? Faraday, shielded wires, etc.

Hi there

I have been fooling around with a mobile electrocardiogram monitor using a MEGA2560, AD8232 and a TFT LCD touchscreen.
As I am nearly finished I have begun to wonder if I should make any effort to reduce e.g. electromagnetic interference.
So, I guess my first question is; would I, in general, stand to gain if I put in the effort? And then quickly following; what can be done? (l am also just curios in general)

I have my project running of a battery which should remove some interference, but what are some other good practises.
Something with the wires?
Short wires? Twisted or shielded wires? Double heat shrink all wires? Better quality? (currently using standard flimsy wires that comes with Arduino sets, but I have some 18AWG electrical wire lying around)
Shielding the entire project?
Enclosing the entire project in a mesh metal box to create a miniature faraday cage?

I apologise if this is all very basic, just seeking more knowledge so I can improve myself and my projects

Best regards, well me

Yo9u won't gain anything by chasing phantoms! If you don't have EMI problems feel lucky.

Do you know how to identify EMI? Do you have the equipment to detect it? IF you do, then you can begin to try to minimize it, if it is causing problems.

Paul

Hi Paul, thanks for your answer.

So are you saying I most likely will have EMI, but it might be better not to worry about it? (perfectly acceptable answer by the way)
Well in my device I would see EMI in my graphed electrocardiogram, but I have no dedicated equipment to detect it no.

There are no general rules of thumbs when working with arduino, to reduce possible interferences?

First, eliminate possible sources. Unshielded fluorescent or LED lights. No cell phones in vicinity. No electric wrist watches. No other electronic devices in the area.

Bypass capacitors on all sensor power leads. Ferrite beads on all sensor leads. Yes, use shielded wire on all wiring, but only ground at one end. Do not ever use the shield as a conductor. Be sure all the ground connections end up at the AC power outlet ground and test to be sure it is really a ground connection.

Without knowing what you are fighting, you must fight all possible interference, real or imagined.

Paul

There are two issues - The EMI from your device affecting other devices. This is what the regulatory agencies are concerned with, but you probably don't care about that. For that, you'd want to shield the Arduino and power supply, and sometimes it's necessary to add filtering (sometimes just a ferrite bead) on any leads/connections coming in & out. ...I've got at least one USB cable will a ferrite bead "lump" in the cable. That has no effect on the performance of whatever device it came with (I don't remember) but it helps reduce radiation coming out of the device through the cable and presumably helps it pass the regulatory requirements.

But, you're concerned about EMI getting into your device. The Arduino and digital electronics are pretty-much immune to EMI. Your analog electronics should be shielded and the leads to the patient should probably be shielded. It might be necessary to shield your analog electronics from your digital electronics.

I assume the biggest problem is the patient acting like an antenna... Have you ever touched an audio amplifier input and heard a buzz? That's the radiation from power lines being picked-up by your body? Or have you ever touched a radio antenna and got a better or worse signal?

So, I assume you need some "serious" filtering to filter-out everything except the "EKG frequencies". Probably just a steep low-pass filter. And that probably means an analog filter, even if there is additional digital filtering.

Since the human body makes such a "good" antenna and that's where most of your noise/interference will come from, I'm not sure if it's necessary to shield the cables.

Thanks for the suggestions and education. Well I somewhat understand ferrite beads and capacitors, could give it a go, just to see if there is any effect. Just to clarify, I do not expect wonders from the little ECG, but I am ofcourse going to try my best.

Filtering sure does seem to be the go to on the forums, but this can eliminate some information from the ECG. That said, it does seem to spand rather broad so if I ever get my head around how to add it, it might also be something to look into.

Very nice to hear your advise in general - will keep in mind for my future projects.

Patient leads are generally shielded with "active shielding". That is to say, the shield is driven by the output of a unity gain buffer from the lead itself. This largely cancels out the capacitance of the shielded cable itself. This is why the connector for a five lead ECG set has ten pins!

You often need a guard circuit to drive the shield to match the signal in the inner conductor for these high impedance
signals - this is standard practice for this sort of stuff. The guard amp keeps the voltage difference between the
signal and the shield zero, thus nulling out the capacitance and leakage currents.

Doubly shielded cable is often used - inner conductor is signal, mid shield is the guard electrode, outer shield
takes all the capacitive interference and conducts those currents away from the input amp section.

For high impedance weak signals good shielding is essential. You also need to short out any RF interference
with a suitable capacitor (100pF to 1nF typically) to ground, before the first semiconductor device,
so you don't end up having an antenna driving a diode detector circuit by accident...

The datasheet for the AD8232 has some useful hints.

Allan

Again, very nice information and learning.
And you have a very VERY valid point Allan :wink:

I think I will try to apply some of the “simpler” solutions, to see if there is any noticeable change.
On the note of improving the AD8232: The datasheet describes the adjustable high- and low-pass filters which would be great to fiddle around with. Unfortunately, it is a bit out of my league. A more hands-on-descriptive approach like “solder here” or “add component here” is more my current style. Just to clarify: it would be fantastic to learn all things possible, but there is only so much time to do with. Perhaps I will give to forum a go, to see if anyone else has fiddled with the filters and such. Then again, a device that basically works, but without the refinements (and complete understanding of why it works) also has a certain charm.

If your asking if it’s okay to use hobby grade wires in a medical device.....

I took apart a IV pump I got at a swap meet and all the wires were solid core comming out of fully epoxy encased control circuit. It was built spend its life being knocked to the ground and keep going..

Mind you, a "fully epoxy encased control circuit" making it impossible to repair smacks more of an attempt at avoiding Chinese knock-offs.

Paul__B:
Mind you, a "fully epoxy encased control circuit" making it impossible to repair smacks more of an attempt at avoiding Chinese knock-offs.

It only slows them down. Epoxy potting can be dissolved, leaving a mess, but the basic components are still there and can be guessed at. I have had a customer bring in a competitor's black epoxy conformal coated board for us to x-ray, so he could see what they were doing.

Paul

Actually I think it was to make it fluid / dust proof as it was designed to have fluid filled bags hanging over it.. Was not that black stuff, but amber.

Either way, it is not repairable.

Of course, it would have to be sent back to the manufacturer anyway ...

Slumpert:
If your asking if it’s okay to use hobby grade wires in a medical device.....

Ha ha, don't worry, I'm not that insane. And I would hardly classify my project as a medical device. If it were, I think my collegues and medical personnel would be a little on edge to say the least :smiley:

Spaziba:
Ha ha, don't worry, I'm not that insane. And I would hardly classify my project as a medical device. If it were, I think my colleagues and medical personnel would be a little on edge to say the least

You are missing the point. This is an electronic device that you propose to physically connect to a person, whether it be yourself or some naïve "volunteer".

That makes it a medical device, whether it be on a commercial basis or not (and now that I come to think of it why else would you be doing it?).

An ECG (it appears you are not American :grinning: ) connects by deliberately low impedance electrodes in order to most accurately measure the signal, and in a fashion that is designed to "connect" to the heart by straddling the chest but the converse of that is that it would take a very small fault current - less than 10 mA - to interfere with the heart function. This demands guaranteed isolation of the input circuitry from any significant voltage.

If the operating voltage is 5 V as on an Arduino, 10 mA corresponds to a resistance of 500 Ohms. This is certainly a resistance you would expect to achieve with commercial ECG electrodes (interesting point - I might just try it out myself at work tomorrow!) or even improvised ones. There are sophisticated current limiting chips designed for this purpose, but a reasonable approximation might be to have 100k series resistors in each input, mounted adjacent to the op-amp buffer which drives the active shield.

Devices such as the FitBit have less of a problem in this respect as they use even lower supply voltages and do not connect in a path which includes the heart.

10mA fault current would require a fair voltage to achieve (50 or more, probably, depending on various factors), but indeed is far above the pain/safety threshold.

0.5mA is quiet unpleasant, as my childhood experiment with a ammeter and many 9V batteries in series showed.

Anything powered from a small low voltage battery is inherently safe, if the voltage is low enough, and
this is one way to go - I think this is true already, but for safety you need to remove/disable any connectors
other than the electrode connectors (people will go and plug things in if there's a matching connector!).

I think that would be demonstrably failsafe. You certainly don't want an accessible USB socket waiting for someone to plug a dangerous cheap eBay mains USB supply into...

Paul__B:
You are missing the point. This is an electronic device that you propose to physically connect to a person, whether it be yourself or some naïve "volunteer".

That makes it a medical device, whether it be on a commercial basis or not (and now that I come to think of it why else would you be doing it?).

An ECG (it appears you are not American :grinning: )...

I see your point, even though I am running low on the naive victi...subjects. The same goes for your definition of medical devices which is a completely other subject legal-wise, and might differ between countries. And on that note, I am fairly far from any of the Americas yes :slight_smile: I would say well done, but i think it was fairly easy to guess.

Paul__B:
...it would take a very small fault current - less than 10 mA - to interfere with the heart function.

I have not really read up on this, but is this 10 mA directly through the heart, or is this when taking the body’s resistance into account? I know a single 9v battery goes a long way in animal experiments when connected directly to the heart – not so much when applied to the skin.

I do not intend anyone suspended in saline solution using my device with skin piercing electrodes, but it is good to be on the safe side. I power the device using 4 AAA alkaline batteries (So around 6v when using new batteries), all components are completely enclosed. But let's assume the batteries all short-circuit, giving full power through wires, Arduino, AD8232, through electrode wires, and finally through the electrodes to the body. Wouldn't something like a resettable fuse provide safety, or am I doing the sound that dogs do up against a wrongfully tree?

MarkT:
10mA fault current would require a fair voltage to achieve (50 or more, probably, depending on various factors), but indeed is far above the pain/safety threshold.

0.5mA is quiet unpleasant, as my childhood experiment with a ammeter and many 9V batteries in series showed.

Anything powered from a small low voltage battery is inherently safe, if the voltage is low enough, and
this is one way to go - I think this is true already, but for safety you need to remove/disable any connectors
other than the electrode connectors (people will go and plug things in if there's a matching connector!).

I think that would be demonstrably failsafe. You certainly don't want an accessible USB socket waiting for someone to plug a dangerous cheap eBay mains USB supply into...

Soooo... remove access to ports, and go nuts?