Golborne Vintage Radio

I think I know most things that I need to know about mains power isolation transformers and their usage. But there is one issue that has me a bit puzzled. There is a strong line of thought which states that if you are using one of these, you must only ever use it to supply ONE load; using it to feed several sockets / loads is an absolute no-no. The short question I am asking is . . . . WHY?

O.K., so before you jump in with a reply, let me expand on things a little.  I can readily appreciate that if two A.C. / D.C. vintage radios were to be so plugged into that supply simultaneously, there is the possibility that the mains input 'L' of one set could become connected to the mains input 'N' of the other one, thus placing a dead short-circuit across the isolation transformer secondary. But anyone who is going to seriously attempt any repair work on such equipment will surely check the A.C. power connections inside the set first,  i.e. is the incoming 'N' (and not the 'L') joined to chassis?  And if found to be so wired - 'L' to chassis - to re-wire it accordingly?
Once that has been done, where is the safety problem? 
(Assuming that if many sockets are in circuit with loads connected, the total V-A demand does not exceed the rated V-A capacity of the transformer.)

So, am I missing something here - something obvious - which, for me, would not be at all unusual?

Al. / Skywave.

An isolation transformer is purely to allow the connection of test equipment to a device under test.

As soon as the earth clip of a 'scope probe (for example) has made contact with a point in the circuit, everything else in the DUT has a potential that is relative to mains earth.

If there are multiple devices plugged into the isolation transformer, then those devices are no longer floating with respect to mains earth. Every part of them will either be at earth potential or some other potential with respect to mains earth. That means that you are at risk of an electric shock - perhaps just from touching the case of another item of gear if you were especially unlucky. It also means that you probably can't connect some other item of test gear that has an earth reference (e.g. a sig gen) while your 'scope is still connected to the first item.

Further to this, most people IME are not aware that the mains earth must not be carried through to the (singular!) BS1363 outlet of your isolation transformer. There have been instances where electric shocks have been received by simply touching the metal case of a DUT, owing to a N-E swap in the mains plug. A chunky 4mm binding post connected to mains earth should be provided, so the operator can firmly earth any part of the DUT as required.

I've attached a BBC guidance note about isolation transformers, which hopefully will be interesting reading to everyone here. Note the mention of RCDs. These generally offer greater safety on a workbench, and should be the default way to supply gear when "working live".

O.K., thanks for your reply. 

All I will add is that for ages I have arranged my test / repair bench to be fed via an isolation transformer unit which feeds multiple 3-pin sockets.  That transformer is internally configured for 120v. - 0v. - 120v. bi-phase output with the 'E' / 0v. centre point connected to supply earth inside the commercially-built isolation transformer unit which has full Government safety approval - plus full ELCB protection and fuses fitted. The idea of the bi-phase arrangement is simply to reduce the magnitude of any 'L' - 'N' electrical shock that I might receive.

I haven't had any mishap yet.

Al.

Al, your arrangment guarantees that the chassis of any "live chassis" set will be at 120V wrt earth. Connecting any earthed testgear to such kit will cause a direct short.

Just plain dangerous.

As Mark says, the best geenral protection for a workbench is an RCD. Use an isolating TX only when you need to conenct earthed testgear to live chassis kit.

I have a dual socket on the output of my iolsating TX but would never use it to feed 2 separate bits of kit. It's currently in use with one socket feeding the heater chain of a live chassis TV. The other is feeding a variac which is supplying HT for the same TV. I've made sure that the polarities are all correct as inadvertantly getting the neutrals to the variac and TV wrong would be unpleasant.

(11-02-2015, 07:36 AM)ppppenguin Wrote: [ -> ]Al, your arrangment guarantees that the chassis of any "live chassis" set will be at 120V wrt earth. Connecting any earthed test gear to such kit will cause a direct short.

Just plain dangerous.

With that arrangement, yes, it is, agreed. But I am aware of that danger (no-one else uses that test bench) and on the very rare occasions when I do need to power an item which has a 'live chassis', I always use a separate, suitably-rated, isolation transformer. I should also add that upon review, I have discovered that I have not connected the incoming supply 'earth' to the 'earth' pin on the supply sockets. The transformer secondary centre-tap 0v. point is connected to those 'E' pins only. That gives me my '0v. reference'. So, "why use that isolating transformer in the first place?", you may well ask. Because with my arrangement, I have a maximum of 120-v. between earth and the supply rails: half of 240-v. Any electrical shock I receive will be less - so a safer work-space. Plus the supply is a balanced arrangement, with the advantages which that brings. That isol. transformer has a RCD fitted - as I mentioned in my earlier post.

Al.

The RCD is useless in your setup. If it's on the primary of the TX, it won't "know" about any earth leakage faults on the secondary. If it's on the secondary, unless somewhere on the secondary winding is earthed before the RCD there cannot be any imbalance for it to detect.

You may regard your own safety as optional but let's make it abundantly clear that yours is not an example to follow. For those without specialist knowledge the guidance given by Mark and in the BBC document are the safest methods. RCD on bench, isolating TX for working with a single specific item when you want to connect earthed testgear.

There is no panaea, there always ways to hurt or kill yourself. The first and most important item is to be aware of possible dangers and work accordingly. RCDs and isol TXs are a secondary level of protection. For example consider the HT supply in a radio with a proper mains transformer and earthed chassis. This will be at 200V to 300V above earth and it will give you a nsaty belt if you touch it. Neither RCD nor isol TX nor fuse will give you any protection in this situation. Counter-intuitively the HT rail on a live chassis set can be safer if fed via an RCD as it will trip before you get a serious shock.

(11-02-2015, 09:46 AM)ppppenguin Wrote: [ -> ]Counter-intuitively the HT rail on a live chassis set can be safer if fed via an RCD as it will trip before you get a serious shock.

This is all good stuff, but its worth also noting that there are RCDs.... and then there are RCDs...

A few things - check your RCD regularly - I find that in the slightly humid conditions of my workshop cellar, an RCD typically only lasts about 3 years before it starts to become erratic and thus needs replacing. Use the test button. Regularly!

Further, if you have an RCD rated for 30mA you will still get a blast for up to 30mS, which for some of our more senior or frail members may be problematic, i.e. other current ratings are available...

Additionally, RCDs do not protect against over-current, i.e. phase-to-phase or phase-neutral faults (depending on usage of single or multiple phases), so ideally use an RCBO which is an RCD with over-current detection in a single package.

Also, RCDs are vulnerable to a supply-side neutral fault or disconnect that would leave the live side connected, thus they should ALWAYS be connected to the supply via a two-pole breaker - the live must be disconnected at the same time as the neutral.

Best idea in a workshop I've found is to use a consumer unit that has a two-pole breaker with an RCBO, and then MCBs for each spur. Thus EVERYTHING is protected, including those silly times when I run an extension cable to power my saw bench when it's moved into the garden because I'm machining a piece of 8x4 that won't fit in the shed !

FWIW, I use a medical isolation transformer, an ISB60 http://www.toroidtech.de/transformer_pro...Series.htm, which has multiple outlets...

EDIT: We should all be working to "IEE Wiring Regulations 17th Edition : (BS 7671: 2011)" anyway, and not trying to re-invent the wheel...

(11-02-2015, 09:46 AM)ppppenguin Wrote: [ -> ]The RCD is useless in your setup. If it's on the primary of the TX, it won't "know" about any earth leakage faults on the secondary. If it's on the secondary, unless somewhere on the secondary winding is earthed before the RCD there cannot be any imbalance for it to detect.

I disagree. The RCDs is on the secondary winding. Let me explain in more detail.
The feeds from the transformer's 'L' and the 'N' - which go to the RCD and thus to the O/P socket -  I shall call phase 1 and phase 2. Each phase is 120-v. RMS w.r.t. the 0v. centre-tap of the transformer, which is fed to the transformer's output 'E' pin and thus to the 'E' pin of every socket outlet. In the event of a leakage current, phase 1 or phase 2 to 'E' - (same as the test item's metal casing) - there will be a current imbalance and the RCD will trip. On test, this has been proven to happen.

-----------------------------------

You may regard your own safety as optional  . . .
Please omit the sarcasm: thank you.

. . . . but let's make it abundantly clear that yours is not an example to follow.
I'm not saying that it is. What I am doing, in effect,  is using an isolation transformer in a method which is not the conventional way of using such a device. But that method is not 'unsafe'.

For those without specialist knowledge, the guidance given by Mark and in the BBC document are the safest methods. RCD on bench, isolating TX for working with a single specific item when you want to connect earthed test gear.

Which I have just dealt with immediately above. And, over the many years, that traditional method is what I have always used when working with a mains-powered item which has one 'side' of the incoming mains connected to the metalwork which the user / tester may come into contact with.

There is no panacea: there are always ways to hurt or kill yourself.
I'm not saying that there is a panacea. Moreover, this topic is about using an isolation transformer, not general electrical safety.

Al.

What are doing is creating your own mains supply, completely floating wrt what comes into your house. You are using a double wound TX to do so. You are also separating your local bench "earth" from mains earth and any extraneous earth such as water pipes. Hence under fault conditions dangerous potentials can exist between your "earth" and any other earthed metalwork. If you can guarantee your workshop to be an earth free zone this is a recognised way of working under certain specialised conditions. Otherwise it's dangerous.

(11-02-2015, 10:15 AM)Skywave Wrote: [ -> ]I'm not saying that there is a panacea. Moreover, this topic is about using an isolation transformer, not general electrical safety.

Since what you are doing is not really using an isolating TX but creating your own local mains supply this raises various safety issues. Judging by previous threads in various forums there is widespread misunderstanding about how to create a safe working environment. Much of this misunderstanding concerns isolating TXs but (if you'll pardon the pun) it cannot really be considered in isolation from other aspects of safe working.

OK, what Al initially described in post #3 was, as he rightly stated, a device to limit the maximum potential w.r.t. mains earth to 120V, thus limiting the magnitude of an electric shock. Of course, similar devices are widely used in the construction trade, for example, where the transformers are 110V with a centre-tap that is connected to mains earth. Hence, maximum potential to mains earth is only 55V, and that brings about a useful increase in site safety - especially a site that might be exposed to the great British weather

However, such devices do not provide the facility to safely connect earthed test equipment to them.

If the centre-tap is earthed, then a downstream RCD can be connected, and does provide a degree of protection.

But, it's important to be clear that just because a transformer is at the heart of this setup, it does not provide isolated mains. For anyone who is considering replicating this, my personal view is that it doesn't actually "bring anything to the party", over and above regular mains outlets that are protected by an RCD in working condition. Apart from perhaps reducing the magnitude of a shock received during the 30ms that it will take the RCD to open... And I suppose that a large, heavy isolation transformer will offer some protection against spikes on the incoming mains.


But...

Al then went on to say (in post #5) that actually, the earth connection of the outlet sockets is not in fact connected to mains earth any more, but continues to be connected to the centre-tap of the isolating transformer. He goes on to say that the maximum potential between mains earth and his "live" and "neutral" terminals can only be 120V, but this is not guaranteed to always be the case:

With the output from the transformer now floating, a connection between the post-TX "neutral" and mains earth will result in his post-TX "live" becoming genuinely live with respect to mains earth. The same would apply to a properly implemented isolation transformer, of course. But here's the rub: the "earth" connections in his post-TX sockets are now at 120V with respect to mains earth, meaning you could get a shock from the metal case of a class 1 instrument!

If you want to work this way, you need to have a totally earth-free environment - nothing at all connected to main earth, and you need to be unable to connect yourself to mains earth (good rubber matting on the floor, etc). It's a complicated way to work, and doesn't bring any obvious advantage over the much simpler way I described earlier.

In other words: Al, please consider revising this setup. Anyone else: please do not replicate this setup

The subject of isolation transformers is a complex one, and it's one that seems to confuse many a fine engineer and technician. In my experience of discussing this subject on the various vintage forums, I find that many people seem to temporarily forget how electricity works, and instead fall back on "it's how we used to do it at {insert name of former employer, the bigger and more well-known the better}, and we never had a problem, mate!". Well OK, but I always try to encourage people to go back to first principles, and consider all of the scenarios, so at least they can fully understand the benefits and limitations of their chosen setup.

And yes, in any thread about the use of isolation transformers, the subject of electrical safety is very much on-topic. Fairly obviously, I would have thought

For what it's worth, I have 3 types of sockets on my bench:

1. The first are red in colour, and these have a double-pole isolation switch (with a red rocker). These are left on 24/7, and they are for the computers and anything else that needs to be continuously powered.
   
2. The second type are white, and these also have a double-pole isolation switch and an RCD. These power all test equipment and DUT (the latter via a 4 way mains distribution unit that has voltage and current monitoring built in. Very handy).
   
3. The final type are isolated. I have a pair of 500VA mains transformers mounted in a 3U rack case (details here), supplying single sockets with no mains earth connection. Obviously, no RCDs involved here.
   

I believe that this setup represents the safest way to work - the colour-coded sockets make it especially unambiguous, and the isolated outputs are clearly labelled (though not in any photos, as that was a recent addition). I rarely use the isolation transformers - they are there for when working on the "live" side of a switched-mode power supply or similar. I've included an image of the "front of bench" outlets for devices under test - these are just above eye-level, so it's really easy to reach up and hit the breakers should the worst happen!

Hope this helps to clarify a few things,

Mark