The IET is carrying out some important updates between 17-30 April and all of our websites will be view only. For more information, read this Announcement

This discussion is locked.
You cannot post a reply to this discussion. If you have a question start a new discussion

Isolation transformer

I’m trying to get my head round isolation transformers and if I’m overthinking Joy. I understand the secondary isn’t earthed on one leg therefore the a shock cannot be received if on leg of the transformer was touched when in contact with the ground. But if one leg was to become accidentally earthed we then have reference to earth. Now the supply is not isolated from earth- on the second fault to earth I can understand if the fault is of low impedance the fuse should blow but if not it can then sit at a potential.

I think I am overthinking this but what do the regs say? 

many thanks in advance for your help getting my head round this topic. 

  • With an isolated supply the both ends of the windings are "floating", that is they are not linked to earth or anything else, so they could be at any voltage. The winding might have an output of say 230V between ends.

    If you catch hold of one end, and there is a path to 0v through you, then you "tie" one end to 0V, Therefore the other end will have 230V on it. 

    Let go and touch the other end, again the other end goes to 230V, your end goes to 0v.

    Only when you touch both ends of the windings will you get a shock. 

    This obviously, greatly reduces the risk of shock, even more so if there is no path to earth within reach.

    The use of isolation transformers and earth free areas was a method of reducing shock risk before RCD's (and their predecessors) became so effective and low cost. I worked in such an area around 1980.

  • As rightly noted, the shock risk is only present once you are touching both ends, either directly or via a fault path.

    Also, it is possible to loosly ground the transformer - i.e. via a high impedance, and detect the voltage across that impedance, so that you can keep running with a single fault, but be alerted that there is a single fault problem, hopefully to be able to do something before it develops to double fault.

    An early version of this was use on the London Underground, where the rail voltages were something like a nominal  660VDC but earthed 1/3 of the way up, via a resistive divider of warning lamps so the negative rail would idle at something like -220V and the positive one at +440V, but if one rail was then forcibly grounded by a fault, then the current flow would be limited by the lamp resistance, and the lamp for one side would go out, and the other come on extra bright -  an alarm then, but power would remain between the rails so that the trains could be kept running to the next station. (more recent sections are upgrading to 750V but the split and earth fault detection remains. Note that the real voltage varies wildly - by several hundred volts actually - as the trains either draw current for traction  or put it back into the line while braking, but that is a complication beyond this discussion. More here)

    The modern insulation monitor on an IT system can be thought of as a descendant of the technique.

    There are a number of ways that isolation transformers may be used, and the regs permit most of them, but do require a bit more thought than a simple TN-x supply.

    Mike.

  • But if one leg was to become accidentally earthed we then have reference to earth. Now the supply is not isolated from earth- on the second fault to earth I can understand if the fault is of low impedance the fuse should blow but if not it can then sit at a potential.

    Quite true. However to get to that situation you need two distinct faults simultaneously - which puts you in a very similar risk area to normal ADS - what happens if the c.p.c. breaks and you then get a L-PE fault? yes the exposed-conductive-part just sits there at 230V waiting to kill anyone that touches it (presuming the circuit doesn't have additional protection, e.g. a ≤30mA RCD).

    Similarly the double insulation approach can be lethal if both layers of insulation fail.

    Ideally where you have more than one distinct exposed-conductive-part on a separated system you'd connect them all together with a protective conductor (unearthed) - so that if you did get two simultaneous faults to PE from different poles, you have something similar to a L-N short circuit which would then either take out the overcurrent protective device in pretty short order, or where the supply didn't have enough oomph, the voltage would likely collapse to something tolerable. The regs sort of require that - only when supplying a single item of current-using equipment can exposed-conductive-parts be left individually floating, where there are two or more items of current using equipment on the same separated system then the exposed-conductive-parts must be bonded together. They just seems to overlook the possibility of there being more distinct exposed-conductive-parts than items of current using equipment - e.g. where supply accessories or wiring systems are Class 1.

    As with most things in life, electrical systems are never 100% safe under all circumstances - it's more a matter of keeping things 'safe enough' (which is of course a bit of moving target as society's attitudes to danger change)..

       - Andy.