Connectors compliant with medical equipment standards

[TimNJ]

Hi all,

I'm working on my EE senior design project which is a data acquisition system for EEG and ECG signals. While I do not have the time or tools to create a product that would pass all the stringent medical standards, I would like to take as many of them into consideration that I can. I'm planning on using coaxial cable (maybe RG-174) for my electrode cables using the shield as a guard. However, I have never seen BNC or SMA used in a piece of medical equipment.

If anyone has any experience in the medical equipment field, why would or wouldn't it be appropriate here? Hypothetically, would it fail the IEC 60601 standard (to use a coaxial connector)?

Thanks,
Tim

[TheWelly888]

I service and repair medical equipment and I can tell you that I have never seen patient connections using coaxial plugs and sockets but plenty of thin coax cables used as leads for ECG. The outer conductor is used for screening of sub 1mV signals found in ECG / EEG. all the screens comes together into one at the plug end and the individual conductors will have a pin on the plug - usually ECG and EEG have more than one lead.

There is NO standard connector for medical equipment! ECG and EEG signal connectors are usually bespoke designs simply because you MUST have electrical isolation between the input and the processing board capable of withstanding at least 250V on the patients lead - don't forget to look up IEC60601 and IEC62353.

I have seen 15 and 25 way D type sockets used and still being used for ECG inputs. Lemo connectors are used as well.

[TimNJ]

I service and repair medical equipment and I can tell you that I have never seen patient connections using coaxial plugs and sockets but plenty of thin coax cables used as leads for ECG. The outer conductor is used for screening of sub 1mV signals found in ECG / EEG. all the screens comes together into one at the plug end and the individual conductors will have a pin on the plug - usually ECG and EEG have more than one lead.

There is NO standard connector for medical equipment! ECG and EEG signal connectors are usually bespoke designs simply because you MUST have electrical isolation between the input and the processing board capable of withstanding at least 250V on the patients lead - don't forget to look up IEC60601 and IEC62353.

I have seen 15 and 25 way D type sockets used and still being used for ECG inputs. Lemo connectors are used as well.

Thank you so much for your expertise. So to clarify a few things (because I'm a dunce!), why couldn't a BNC connector be isolated from the processing board? Because it's usually tied to a metal chassis?

By isolation between the input and processing board, does that mean between the analog front-end circuit and the processing board? I'm thinking the analog front-end will basically be an instrumentation amplifier and optoisolator. What's the point of this specification? (How does it make it more safe?)

Thank you!

[TheWelly888]

Thank you so much for your expertise. So to clarify a few things (because I'm a dunce!), why couldn't a BNC connector be isolated from the processing board? Because it's usually tied to a metal chassis?

Not necessarily to do with isolation because BNCs are physically large connectors which can take up way too much panel space when you have multiple connections to make and it is also inconvenient for the clinical user (nurse or technician) to be making multiple connections with scope for error. Plugs with multiple pins and sockets reduces errors resulting from misconnections.

By isolation between the input and processing board, does that mean between the analog front-end circuit and the processing board? I'm thinking the analog front-end will basically be an instrumentation amplifier and optoisolator. What's the point of this specification? (How does it make it more safe?)

Yes, the isolation can be between the analogue front end and the processor board (remember to isolate the power supply as well!)

The point of the specification is to reduce currents through the patient's skin to levels that will not cause injury and you will be surprised that level is as low as 10 micro amps! ie 10^-5 amp. Also the equipment is expected to keep the current down to non lethal if the patient somehow gets into contact with 240V mains electricity (imagine a patient attached to an ECG monitor who lashes out at a bedside lamp and contacts mains as an extreme example) Galvanic isolation helps with this.

This website has a list of short articles on electrical safety in medical equipment which I hope you will find useful. http://www.ebme.co.uk/articles/electrical-safety

Hope you find it useful. Good luck!

[Pjotr]

Since you do not have the time or tools to create a product that would pass all the stringent medical standards I would say cut corners directly: Use electrodes + cabling from established manufacturers in this field. For safety use approved isolation amplifiers off the shelf. AnalogDevices has a long history of it, has a large range of building blocks for ECG/EEG sensing you can put straight on a PCB and pricing is reasonable. TI (ex Burr Brown) has also some.

[TimNJ]

Thank you so much for your expertise. So to clarify a few things (because I'm a dunce!), why couldn't a BNC connector be isolated from the processing board? Because it's usually tied to a metal chassis?

Not necessarily to do with isolation because BNCs are physically large connectors which can take up way too much panel space when you have multiple connections to make and it is also inconvenient for the clinical user (nurse or technician) to be making multiple connections with scope for error. Plugs with multiple pins and sockets reduces errors resulting from misconnections.

By isolation between the input and processing board, does that mean between the analog front-end circuit and the processing board? I'm thinking the analog front-end will basically be an instrumentation amplifier and optoisolator. What's the point of this specification? (How does it make it more safe?)

Yes, the isolation can be between the analogue front end and the processor board (remember to isolate the power supply as well!)

The point of the specification is to reduce currents through the patient's skin to levels that will not cause injury and you will be surprised that level is as low as 10 micro amps! ie 10^-5 amp. Also the equipment is expected to keep the current down to non lethal if the patient somehow gets into contact with 240V mains electricity (imagine a patient attached to an ECG monitor who lashes out at a bedside lamp and contacts mains as an extreme example) Galvanic isolation helps with this.

This website has a list of short articles on electrical safety in medical equipment which I hope you will find useful. http://www.ebme.co.uk/articles/electrical-safety

Hope you find it useful. Good luck!

Thanks for all of the useful information. This is going to be a battery powered device so I guess it will be isolated by nature. Not sure if the isolation rule still applies there. Great link!

[TimNJ]

Since you do not have the time or tools to create a product that would pass all the stringent medical standards I would say cut corners directly: Use electrodes + cabling from established manufacturers in this field. For safety use approved isolation amplifiers off the shelf. AnalogDevices has a long history of it, has a large range of building blocks for ECG/EEG sensing you can put straight on a PCB and pricing is reasonable. TI (ex Burr Brown) has also some.

Thanks. Surprisingly, looking through a bunch of instrumentation amplifier datasheets, one of the most popular application circuit is an ECG amplifier. I do see that TI literally has an analog front end chip for ECG. ECG, however, is relatively easy since the signals might be as large as 5mV. On the other hand, EEG might peak at 100uV. The noise/interference susceptibility is much higher for EEG. With that being said, there are a number of options to address this. The first is active electrodes which places a tiny instrumentation amp directly at the probe. The next is to use a shielded cable and use a driven guard. As you suggest, I want to use off the shelf (passive) probes, but I want to have acceptable performance, so I am planning on going to shielded cable route. There seems to be a decent amount of literature/documentation on it so it seems plausible to design and get a board made by next May. (My two other teammates are doing the embedded DSP work and app software.)

I will look into the isolation amplifiers. I was planning on using an IL300 linear optocoupler but if there's an easier way, then I'm game!

[Pjotr]

It is a couple of years ago I was into this. I remember some of the Analog Dev isolation bricks also had +15V and -15V outputs of a couple of mA at the isolated side. It was intended for use of extra signal conditioning at the isolated side.

[m98]

Thanks for all of the useful information. This is going to be a battery-powered device so I guess it will be isolated by nature. Not sure if the isolation rule still applies there. Great link!

No, in this case, you don't need any isolation, as long as there is no other connector on the device. But even a single USB-port for charging and data exchange will again ruin that and require an isolated patient frontend. Also, depending on the application, consider to make it defibrillator safe (remember implanted defibrillators).

[TimNJ]

Thanks for all of the useful information. This is going to be a battery-powered device so I guess it will be isolated by nature. Not sure if the isolation rule still applies there. Great link!

No, in this case, you don't need any isolation, as long as there is no other connector on the device. But even a single USB-port for charging and data exchange will again ruin that and require an isolated patient frontend. Also, depending on the application, consider to make it defibrillator safe (remember implanted defibrillators).

Man being safe just makes everything so difficult!  There are no plans for USB connectivity but charging is a different story. I wonder if there would be any performance advantage to having separate analog and digital grounds anyway.

It's definitely going to be more work to include isolation but it may be worth it. I suppose someone might want to have it plugged in while using it...but I also know the danger of requirements/scope creep.

[stj]

you need to talk to connector company's,

as far as probes and sensors go, the plugs are often sealed because they must put the whole assembly in an autoclave to sterilize it..

you cant use radiation like you can to sterilize passive stuff like forceps or tweezers.

[TimNJ]

you need to talk to connector company's,

as far as probes and sensors go, the plugs are often sealed because they must put the whole assembly in an autoclave to sterilize it..

you cant use radiation like you can to sterilize passive stuff like forceps or tweezers.

Good points. Didn't know about that. This is definitely not going to be a production model, so I might be able to let that slide for now.

[Pjotr]

This is definitely not going to be a production model, so I might be able to let that slide for now.

Nevertheless you have to guarantee it is safe for the patient. For that you have to do the required risk management as well. In this case you will mainly concentrate on leakage current. What the requirements are depends also on the class of medical equipment --> IEC 60601-xxx.

So where you put isolation barriers is an investigation by itself. You can do it at the analogue level by isolation amps. But at the digital level after AD conversion, can be more economic to day. Or even isolate the whole apparatus if the application tolerates it.

[gasman5280]

I am not an electrical engineer but rather an anesthesiologist who tinkers in electronics.  Ten milliamps applied in the right spot can put a patient into cardiac arrest (this would have to be done via an open break in the skin or via an invasive line providing some sort of conductive path). Most ORs in the US have line isolation monitors that measure leakage current (though they are no longer required by code.  See:  https://vam.anest.ufl.edu/simulations/orelectric.php).  Electrical isolation is critical.

There is no standard connector for medical equipment though some monitors are somewhat standardized, most often they are not interchangeable.  Most sensors do not get routinely autoclaved, they are either thrown away, reprocessed, or wiped down with disinfectant.

Some equipment needs to non-ferromagnetic and/or able to operate around high magnetic fields (i.e. in an MRI scanner). Hope this helps. 

ISO 13845 and 60601 apply to medical device safety, you may be able get access to them through your university.  If you want to bring prototype or experimental equipment into a patient care area, you will have to go through the healthcare facility's biomedical department (if your facility has one).

Good luck with your project.

[NiHaoMike]

Man being safe just makes everything so difficult!  There are no plans for USB connectivity but charging is a different story. I wonder if there would be any performance advantage to having separate analog and digital grounds anyway.

It's definitely going to be more work to include isolation but it may be worth it. I suppose someone might want to have it plugged in while using it...but I also know the danger of requirements/scope creep.

You can use wireless charging to provide isolation. Then you'll have to validate that the wireless charging does not cause interference. Or if there's no need for it to be in use for very long times, design it so that it would be physically impossible to connect the charger and probes at the same time.

[ian.rees]

If the device is battery powered, it's probably easiest to handle isolation by making the battery removable, and the charger a different assembly.  -Ian- 

[TimNJ]

I am not an electrical engineer but rather an anesthesiologist who tinkers in electronics.  Ten milliamps applied in the right spot can put a patient into cardiac arrest (this would have to be done via an open break in the skin or via an invasive line providing some sort of conductive path). Most ORs in the US have line isolation monitors that measure leakage current (though they are no longer required by code.  See:  https://vam.anest.ufl.edu/simulations/orelectric.php).  Electrical isolation is critical.

There is no standard connector for medical equipment though some monitors are somewhat standardized, most often they are not interchangeable.  Most sensors do not get routinely autoclaved, they are either thrown away, reprocessed, or wiped down with disinfectant.

Some equipment needs to non-ferromagnetic and/or able to operate around high magnetic fields (i.e. in an MRI scanner). Hope this helps. 

ISO 13845 and 60601 apply to medical device safety, you may be able get access to them through your university.  If you want to bring prototype or experimental equipment into a patient care area, you will have to go through the healthcare facility's biomedical department (if your facility has one).

Good luck with your project.

Thanks! Really appreciate the insider information. I think I will see if those standards are available. I think we'll just test it on ourselves  , but I'd like for all of us to make it out alive if possible. 

[TimNJ]

If the device is battery powered, it's probably easiest to handle isolation by making the battery removable, and the charger a different assembly.  -Ian-

Hey that's a good idea! So the isolation we are talking about here is preventing the probe tips from going live (120/240VAC) and finding a return path through the earth?

What do you guys think about using an isolated DC-DC converter (to power the front-end) and a linear optocoupler (for the biosignals)?

Here's the type of converter I was thinking about: http://power.murata.com/en/products/dc-dc-converters/isolated.html

I don't want to complicate the project (because I know I'm really good at that) but if I'm going to get a PCB made (will be my first one, though), I figure I should be able to get this all working by May. I may be optimistic especially since I'm inexperienced and trying to design a board that will amplify uV level signals without signal integrity issues. I do have a lot of good resources like Henry Ott's EMC book (which I am perusing).

Thanks for all of the great input.

[m98]

Right, use one of those isolated DC-DC converters for the frontend, and carry the digital output signal of the frontend (ADC / integrated frontend) over an optocoupler with at least 400V isolation voltage. Alternatively, you could just isolate the processor board from the power supply by using an isolated DC-DC converter right where the power comes in.