Test Voltages for a Pulse Transformer used to Isolate the Patient Connection

[Pads38]

Our device uses a transformer to isolate the patient connection from the rest of the circuits. It provides 2 MOPP from primary to secondary and the frame is earthed.

This device provides high voltage outputs so the primary circuits are connected to a (intentionally) high impedance voltage source and the output transformer then steps up the very brief pulses.

If one uses the concept of the working voltage being the sum of the peak voltages in the two windings then the answer is very high with the consequence that the requirements for dielectric strength test voltage are huge (in the order of 8kV).

Is there a better way of working out the working voltage?

I'm also struggling to work out the requirements for the turn-layer tests (15.5.2) for a pulse transformer where the duty cycle is so low. If any one has ideas on that I would be grateful.

Further question:
The pulse transformer provides 2 MOPP from the primary to the (applied part) secondary. But what is the requirement for between the secondary and the earthed frame? Does the earth of the frame count as 1 MOPP? so that the insulation just needs to provide 1 MOPP? (consider the insulation examples given in Annex J figure J.6 and J.7).

Many thanks for any suggestions.

 

[Stijloor]

Our device uses a transformer to isolate the patient connection from the rest of the circuits. It provides 2 MOPP from primary to secondary and the frame is earthed.

This device provides high voltage outputs so the primary circuits are connected to a (intentionally) high impedance voltage source and the output transformer then steps up the very brief pulses.

If one uses the concept of the working voltage being the sum of the peak voltages in the two windings then the answer is very high with the consequence that the requirements for dielectric strength test voltage are huge (in the order of 8kV).

Is there a better way of working out the working voltage?

I'm also struggling to work out the requirements for the turn-layer tests (15.5.2) for a pulse transformer where the duty cycle is so low. If any one has ideas on that I would be grateful.

Bump!

Can someone help with this?

Thank you!!

Stijloor.

 

[Peter Selvey]

For the first question: directly adding primary and secondary voltages was an error in the second edition, which I believe has been corrected in the 3rd edition. That is, the 3rd ed uses "working voltage", which means the true working voltage measured at the point which stresses the solid insulation.

For transformers, this would involve measuring the pri-sec voltages between each input / output terminals (a simple transformer with 2 input, 2 output terminals would have 4 measurements). For convenience, the maximum of these measurements becomes the reference voltage for the transformer and the dielectric strength tests based on this value.

Test of 15.5.2 can be skipped if you do the short circuit test in 15.5.1.2 directly at the transformer terminals (3rd ed is a bit confusing here that there is no reference to this in 15.5.2). It is normally skipped in any switching or pulse transformers.

1 or 2 MOPP to earth will depend on what happens if the MOPP is shorted (SFC), keeping in mind that the patient can be earthed in normal condition. This means there is a potential loop path through earth in the SFC, which typically means a second MOPP must be somewhere in the system, either in the transformer or in the applied part.

For most Type F equipment, this situation always exists, but the secondary voltage is very low (e.g. 5V). Since an F type applied part must always have 1 MOPP for mains (e.g. 230V), the isolation requirements for 1 MOPP @ 230V is greater than 2 MOPP @ 5V. So the 2 MOPP @ 5V is often not mentioned. But, if the secondary voltage is high (e.g. 100V), the situation is very different and 2 MOPP requirement needs to be studied carefully.

 

[Pads38]

Many thanks Peter.

Having just tested the voltages as you suggest we now have a better understanding of what is going on. And have been able to significantly reduce the test voltage for the primary to secondary test.

 

[sonic85]

For the first question: directly adding primary and secondary voltages was an error in the second edition, which I believe has been corrected in the 3rd edition. That is, the 3rd ed uses "working voltage", which means the true working voltage measured at the point which stresses the solid insulation.

For transformers, this would involve measuring the pri-sec voltages between each input / output terminals (a simple transformer with 2 input, 2 output terminals would have 4 measurements). For convenience, the maximum of these measurements becomes the reference voltage for the transformer and the dielectric strength tests based on this value.

Test of 15.5.2 can be skipped if you do the short circuit test in 15.5.1.2 directly at the transformer terminals (3rd ed is a bit confusing here that there is no reference to this in 15.5.2). It is normally skipped in any switching or pulse transformers.

1 or 2 MOPP to earth will depend on what happens if the MOPP is shorted (SFC), keeping in mind that the patient can be earthed in normal condition. This means there is a potential loop path through earth in the SFC, which typically means a second MOPP must be somewhere in the system, either in the transformer or in the applied part.

For most Type F equipment, this situation always exists, but the secondary voltage is very low (e.g. 5V). Since an F type applied part must always have 1 MOPP for mains (e.g. 230V), the isolation requirements for 1 MOPP @ 230V is greater than 2 MOPP @ 5V. So the 2 MOPP @ 5V is often not mentioned. But, if the secondary voltage is high (e.g. 100V), the situation is very different and 2 MOPP requirement needs to be studied carefully.

I just need clarify one thing. In the IEC 60601-1 2nd ed, the transformer Dielectric Test (equivalent to test of 15.5.2) normally skipped in any switching or pulse transformers. But in the IEC60601-1 3rd edition, it should not be skipped. If you look in Annex (General Guidance and rationale) under Sub-clause 15.5, you will find a rationale stating, for switching transformer (relied on to provide separation between Operators, Patients, etc.. and Hazards), the Dielectric test with 2X frequency (switching frequency) and the 2X voltage shall be conducted.

 

[Pads38]

The difference between Ed 2 and 3 is that the scope of the transformer section is extended to include all transformers that provide 'separation', whereas Ed 2 only mentions mains transformers.

As Peter pointed out test 15.5.2 (titled dielectric strength - usually thought of as the turns-layer test) can be avoided if a direct short on the output does not cause a hazardous condition (see 15.5.1.2).

 

[sonic85]

As Peter pointed out test 15.5.2 (titled dielectric strength - usually thought of as the turns-layer test) can be avoided if a direct short on the output does not cause a hazardous condition (see 15.5.1.2).

Yes, but this does not give you an out on not doing the Transformer dielectric test 15.5.2. There is nothing in the standard says that you can waive the Dielectric test if the Short Circuit results shows no hazards. So you still have to conduct the test unless you add a rationale in your RM (Risk management) or using a triple insulated wires

 

[raymond02302]

transformer dielectric test 15.5.2 is not a must. The test is to ensure the insulation between layers. We can waive this test, but we must do the overload/short circuit test at just the exist of the transformer before the protective device.

 

[Peter Selvey]

Actually after reading the standard closely, I agree with sonic85, there is no exclusion of the 2x test. It is a case where our reasonable expectations might have blinded us to the actual words in the standard.

It is clear the 5x test can be excluded, but the 2x test still remains. Risk management or the use of triple insulation cannot exclude this test.

During a recent CB scheme audit, the auditor made clear that for the 3rd edition, RM can only be used for equivalent safety, not to modify the standard to the point of ignoring requirements. It's still OK to modify or ignore the standard at a regulatory level, but the difference is you cannot claim to "comply" with the standard. Non-compliance with a standard usually invokes special regulatory provisions.

Triple insulation would seem to be a case of obvious compliance, i.e. don't actually do the test because it's a waste of time, but formally the compliance criteria still applies.

Having worked in the development of a high frequency insulation tester around 400kHz, this would be a difficult test for the average test lab to do, although it is feasible, as switching frequencies are usually less than 75kHz. One problem is that at higher frequency, stray winding capacitance becomes a significant load. The 2x voltage, 2 x frequency test would result in around 8x higher capacitive load (capacitive load is proportional to voltage squared, and frequency). A practical test system might drive the secondary to avoid building a high voltage high frequency amplifier.

Or, it could be just a mistake in the standard that will be fixed in Amendment 1. Time will tell.

 

[Pads38]

I've also re-read the standard and agree that the wording of 15.5.1.2 does not give an opt out for the test of 15.5.2.

Just checked a draft of amendment 1 (62A/747/INF) and both 15.5.1.2 and 15.5.2 are unchanged.

(Changes are a minor change to 15.5.1.1 to improve clarity and a change to the construction requirements of 15.5.3 effectively back to the Ed 2 requirements BUT this is a draft and has not been finalised.)