The capacitors in single-fault condition

[HamidRNZ]

Hi,
I'm working on a medical product where we have several capacitors and a 250 mA fuse.
How can we test single-fault conditions on the capacitors?
Shall we short-circuit them or change them with a resistor (to drain up to 250 mA)?

Thanks in advance for your help.

 

[HamidRNZ]

Do I need to perform a test of the single fault condition of every component, such as LDOs etc., to see how much they can increase the surface temperature?

 

[FelipeSchneider]

You need to better specify few points:

• Where are the capacitors place at?
• Are they connected to the patient (or have a risk to be)?

If they are not certified X or Y capacitors I would, initially, say that you should evaluate what happens if they fail and create a short-circuit or an open circuit.

 

[HamidRNZ]

You need to better specify few points:

• Where are the capacitors place at?
• Are they connected to the patient (or have a risk to be)?

If they are not certified X or Y capacitors I would, initially, say that you should evaluate what happens if they fail and create a short-circuit or an open circuit.

The capacitors are not connected to the patient and are only on the power line (Ceramic 0402 or 0201 capacitors).
My question also is if the surface temperature exceeds the limit when one of the components fails or burns, such as an LDO or the microcontroller, do we have to do anything? Do we have to consider this failure for surface temperature according to IEC 60601-1?

 

[Peter Selvey]

The focus of the standard is on parts that are critical for safety and in particular parts that are used to protect against electric shock and other serious harm.

For the purpose of a third party test lab (not involved in the design) for other parts such as electronics (SMD capacitors, LDO, micros, resistors etc) it is assumed that a typical design will keep these well away from limits, mainly because there is not time to assess every parts, and because the parts are usually used well below specifications and are not located near heat sources. In special cases there might be for example a power resistor or power transistor, or electronics located near a heater that might get some attention from the lab, but otherwise it is assumed to be OK.

However, under Clause 4.8 and under normal condition, all parts should be used within specified temperatures. So if you know there is a 0402 capacitor that is getting to 108°C and the part manufacturer specifies the capacitor as maximum 85°C, it is a failure in IEC 60601-1, as well as being pretty bad design. The standard allows the use of risk management, but really there is no reasonable justification in this case.

Under abnormal/fault condition there is no temperature limits for electronic parts used in practice (in theory there is, but not in practice). The part itself can anyhow be considered to fail, and flammability is covered by the use of flame retardants in the PCB, wiring, case.

In the case of high risk systems (where faults in electronic control system could otherwise cause serious harm, death) there is normally a second independent system (protection system) that can prevent serious harm. Of course, consideration should be given to a single fault leading to both control and protection system electronic components exceeding temperature limits.

 

[HamidRNZ]

Thanks for your quite good reply.

If an LDO burns and takes 60 mA (Fuse cuts in 250 mA, then the fuse will not react) can heat up the surface temperature (our conservation is only surface temperature). According to 60601-1, the surface temperature of the applied part shall not exceed 43 degC (plastic enclosure), even in S.F.C. But the surface temperature will exceed 43 degC. Can it be taken as a failure of the system for surface temperature? The surface temperature can increase to 60 degC.

Does the Single Fault Condition apply only to the critical components, or any other components can be considered?
What is the HARM? Exceeding the surface temperature can be a HARM?

Thanks in advance.

 

[Peter Selvey]

This is a little complicated.

It is good to stay below 43°C in normal and SFC for applied part temperature, to keep it simple.

If the temperature reaches 60°C in SFC, there may be a few possibility:

1) it may be that this temperature is reached when tested in open air. If tested on a real person and it is only a small heat source, the temperature rise on the body may be << temperature rise in air as the human body can absorb the heat.

2) it may be due to specification of 40°C use, which means that only 3°C rise is over the limit. This is not correct, it should be temperature rise above body temperature, which can be assumed to be 35°C for skin (see IEC 60601-2-37 for ultrasounds)

3) if may be that the use is short term, especially in SFC. Temperatures around 41-45°C can only cause harm if the contact time is very long (hours). However temperatures above 45°C are increasingly dangerous and a real temperature of 60°C is a concern even for short periods.

In my opinion, applied parts contacting unbroken skin should be allowed 8°C rise (NC, SFC) when tested in open air. This assumes skin temperature of 35°C and allowable limit of 43°C. If the open air test exceeds 8°C, there should be the option to test with an actual human body or simulated tissue. However, in this case the allowable rise should be reduced to 5°C to allow for uncertainty in the test. Also higher limits can also be considered if the application time is reliably short.

 

[HamidRNZ]

This is a little complicated.

It is good to stay below 43°C in normal and SFC for applied part temperature, to keep it simple.

If the temperature reaches 60°C in SFC, there may be a few possibility:

1) it may be that this temperature is reached when tested in open air. If tested on a real person and it is only a small heat source, the temperature rise on the body may be << temperature rise in air as the human body can absorb the heat.

2) it may be due to specification of 40°C use, which means that only 3°C rise is over the limit. This is not correct, it should be temperature rise above body temperature, which can be assumed to be 35°C for skin (see IEC 60601-2-37 for ultrasounds)

3) if may be that the use is short term, especially in SFC. Temperatures around 41-45°C can only cause harm if the contact time is very long (hours). However temperatures above 45°C are increasingly dangerous and a real temperature of 60°C is a concern even for short periods.

In my opinion, applied parts contacting unbroken skin should be allowed 8°C rise (NC, SFC) when tested in open air. This assumes skin temperature of 35°C and allowable limit of 43°C. If the open air test exceeds 8°C, there should be the option to test with an actual human body or simulated tissue. However, in this case the allowable rise should be reduced to 5°C to allow for uncertainty in the test. Also higher limits can also be considered if the application time is reliably short.

Then a failure in an LDO that can end up exceeding the surface temperature should be considered. Am I right?
My concern is that electronic circuits are complex, and anticipation of all failures is almost impossible.
The contact area is almost too small, almost double of a coin.

 

[Peter Selvey]

Yes, the failure should be considered. My point is that the limit of 43°C is a bit simplistic.

If the part when actually in contact with the patient skin or tissue and the the patient skin/tissue (not the LDO) reaches 60°C in single fault condition then it is a problem. It will cause a burn very quickly.

Usually the solution is to avoid having that kind of part next to the patient skin/tissue. That way you don't have to worry.

By the way, what is "LDO"? I was thinking low drop out regulator but realised it is O not R.

 

[HamidRNZ]

A low dropout regulator. Thanks.