Is it OK(ish) to treat garage MCBs as switches?

[JohnW2]

61095, and 60947-5-1 (which is one I work on) have extra tests that are applied when the manufacturer wishes to claim that his device is suitable for isolation.

Thanks. As I said, BS7671 seems to think that they are not suitable for isolation (which I suppose is true unless those extra tests are undertaken). Is there something about these devices which means that thy are not necessarily suiable for isolation?

50428 devices are electronic. 60669-1 doesn't seem to address isolation at all.

Again, BS7671 seems to think that many of the devices to 60669-x-y are not suitable for isolation - is that perhaps just because the Standards don't say that they are suitable? What sort of devices are these?

... and, as a matter of interest, what sort of tests of 'isolation' are usually specified?

Kind Regards, John

 

[stillp]

Designing devices suitable for isolation involves separation distances, creepage and clearance distances, suitable insulation materials (comparative tracking indexes) and some other things. That costs money, and is largely unnecessary if there is no commercial advantage in claiming suitability for isolation. The verification would include examination of the materials and the distances, as well as an impulse voltage test. Some pre-conditioning is usually specified to ensure that the distances are adequate when pollution and temperature effects are present.

 

[JohnW2]

Designing devices suitable for isolation involves separation distances, creepage and clearance distances, suitable insulation materials (comparative tracking indexes) and some other things. That costs money, and is largely unnecessary if there is no commercial advantage in claiming suitability for isolation. The verification would include examination of the materials and the distances, as well as an impulse voltage test. Some pre-conditioning is usually specified to ensure that the distances are adequate when pollution and temperature effects are present.

Thanks - so, as I rather suspected, it's really primarily a matter of design and construction, with no much in the way of 'testing' that can be done?

As I've said, despite what you say about 'commercial advantages' (or the lack of them), BS7671 appears to believe that most devices used for switching are also suitable for isolation.

Kind Regards, John

Kind Regards, John

 

[stillp]

The impulse voltage test is considered to be an adequate verification of the isolation capability for many devices.

I think it would be more precise to state that most of the devices listed in BS7671 as suitable for switching are also suitable for isolation, since there are many devices used for switching that don't get a mention in that table.

 

[plugwash]

Thanks - so, as I rather suspected, it's really primarily a matter of design and construction, with no much in the way of 'testing' that can be done?

The trouble with insulation testing is at any reasonable test voltage (even the 4KV used in medical grade type testing) there are insulation structures that will easilly pass the HV test but are nevertheless unsuitable for safety insulation.

 

[JohnW2]

The impulse voltage test is considered to be an adequate verification of the isolation capability for many devices.

I can believe that. I asked the question because I couldn't really think of many ways to test isolation capability.

I think it would be more precise to state that most of the devices listed in BS7671 as suitable for switching are also suitable for isolation, since there are many devices used for switching that don't get a mention in that table.

Well, yes - particularly given that BS7671 was my only source, that is really what I meant. However, given the context of this discussion (and, indeed, of this forum), I would imagine that the list in BS7671 probably includes most of those switching devices which are likely to be encountered in a domestic setting, doesn't it?

Kind Regards, John

 

[JohnW2]

Thanks - so, as I rather suspected, it's really primarily a matter of design and construction, with no much in the way of 'testing' that can be done?

The trouble with insulation testing is at any reasonable test voltage (even the 4KV used in medical grade type testing) there are insulation structures that will easilly pass the HV test but are nevertheless unsuitable for safety insulation.

Fair enough - but, in the final analysis, how else can one actually 'test' suitability for isolation other than by demonstrating an ability to 'resist voltage'?

Kind Regards, John

 

[stillp]

The trouble with insulation testing is at any reasonable test voltage (even the 4KV used in medical grade type testing) there are insulation structures that will easilly pass the HV test but are nevertheless unsuitable for safety insulation.

Could you give some examples?

 

[plugwash]

Thin layers of dielectrics that are in some way subject to wear and tear is a big one. It's easy to make a dielectric that is thick enough to withstand the test voltages but not thick enough to withstand normal wear and tear.

Contamination is also a big issue. Dry air and clean surfaces make for very good insulation, contaminated surfaces much less so. AIUI this is the reason that specified creepage and clearance distances are as big as they are.

Another big issue is mechanical stability. Making sure that those creepage and clearance distances are actually maintained through a lifetime of use.

 

[JohnW2]

Thin layers of dielectrics that are in some way subject to wear and tear is a big one. It's easy to make a dielectric that is thick enough to withstand the test voltages but not thick enough to withstand normal wear and tear. ... Contamination is also a big issue. Dry air and clean surfaces make for very good insulation, contaminated surfaces much less so. AIUI this is the reason that specified creepage and clearance distances are as big as they are. ... Another big issue is mechanical stability. Making sure that those creepage and clearance distances are actually maintained through a lifetime of use.

Well, yes, but they are all essentially matters of design,which clearly require a lot of consideration. As I said, when it comes to specifying tests, I can't see what one can really do other than demonstrate that it can 'resist voltage', under the intended in-service environmental conditions. I suppose one might try to repeat those tests in devices which had been subjected to mechanical stresses and environmental contamination - but I would have thought that it would be extremely difficult to specify such stresses/contamination in a manner that would result in particularly meaningful tests.

Kind Regards, John

 

[plugwash]

Maybe I misinterpreted your question, I thought what you were asking is "given a random switch is there any reasonable way to test if it's suitable for safety isolation?" and the answer to that is no. You ran run HV insulation tests on it but passing those tests is no gaurantee of it's suitability.

 

[JohnW2]

Maybe I misinterpreted your question, I thought what you were asking is "given a random switch is there any reasonable way to test if it's suitable for safety isolation?" and the answer to that is no. You ran run HV insulation tests on it but passing those tests is no gaurantee of it's suitability.

Yes, I think you misinterpreted - maybe because I wasn't clear enough. I was really suggesting much the same as you - that there is not all that much (beyond the obvious) that one can do in the way of specifying tests for 'suitability for isolation' - so that the main criteria for deciding that suitability will probably have to relate to design, rather than 'tests'.

I suspect that tests alone would probably suggest that almost any device suitable for functional or emergency switching was also 'suitable for isolation' (as are most of those mentioned in BS7671) - but, as you have said, that probably isn't necessarily the case. Indeed, if it were, then in terms of the devices used, we probably would not need to consider 'isolation' as something different from 'switching', even though (as stillp has pointed out) the purpose/intent of those two things are very different.

Kind Regards, John

 

[stillp]

Maybe I misinterpreted your question, I thought what you were asking is "given a random switch is there any reasonable way to test if it's suitable for safety isolation?" and the answer to that is no. You ran run HV insulation tests on it but passing those tests is no gaurantee of it's suitability.

That's why the tests require preconditioning.