workbench safety: mains isolation

[tweek]

The last year or so has kept me away from the electronics hobby for various reasons.  In that time though I managed to get my hands on a house and I've been allocated by the lady of the house 100sq ft of my very own.  Workshop time!  Sorry Dave and Chris this is pure coincidence, and WOTW had nothing to do with it.     Anyhow I've racked up a queue of several projects I want to build and at least one of them (off-line SMPS for my bench) will require work with mains voltage, and potentially require me to poke at the thing with an oscilloscope.  While I intend to isolate the output, obviously some of the guts of the power supply aren't.

My understanding is that this is a problem for test gear where the ground clip is the actual supply ground, and isolation must be used.  Funny thing is I don't really see this discussed much.  Does anyone maintain isolated power at their workbench?  If so, can anyone comment on whether it's more prudent to isolate the DUT, or the test gear itself.  It seems that it would be easier and cheaper to isolate the test gear, but there is an additional potential safety benefit from instead isolating the DUT, on the off chance that you, for some reason, become not isolated if you know what I mean.   

Recommendations on isolation transformers is also welcome!

Any thoughts?

Thanks!
Phil

[Mr J]

probably a good idea added safety. Another idea...power kill switch for the bench. On mine I use a DPDT switch to cut both hot and neutral located on the right side of the bench with a big red facepate.   

[SG-1]

I use this for mains isolation, it allows one to dial in the trip current.  It was designed to test power tools & such for insulation failure.  It was used & abused when I found it, but still working.

[Psi]

An RCD is another option

[JohnS_AZ]

I isolate the DUT simply because I have a number of pieces of test gear that I would have to isolate, but I'm usually only working on one DUT at a time.

BTW ... I've always had a big kill switch on the front of my benches. A really good idea.

[SG-1]

When working with mains you should consider wearing safety glasses in case something vaporizes. 
Wear some kind of gloves when probing around. 
Modify your multi-meter probes something like this to reduce the likehood of causing a short circuit or ground fault.

[alm]

Connect the DUT to an isolation transformer, not the test equipment. You want IEC class I equipment (anything with a grounded plug) to be grounded. The equipment was designed with the assumption that ground is safe to touch under all circumstances, a floating ground may be connected to dangerous voltages. Say you accidentally clip the ground clip of your scope to live (or something that gets connected to live due to a fault), and then touch neutral with your other hand. An isolation will not protect against this, but a grounded scope would. Not grounding test equipment will also compromise signal integrity. If you want test equipment to float, buy test equipment properly designed for this, like a Tek TPS2000 series scope or a Fluke Scopemeters. All DMM's I know are already floating, even bench DMMs with grounded plugs, so no point hooking these up to an isolation transformer.

[Kiriakos-GR]

Recommendations on isolation transformers is also welcome!

Any thoughts?

No thoughts just an question, how many watts you need ?

The copper price are flying sky high , and the transformers price follows that route too.
And so I suggest you to build just an " partial "  isolation by transformer ,
like an special wall plug , so to use it with some devices that can cause problems.

Like old radios with tubes , or old televisions with tubes.

In this case , you will need something small and affordable.  ( lets say 500W Max, so to operate for long time , and remain cold too)

[RayJones]

I'm personally in favour of a portable isolation transformer that is specifically used for a specific task at hand.

i.e. ie requires the conscious effort to utilise the transformer when you expect it to be required.

If you are always isolated, this may lead to complacency should you move elsewhere, and being portable you can easily move it elsewhere if the job cannot be moved to the bench...

Nevertheless - RCD - YES YES YES.
In Australia it is now mandatory that all houses (with recent wiring work performed) must have an RCD fitted.
You can always use another one, local to you bench, just to be sure.

Besides all that 1000 square feet - NIRVARNA - my shed at home is only 600 sq ft and that usually houses two cars :-(

[Psi]

For mains power on my desk i built a control box to distribute power to all my gear.

With a master on/off switch and neon i can turn off all my gear with one switch and can't miss seeing the bright master neon at night, so it's pretty hard to leave anything on that might burn the house down.
(which is good because i seem to attract nearly-burning-the-house-down events   Electric blanket fire a while ago and that garden slug shorting out the wall socket that i posted on here a week or so back)

The master switch powers 4 outputs, each with their own switch and neon.

Output A is a cable hardwired to a 4-way powerboard

Output B, C and D are 3 pin mains sockets on the back of the box. So i can have testgear/equipment plugged directly in at the back with their own switch or i can plug in extra power boards and have multiple devices on each output.

Its quite useful, i have the built-in power swiches set to ON for all my gear. That way i can just turn on the master switch and whatever output switches i need.
A for Lighting,  B for Soldering/rework, C for Test gear and D for the Variac and Isolating transformer (which i use if i need 230v for projects).

Also included in the control box is a fuse (since i dont need full 10A at the desk) and an earth terminal for my anti static strap.
I plan to put a RCD on the input to the control box at some point as well, currently it's just mains from a wall socket.

[tyblu]

Fairly long answer here: http://electronics.stackexchange.com/q/8534/2118

The short version: isolate DUT, monitor floating chassis voltage relative to Earth ground, and don't touch anything while DUT is powered and/or you are taking measurements. (Turn off power, clip on probe, set trigger, turn on power, step back, turn off power, look at chassis voltage, equalize chassis if needed (if capacitive), then read test gear/'scope.)

[jahonen]

Yes, I have one on my bench. It is actually an isolated variac with output power of 630VA 0-250 VAC nominally. I used to repair all kinds of equipment in my previous life, and for measuring the non-isolated primary side with an oscilloscope safely, the DUT isolation is a must. Having just a RCD does not allow that, although it naturally improves safety otherwise. I still use it whenever working with low-power mains powered equipment, even if I don't necessarily need the isolation. Note that there must be only one DUT connected to the isolation transformer output, or the situation becomes hazardous again.

I believe the isolation transformer reduces CAT-requirements due to that it increases the resistance so short circuit current is reduced compared to direct mains connection. Of course that depends how stiff transformer you'll have.

Isolating just the measuring equipment is dangerous, but if you can afford it, both can be isolated from mains (DUT and mains powered measuring equipment). Also do not use same isolation transformer for both DUT and your measuring equipment. I have isolation on the measuring equipment too, haven't seen any odd behaviour due that.

Regards,
Janne

[mikeselectricstuff]

I'm personally in favour of a portable isolation transformer that is specifically used for a specific task at hand.

i.e. ie requires the conscious effort to utilise the transformer when you expect it to be required.

Definitely agree with that one. Unless all you do is SMPS design, most of the time you won't need an isolated supply, so bring one out when you need it.
A variac (with V and I metering) is also a very useful addition to an iso transformer for debugging mains stuff - you often want to start things slowly to watch for smoke, and at some point will need to test your hardware over a range of mains voltages.

One word of caution on isolating transformers, especially smaller ones. It's easy to assume that the output voltage is the same as the input, however this will usually only be the case near full load, and  at light loads, the output voltage can be higher - a long while ago I was debugging a mains fan controller and was puzzled for a while as to why it wouldn't work properly on the isolating transformer - turned out it was putting out about 285V at no-load.

[Psi]

yeah, it can also be an issue with variacs

Some countries have 240V mains with houses close to the step down transformer getting near 250V.
When building and testing mains gear you normally want to test between 210v and 250V and want a variac that can deliver that. So a lot of variacs are designed to be able to produce 250v from 220v.  
Setting your variac to max if you already have 250v gives obvious results.

[Zero999]

An isolation transformer is a good idea but you don't need a large one, 250VA is fine for most applications and can be used on larger loads for short periods.

One thing to bear in mind is that an isolation transformer will defeat an RCD so there's no point in having both although all test equipment should be powered of an RCD.

Make sure the device being tested is protected with a suitable fuse. Here in the UK it's not a problem as all plugs are fused but in other countries it's not the case.

[JohnS_AZ]

RCD?

[tyblu]

GFI in Aussie-talk (residual current device).

[Simon]

I always use a portable RCD, that way if you touch a live by mistake your saved but make sure it is a trip at 30mA or less sort. if you make sure you clip your scope ground to neutral you should be ok but an isolation transformer is a good idea.

[JohnS_AZ]

If you make sure you clip your scope ground to neutral you should be ok ...

Is that really what you meant to say?

Ground is not the same thing as neutral. If, by some freak failure, you loose your neutral connection and you have your scope ground tied to neutral you'll have all of the AC current coursing through ground. Not a good thing.

[tyblu]

Neutral isn't 0V, either, due to unequal split-phase loading. Perhaps you mean isolated neutral?

[tweek]

Wow what a great response to my first thread!  Y'all sure know how to make a guy feel welcome.  Definitely a few angles in here that I had not thought of before.

Recommendations on isolation transformers is also welcome!

Any thoughts?

No thoughts just an question, how many watts you need ?

I figure 400-500VA is probably more than enough for any single device I would want to test.  I was definitely looking for a portable unit as some have suggested, which I would only use when I need it (which honestly will likely not be very often).  I'm thinking about a unit similar to this:

http://www.tripplite.com/en/products/model.cfm?txtModelID=227

The MSRP is kind of high but I've seen them for around US $200.

I'm also considering the 250VA model which is considerably cheaper.

Nevertheless - RCD - YES YES YES.
In Australia it is now mandatory that all houses (with recent wiring work performed) must have an RCD fitted.
You can always use another one, local to you bench, just to be sure.

Besides all that 1000 square feet - NIRVARNA - my shed at home is only 600 sq ft and that usually houses two cars :-(

Here in the US GFI/RCDs are still only required in specific areas and they generally pop out at 5mA.  AFCI breakers are now pretty universally required on branch circuits to all areas of the home where people might be regularly and I think some of these units also provide ground fault protection.  My house was built a couple years before that was universal but I'm considering up-fitting at least some of the circuits here with those devices.

Wow 1000sq ft awesome!  That is way more generous than the 100sq ft my wife actually let me have!

BTW ... I've always had a big kill switch on the front of my benches. A really good idea.

Do you use a large rocker switch as Psi showed in his picture or do you have a "big red button" type arrangement with some sort of shunt trip or a relay that kicks out when you slam it?  I used to have the latter in the workshop I had as a kid.  Big red buttons rock!  Bonus if it flashes.

If you make sure you clip your scope ground to neutral you should be ok ...

Is that really what you meant to say?

Ground is not the same thing as neutral. If, by some freak failure, you loose your neutral connection and you have your scope ground tied to neutral you'll have all of the AC current coursing through ground. Not a good thing.

Also even in the best case where your neutral is intact, if you were using a RCD I think you would likely cause it to trip .  The only time it's really OK to consider ground = neutral is generally at the service entrance where all your bonding points are tied together.

Reminds me of a story though...this past Thanksgiving I was at my in-laws and we got on the topic of electric work in old farm houses.  He told me a story about how one time he lost the neutral up at the barn so he just hooked it to a copper pipe that was going back to the house.  Problem solved!  Still picking up the rubber bands that popped in my head that day.   There was a book I was reading a while back (can't remember the name right now) that talked about how the problem with most safety systems and procedures is that probability-wise there is generally little to no penalty for defeating them.

Thanks,
Phil

[Zero999]

How about getting a transformer with two 120V secondaries so you can test 240V as well as 120V devices?

[Neganur]

Safety regulations here demand both DUT and measuring equipment connected to isolation transformers, if you go anywhere near mains power.

It's always a good idea to have your scope connected to one anyways. You end up creating ground loops more often than you think if your signal generator and scope use the same mains for example.
I use a fairly inexpensive 140VA one for my scope. And an old german variable 1000VA, 0-250V for the DUT.
Variable is a must btw, when you intend to repair old gear - slowly increase the voltage to not upset those capacitors

Handheld multimeters are always safe in this regard.

[Simon]

I was told once in Italy that if you got a shock off a tap it was because someone was using the earth as neutral to evade logging their consumption on the meter, To be honest I don't know why we don't earth the neutral in buildings ? that way we know that neutral is neutral

[JohnS_AZ]

While a little off topic, here is the electrical diagram of my new workbench. Some will certainly say that it is over the top, but I already own everything in the diagram, so why not use it?

My isolation transformer is a stand alone, so it's not included in the drawing. I always use it for HV DUT, but I never have for my test equipment. If one side is isolated, never saw the necessity to isolate the other.