Cost-effectiveness of RCDs,SPDs, AFDDs etc., in perspective

[JohnW2]

John it is always 'scraping barrels' so far as you are concerned - always it would never happen.

No, not "never" - but, so often in relation to these discussions, "incredibly rarely".

I have spent much of my life professionally involved in activities which include the quantification of risks, and judgements (in relation to benefits and/or the cost of mitigation) about whether it is 'worthwhile'/sensible to take measures to mitigate those risks. If one took the view that any degree of risk, however minute, had to be addressed (regardless of 'cost', in the widest of senses), then the world as we know it would probably come to a halt.

A problem with such judgements is that there is so much variation in attitudes to risk, not only between individuals/ organisations/ governments/ whatever but also major (seemingly irrational) inconsistences within them. I'm sorry to be mentioning Covid again, but it is hard to avoid observing that the BBC proclaimed that 1,600 UK road deaths in a year was "scandalous" without their apparently having said anything about the fact that about that same number of UK 'Covid deaths' had occurred in the week during which they televised that proclamation.

Accidents happen when the dangerous circumstances conspire together and happen. Accidents are almost always a series of things going wrong and could have been avoided by changing just one thing in the series.

Indeedd so, and it is the probability of that series of events having gone wrong which is the crucial issue.

Are you saying you have never had a belt - if you say no, I will not believe you.

Of course I have, albeit (perhaps because of the increased 'caution' that comes with increasing age), not in the most recent many years. However, whilst I don't know how typical my experience is, I cannot recall any occasion on which I've received a shock as the result of (or solely the result of) any electrical fault (in an installation or appliance/equipment) - in every case I can think of, the shock would not have happened had it not been for some 'foolishness' on my part.

If you had been the one walking in that flooded banking hall, I am certain that would have changed your mind.

As I said, I don't think I would have touched anything in (and probably not even entered) a "flooded banking hall with a soaking wet carpeted floor" if I had know that the electrical installation within the hall was still live!

That had to have a whole series of improbable things to go wrong and a fault condition to exist, plus a certain level of incompetence, carelessness stupidity in bypassing obvious and clearly signed safety equipment, plus me to walk in, in the dark and already soaking wet from the rain outside. Any one of those improbable steps not lining up and it would not have happened. but there you go.

All true - but, again, the crucial question is how probable it is that all those things would come together. Perhaps more to the point (I've rather lost track of how this bit of the discussion started!) what I you suggesting could/should have been done to reduce that probability even further?

Kind Regards, John

 

[Harry Bloomfield]

My understanding is that the principles are broadly similar, medical devices just tend to have stricter rules.

I thought I had read somewhere, there had to be double isolation between mains and patient connections of the equipment - basically two transformers in series and yes, SMPSU's do have transformers.

 

[bernardgreen]

basically two transformers in series

or a transformer where the primary and secondary windings are on different legs of the core.

 

[Harry Bloomfield]

Of course I have, albeit (perhaps because of the increased 'caution' that comes with increasing age), not in the most recent many years. However, whilst I don't know how typical my experience is, I cannot recall any occasion on which I've received a shock as the result of (or solely the result of) any electrical fault (in an installation or appliance/equipment) - in every case I can think of, the shock would not have happened had it not been for some 'foolishness' on my part.

I had one this morning - I finally got around to looking at that new and failed mixer unit from Lidl I couldn't see any damage on it's speed control/switching pcb, so I decided to apply power straight from the mains to make further checks, pushing the buttons and steadying it one hand, other hand safely out the way, insulated floor and bench, I got a slight tingle. No fuss, I normally have very high resistance skin. I made some progress though, there was a tiny 1.6amp pcb fuse which had gone o/c of a style I have not met before. Almost certainly the thyristor had taken out the fuse, but it tested as it should. Fit a replacement fuse and test again..

 

[JohnW2]

The problem is the stray capacitance interacts with the switching operation of the converter, so create what is essentially a radio transmitter, to prevent unacceptable radio emmisions SMPSUs have to deliberately add more capacitance between input and output.

Fair enough.

There are clearly limits on what faults can be considered "single faults".

Maybe, but I would have thought the most likely of 'single faults' (in the everyday sense) would be the breakdown of insulation (e.g. between primary and secondary of a transformer, or of a capacitor between input and output).

My understanding is that the principles are broadly similar, medical devices just tend to have stricter rules.

Yes, that is probably the case, although I might suspect that the requirements for consumer products might be consdierably 'less strict'.

I think there may be an issue that BS7671 is over-simplifying and/or translating imperfectly from "appliance designer terminology" to "electrician terminology". It is very much possible for a protective earth conductor to be part of the protection and leakage management strategy even if it is not exposed directly to touch by the user and AIUI such an appliance is still considered class 1. This is common with laptop power bricks.

Maybe, and that would make total sense. If live parts are surrounded by earthed metal, then that seems to be Class I, even if that earthed metal is surrounded by (electrically redundant) insulating material, and therefore not 'exposed'. However, as I said, if it hasn't got exposed-c-ps to earth, it seems that BS7671 would not regard it as Class I.

If the user is protected only by basic insulation it would be class 0 and hence "not allowed".

Maybe that's the answer to my question (simply 'not allowed').

In practice, 'double insulation' seems fairly uncommon, such that many/most Class II items must be relying on a single layer of insulation qualifying as 'reinforced'. I have to say that (consistent with your comment "Sadly I suspect though that there is also a lot of equipment out there that fraudulently claims class 2 status.") I often wonder whether some of the plastic enclosures concerned really are robust enough to count as 'reinforced'!

Kind Regards, John

 

[JohnW2]

I thought I had read somewhere, there had to be double isolation between mains and patient connections of the equipment - basically two transformers in series and yes, SMPSU's do have transformers.

Wherever possible anything connected to a patient (and particularly if 'invasively' connected) is opto-isolated.

Kind Regards, John

 

[JohnW2]

I had one this morning - ... I got a slight tingle.

I wasn't really thinking of 'slight tingles but, as I said, it's now many years since I had a shock which I would think really qualified as a 'belt'.

The worst shocks I received in my 'foolish youth' (many decades ago!) were from what would today be called 'HV' - DC voltages in excess of 1,500V - but I somehow survived them!

Kind Regards, John

 

[Harry Bloomfield]

Wherever possible anything connected to a patient (and particularly if 'invasively' connected) is opto-isolated.

Kind Regards, John

Yes, on the sensor side and battery powered, but the mains powered units have two transformers or possibly the method mentioned by Bernard - two separate windings on a common core, though suspect that technique was only used on the much older equipment, pre-SMPSU's becoming common.

 

[Harry Bloomfield]

The worst shocks I received in my 'foolish youth' (many decades ago!) were from what would today be called 'HV' - DC voltages in excess of 1,500V - but I somehow survived them!

It's all a matter of not just the voltage, but also the amount of current/source impedance and the path where it flows through your body. Worst possible is from one side of the body across to the other side, or corner to corner.

 

[JohnW2]

Yes, on the sensor side and battery powered, but the mains powered units have two transformers or possibly the method mentioned by Bernard - two separate windings on a common core, though suspect that technique was only used on the much older equipment, pre-SMPSU's becoming common.

Yes, I think we're somewhat talking at cross-purposes. Whilst the opto-isolation I mentioned is relevant to connections to patients, it obviously is not in relation to PSUs.

Most of the medical devices I see use 'off-the-shelf IEC 60601-compliant PSUs. That Standard specifies the required performance (particularly in relation to 'leakage currents', under both normal and single-fault conditions) but, to the best of my knowledge, says nothing about how that has to be achieved. I therefore usually don't know (and don't need to know) what is going on electrically within those units to achieve the required performance - so you and/or bernard may well be right.

As you probably know, there are three 'Classes' (B, BF and CF) of medical equipment, and it is Class CF (equipment directly connected to patients) which is the most demanding - in general with a maximum permitted 'patient leakage' of 10μA for both AC and DC. For the other two classes, although the DC limit usually remains at 10μA, but for AC is 100μA.

Kind Regards, John

 

[JohnW2]

It's all a matter of not just the voltage, but also the amount of current/source impedance and the path where it flows through your body. Worst possible is from one side of the body across to the other side, or corner to corner.

Indeed. With the 'HV' PSUs I was talking about, the 'source impedance' was certainly low enough to fry me, had contact been good enough and had persisted for long enough!

Probably the most 'dangerous' thing I ever played with (but, fortunately, never got a shock from!) in my youth was a TV, I seem to recall Phillips and with a model number of "TV1", which came in a floor-standing mahogany cabinet. Virtually all subsequent TVs derived the 'EHT' for their CRT from the line-output-transformer - a very high-resistance source that seriously limited the current that could flow through a victim's body. However, this "TV1" had a mains-powered EHT (I think ~12kV) PSU (big transformer and lots of voltage multipliers) and so was potentially very lethal!

Kind Regards, John

 

[plugwash]

Fair enough.
Maybe, but I would have thought the most likely of 'single faults' (in the everyday sense) would be the breakdown of insulation (e.g. between primary and secondary of a transformer, or of a capacitor between input and output).

Which is why if you are doing this properly you don't just use any old capacitors and transformers. In the transformer, you either have split bobbins (safer and better for EMC but less efficient) or you put layers of insulating material or even earthed metal screens between the primary and secondary coil.

For the capacitors there are methods of making capacitors that are very unlikely to fail short. For example capacitors with three (or more) sets of plates rather than two so that the plates attatched to the terminals don't directly overlap. For ceramic capacitors use of "soft terminations" can dramatically reduce the risk of the ceramic cracking.

I often wonder whether some of the plastic enclosures concerned really are robust enough to count as 'reinforced'!

I'd be more worried about the transformers and capacitors not being up to scratch of about the PCB designs not respecting appropriate creepage and clearance distances than about the plastic case not being up to scratch. If the plastic case falls apart then at least you can see it has fallen apart.

 

[Harry Bloomfield]

Probably the most 'dangerous' thing I ever played with (but, fortunately, never got a shock from!) in my youth was a TV,

in my youth - Putting my finger tip in live BC lamp sockets and near colour TV EHT's to strike an arc, was one of my daft 'party tricks'. I would barely feel a tingle, if my skin was dry.

 

[JohnW2]

Which is why if you are doing this properly you don't just use any old capacitors and transformers. In the transformer, you either have split bobbins (safer and better for EMC but less efficient) or you put layers of insulating material or even earthed metal screens between the primary and secondary coil.

Indeed - but, again, BS7671 does not seem to be very concerned about such (seemingly obvious considerations - requiring of a Class II item only that the double/reinforced insulation provides an adequate barrier against physical contact with live parts. As tyiou have suggested, it seems that BS7671 has a rather restricted view of this matter!

For the capacitors there are methods of making capacitors that are very unlikely to fail short. For example capacitors with three (or more) sets of plates rather than two so that the plates attatched to the terminals don't directly overlap.

That sounds a bit like the 'trick' of putting two capacitors in series. However, in both cases, a problem is that one doesn't know when a 'first level failure' has occurred.
Edit: I could have added that one can obviously also 'put transformers in series (a second 1:1 one after the first), for the price of a bit of loss of efficiency - but, again one probably wouldn't know if/when the insulation of one of them had failed.

I'd be more worried about the transformers and capacitors not being up to scratch of about the PCB designs not respecting appropriate creepage and clearance distances than about the plastic case not being up to scratch. If the plastic case falls apart then at least you can see it has fallen apart.

Yes, I can't disagree with any of that (but the above comments about BS7671's view still apply). There are also some plastic-encased things that don't even have any outputs/sockets (or anything else touchable and conductive), in which case it doesn't really matter what happens to the transformers, capacitors or PCBs within them!

Kind Regards, John

 

[bernardgreen]

Whilst the opto-isolation I mentioned is relevant to connections to patients, it obviously is not in relation to PSUs.

I recall a design for a medical PSU that created the safe power for the sensor and it's microwatt amplifier using a photo-voltaic cell illuminated by a light source.