Wierd one....

One of the CE marking EMC related tests for double insulated down lights now requires you to look at both the 50Hz and HF currents that flow when an earth tail is added (usually with a  jumbo "jubilee" clip around the perimeter of the show face of the fitting.) I presume the test has been added to try to spot this sort of problem.


These (highly simplified) schematics show how the HF interference currents can end up circulating and putting significant unwanted voltages outside the LED driver, in earth paths that may or not be actual wiring or may be you and the lighting grid.. the middle one shows why double wound mains xformers are sometimes recommended to clean up EMC noisy loads (sometimes suggested for VFDs, note that as the whole point of the SMPS approach is to not need a mains transformer this is not done very often ) and in the last one, how inter-winding screening can help by splitting the capacitance between the winding layers into two capacitors with the centre point able to divert much of the problem current.

Thanks to everyone for the replies....


Its going to be interesting hearing what Ansell have to say.......

I will bet they will try to BS it.....

Reminds me of when whilst at sea and passing close to land, off Cape Town or Singapore Straits come to mind, I could get good mobile phone coverage both for voice and data.  Problem on one ship was that from my cabin, I needed kneel on my bed with my head up to the window glass to avoid loosing the signal - after all I lived in a large steel can... Besides often being able to "feel" the GSM signal just above my ears, I found that if I phoned whilst still connected to the charger, I was getting a tingle from the external aerial socket which was on the rear of my Ericsson GH688; my other hand being on the steel window frame.  From memory that too was about 110 volts between socket and steelwork. Swapping the polarity of the plug made no difference, which with no neutral due the sockets being across two phases of a 3-phase 230v supply, was no surprise. I cured both by purchasing a car mobile phone aerial with a magnetic base. Saved twisting my knee as well!

Clive

My rep spoke to Ansell...


They apparently used the phrase  "GHOST VOLTAGE" .....


HAHAHAHAHAHAHAHAHA.............


Going to have to find that  in the regs....lol


My plumber colleague must have been bitten by a ghost then...


I am awaiting a document from them........

Ah well, at least they've taken things in the right spirit then!

  (sorry)

    - Andy.

Going to have to find that  in the regs....lol

Well I suppose in a way it's a section 543.7 kind of thing - even if in this case the leakage current isn't passing back through the c.p.c.


I does occur to me that danger from electric shock depends on frequency as well as current/voltage. 50Hz a.c. is dangerous at a lower level than d.c. So what about these mutli-kHz drivers - should they be subject to the same 3.5mA/10mA limits as 50Hz? Might 543.7 be in need an update to catch up with high frequency switching technology (not just LED lights but VF motor drives and possibly the ubiquitous switchmode PSU etc).


I would be interesting to know what Ansell's analysis and recommended solution is.


   - Andy.

Higher frequencys should be less dangerouse because of skin effect 50 cycle currents will easily get into a person  to do there damage whereas 50 Kcs will travel over the skin maybe only going in to skin depth so not able to get to the heart

yes, but the current may read quite low on a meter that is only really tested for 50Hz - unless a true RF current measurement is being made, with the right instruments, do not assume that the current really at a safe low level. A 'normal' mains frequency meter may read either high or low.


Also the 'skin depth' for current flow  is a very funny function of frequency and which organs are involved- some part of the human body are very much better conductors than others and which bits get cooked are hard to predict in advance. There are several lifetimes work cataloguing this sort of thing. The biggest problem is that as the frequency rises there is a much lower level of nervous sensation, so you may not realise you are being shocked or irradiated at some MHz, in the same way you would feel it at say a few hundred Hz.

There may also mean there is little or no warning muscle twitch either, but there may be other more insidious effects. At certain frequencies the hot bacon smell from burning of the skin can sometimes be the first symptom.

I do not think that we are anywhere near that in this case, I'd be surprised if more than a few mA are going adrift - I expect the power supply to be chopping at maybe  10-200kHz and no more than 1000pF of transformer winding to winding capacitance, but at far high powers there have been some very nasty accidents with radio frequency plastic sheet welders and similar, that did not feel tingly.

I just love thread drift. ?


Surgical diathermy - high frequency, but also depends upon current density. So with the old-fashioned monopolar kit, you had a ground electrode (for want of a better word) which was the size of a paperback book, usually on the thigh so no burning there. By contrast, the surgeon's instrument passed all the current through a narrow spot and thereby cooked the tissues as required.


I don't know about Mike's "hot bacon", but you would certainly know about it if you got your fingers in the way. ?

Having watched in horror as an enthusiastic American tested out a repaired RF 'scalpel' burning patterns into a steak, to make sure it was performing as expected, I'd say RF surgery can be very much the same. Even more scarily similar units turn up on Ebay, which makes you wonder who buys medical kit of unverified origin.