RCD or Isolation Transformer?? What is best to use for safety

[jastreb]

HI everyone,

I hope you can help me to understand where I should use each one. So I am a hobbyist and I would like to start playing with A?V receivers and TVs etc and pull apart and try and fix myself. I currently have a plasma TV and a denon amp I would like to start with.

So these being potentially lethal, I have purchased one of those portable RCD devices such as this
http://www.gasweld.com.au/toolex-safety-box-port-10amp-clipsal-4-outlet-583015?fee=23&fep=161616&utm_source=Toolex+Shopping+Feed&gclid=Cj0KCQjwi97NBRD1ARIsAPXVWWAvXagJ4KB__35Asggrsj4r7VPM2BNaG2dk-nlI7wahAUaibptlG5AaAoqUEALw_wcB

But I have been reading some more and what seems to be suggested is an isolation transformer. So what is better from a safety perspective. My understanding is that an RCD will not work in serial with the isolation transformer.

Thanks for advance and please excuse my question if it seems obvious.

[krystian]

If you use isolation transformer it can still shock you and have enough current capability to kill you. The difference is, you have to touch two points in the circuit you are working on and they have to have different potentials. And if you do this, RCD won't help because voltage is not referenced to the ground. In my lab I would use RCDs with some small engaging current (3mA). But it is still only additional protection which may not help.

I wouldn't recommend working on devices connected to line voltage without training.

Working with line voltage is never safe.

[Brumby]

Anything that has mains power running into it can easily be lethal, so it is wise to be cautious.  But, don't forget that even some battery operated equipment can have dangers as well.

Having said that, the first step in safety is to know what you're sticking your nose into.  Have someone show you or supervise you, if at all possible.  Even if you don't have a circuit diagram, at least have a rough idea of how the power supply side of things works.  If there is a nice heavy power transformer involved with separate input and output windings, then that offers some protection.  If it's a SMPS, then get to know what to look for and what to keep away from.  With live chassis - be VERY careful.  In fact, I would recommend steering clear of these until you were much better versed.

I would always recommend an RCD on any mains coming into the workshop - whether it's a main one in the meter box or a separate one like you have purchased.  A separate one that is a little more sensitive than the one in the meter box would be preferable - so that any "oops" in the workshop trips your local one, before the household one (hopefully).

An isolation transformer always sounds like a good idea - but you have to be particularly careful in how you use it.  If you are not careful, you can actually end up with situations where the risk is greater.

Using an RCD with an isolation transformer is not necessarily going to give you a safer environment.  Putting the RCD before the isolation transformer is not going to give you any protection from the secondary, since there will be no fault path that it will be able to detect.  Putting the RCD between the isolation and the device under test is not going to work, unless you establish a fault path avenue that will allow the RCD to function.  The simplest way is to earth reference the secondary - which immediately removes the benefit from having an isolation transformer.

If you have an isolation transformer, then it is probably safer to use it as a door stop ... and if you have a situation where you think it might be useful ... present it here before you try, so that we can give you some feedback.

The one-hand rule is also a good one to learn.... When poking around mains and high voltage circuits, keep one hand in your pocket.  While this may sound silly - it is not - and is based on a very significant risk if it is not followed:  If you have two hands on a device with dangerous voltages around and one touches the live point and the other touches ground, then the fault current will pass through one arm, across the chest and down the other arm.  It's that across the chest bit that is the real worry ... because your heart is right in the way......

Care is essential, but knowledge is important in staying safe.

[Assafl]

David Aurora is a mushroom (fungi / mycology) expert who wrote one (if not the) best book on North American Mushroom and mushroom foraging. In addressing safety he set down rules (and I paraphrase from memory) that one should never forage until one is able to identify the lethal mushrooms (mainly Amanita Phalloides - the "death cap" and the "destroying angels" like Amanita Virosa). It is far more important to NOT PICK a bad Amanita than overlook a tasty one.

Same is here. It is far more important to know what constitutes danger and what doesn't. And it isn't always a matter of voltage or current.

A few examples:
1. Main powered. Always dangerous. Understanding grounding and understanding if and where the mains ends is critical. In many power supply circuits the PCB is clearly marked. But in some it isn't. Personally I don't trust the markings. If I am not sure I measure. As an example, the mains side of a switch mode power supply may have a small 16V electrolytic. Isn't 16V safe? Nope. If it is used to stabilize voltage to the SMPS control chip it is 16V but it may float on 220V mains.

Similarly, you may read that Magnetrons use 3V heaters. But the 3V is at a -2000VDC from ground since they use the heater pins as the cathode. (That is why the two heater pins use a big white insulator - it is a dead giveaway that something ain't kosher here).

2. Capacitors. They can pack a punch. Obviously tiny pF ones meh. But the big electrolytic ones? They may have enough to make a high gauge wire disappear like a fuse. With a pop and exploding gasses. And that assumes the voltage on them is low. In a very high power Amplifier, the difference between the positive and negative rails may exceed 100V. That is dangerous. A particularly vile one is the ugly cap in a Microwave oven 2000v (or more) at 1.8uF. Calculate the coulombs and joules and compare to a defibrillator. Not quite there but not too far. Just one of the reasons microwave ovens are one of the most dangerous devices out there. Also the small value high voltage caps in SMPS. Nasty little things (if charged).

Capacitors can also explode. I've had an exceptionally high quality electrolytic from Cornell Dubilier ooze dreck for days on my desk (it would bubble and foam as if angry). Tantalums capacitors can blast their heads off (especially if reverse biased) - and stink. If it hits you in the eye.... So how much voltage does your Ohmmeter give out when you reverse bias the Tantalum (in reverse - by mistake - of course)?

3. Voltage doublers, sparking circuits, fluorescent supplies, etc. - usually will give a nasty kick. But if they get a capacitance charged - it may be more than nasty. Learn how to identify flybacks (rare today), doubler cap-diode chains, small isolated transformers, HV nipples, weird flexible HV cables, etc. 

4. Mixed isolation circuits. The nastiest I probably ever worked on was repairing my Brother In Laws treadmill. It did have an isolation transformer, but it also rectified the 220V AC into DC and used a capacitor to smooth the motor/generator voltage. It was very unclear what was mains and what was isolated on the PCB. I used an isolated channel oscilloscope (a Fluke 196 I keep handy for these types of circuits) and a handheld DMM to measure. It ended up being a faulty emergency break SCR that shorted. Washing machines, dryers, cheap espresso machines, etc all fall into the category of "partially isolated".

5. SMPS, VFD and the like. Just don't until you have sufficient experience. The reason is that they are nasty to probe and can pack high energy. Not just shock hazard, but also venting capacitors and other flying crap. In any case I like to remove, figure out the controller, the topology, where the isolation line is - and only then probe. I do my homework: I rarely have to probe more than 2 points to verify what the problem is - and that homework leads to safety.

Learn to identify the components, circuit topology and the safe way to probe - and only then start and work very slowly and thoughtfully. Mindfulness is key. 

Oh - and RCD and Isolation and Variacs - Nice tools. But if they ever trip or work - you've failed. I don't mind having one (and I do) - but they never tripped. If they do - thank them for your life and reconsider if this is a hobby for you.   

[technogeeky]

It's worth noting that an isolation transformer is really there to protect equipment, not the user. The goal is to make sure that you can place your oscilloscope ground clip, for example, on any node inside a DUT (Device Under Test) without worrying that you will create a short through the ground clip.

[Ian.M]

It's worth noting that an isolation transformer is really there to protect equipment, not the user. The goal is to make sure that you can place your oscilloscope ground clip, for example, on any node inside a DUT (Device Under Test) without worrying that you will create a short through the ground clip.

Yes. As soon as you clip on that scope ground clip, (or any other grounded test equipment lead), or even connect an aerial lead (as many aerial systems are grounded), that high voltage high stored energy circuit (D.U.T. - Device Under Test) on your bench that you thought was isolated has just become *FAR* more likely to kill you.  The odds of it killing you before your grounded a node were pretty small, as you'd have to be fairly dumb and unlucky to get a high energy shock across your chest by touching it with both hands. 

If you want to retain the increased safety of a fully floating D.U.T. when using an isolation transformer, *ALL* your other test and measurement connections must be isolated or fully floating and double insulated so it doesn't introduce a ground to the D.U.T.   If you mistakenly use another isolating transformer to float your normally grounded test equipment you've just increased the area and scale of risk considerably, as in that scenario, the grub screw in a knob you are adjusting on the test equipment can now kill you - one slip with the hand you are probing with while the other hand is on the knob will do it.

Therefore the isolation must be between the D.U.T. and the test equipment - an isolated differential probe for the scope, and a signal isolating transformer on any signal source outputs (or if its a HF source, high voltage coupling capacitors in both signal and shield wires), USB isolators on PC connected debuggers, or logic analyser pods etc.

Its also worth noting that its extremely unwise to just clip the probe ground to *ANY* node.  If you clip it to a fast switching high voltage node (e.g. the collector or drain of a chopper transistor in a SMPSU, a CRT line output stage or similar), there is likely to be enough interwinding capacitance in the isolating transformer and capacitance to ground  to cause problems ranging from damage to or destruction of the D.U.T. up to damage to test equipment and possible failure of the isolating transformer due to insulation breakdown.

[rstofer]

It's worth noting that an isolation transformer is really there to protect equipment, not the user. The goal is to make sure that you can place your oscilloscope ground clip, for example, on any node inside a DUT (Device Under Test) without worrying that you will create a short through the ground clip.

But not the second one!  They both need to connect to the same point.  Use the scope's A-B function instead!

Using an isolation transformer is such a bad idea that it should be banned from the planet and never discussed in forums!

An RCD ordinarily trips at 30 mA and this is a LOT.  In the US, the GFCI trips at 15 mA and this can still do some damage.  It all depends on the path.  If it goes through the heart, you are probably screwed either way.  That's why the "One Hand In Pocket" rule.  It prevents current flowing through the heart.

Really, without a formal training course in electrical safety, beginners should stick to < 30V.  I'm talking about a FORMAL COURSE, not a bunch of random musings on the Internet.  Find a qualified instructor!

[technogeeky]

It's worth noting that an isolation transformer is really there to protect equipment, not the user. The goal is to make sure that you can place your oscilloscope ground clip, for example, on any node inside a DUT (Device Under Test) without worrying that you will create a short through the ground clip.

But not the second one!  They both need to connect to the same point.  Use the scope's A-B function instead!

Using an isolation transformer is such a bad idea that it should be banned from the planet and never discussed in forums!

An RCD ordinarily trips at 30 mA and this is a LOT.  In the US, the GFCI trips at 15 mA and this can still do some damage.  It all depends on the path.  If it goes through the heart, you are probably screwed either way.  That's why the "One Hand In Pocket" rule.  It prevents current flowing through the heart.

Really, without a formal training course in electrical safety, beginners should stick to < 30V.  I'm talking about a FORMAL COURSE, not a bunch of random musings on the Internet.  Find a qualified instructor!

I specifically mentioned this in the other thread.

I suppose the real rule is: there is no case when you need to connect more than one ground clip except for shielding purposes, and in that case, all ground clips must be connected to the same potential.

I assume it's possible to get small ground loops if you are probing 4 signals that are far apart on a PCB, but only one probe is grounded.

edit:

One case where the MATH A-B function isn't very good (or, put another way, where an analog oscilloscope is better than some digital scopes) for this is on the e.g. Rigol 1054z. Unfortunately every time I try this technique with mains AC I get a very choppy pattern which is very laggy. Perhaps I'm doing it wrong.

[Zero999]

Both isolation transformers and RCDs have their place. I wouldn't agree with the statement about banning isolation transformers. The main advantages they have over an RCD, is that one has to touch the circuit in two places to get a shock and they're passive components, so there's less to fail. Yes it is true that connecting a piece of test equipment's earth to the secondary of an isolation transformer, defeats the safety benefits.

One thing that can be done with isolation transformers, is to add a circuit to sense when the secondary is earthed and light a neon, that way the user is aware that isolation transformer is no longer protecting them. The circuit can be as simple as connecting one end of a neon lamp to both sides of the secondary, via some high voltage (>4kV impulse rated) 470k resistors and the other side to earth.

[Brumby]

My main concern about an isolation transformer is that it can give you a false sense of safety.

If you use one, you will be able to make one external connection without damaging anything or getting any indication that you have just made one point live that you would normally expect to be innocuous - such as the chassis.  Then, when you move to another point in the circuit and forget what potential the chassis is at, you find out the hard way.

An RCD can't protect you, because there isn't an independent fault path and if the chassis was properly earthed in the first place, you would likely have found your attempted connection blowing a fuse or some other dramatic indication.


I've been playing around with electronics since I was 8.  Didn't get into mains until 10 years later ... and I still don't like isolation transformers because of the extra care required.  I cannot endorse them for a beginner in any way, shape or form.

[vk6zgo]

If messing with the Mains side of equipment power supplies is the most dangerous thing you do, you will live a long uneventful life.
Every time you become a road user of any kind, you are subject to a much greater accident risk.
Even something like cleaning your gutters is more dangerous, to say nothing of using any kind of power tool.

Tube/valve equipment , certainly, in Oz, normally had a power supply consisting of a transformer, the secondary of which, provided heater voltage, &, most importantly, a high ac voltage which was then rectified to give the tube's required dc HT voltage.

These secondaries were normally used with a  two diode full wave rectifier, requiring a centre tapped secondary which provided, commonly, "285 volts a side", or for older equipment, "385 volts a side" for radios, & around "350 volts a side" for earlier TVs.

Obviously, the complete secondary would be twice that, giving  570volts, 770volts, & 700volts respectively for the above examples.
And they weren't little "pipsqueak" transformers, either, ----many were around the size of many isolation transformers.
There were also quite unpleasant dc voltages available.

People worked on this stuff with perfect peace of mind, as they knew " where not to stick their fingers".
Beginners commonly made Mains operated stuff without mishap, as they learnt "first up" that they were working with equipment which contained nasty voltages.

The traditional "ac/dc" design, such as the "All American Five", were hated, & dubbed "death radios" by Australian EEs & Techs, as they really do have the chassis connected to one side of the Mains .
Active or Neutral,? How lucky do you feel?

Modern equipment with SMPS is a totally different thing.
The "guts" that performs the desired function is isolated from the Mains, just as with a transformer/rectifier linear supply.
Unless there is a SMPS fault  you should not be involved with any circuitry which has its "common" rail connected to the Mains., so can use your 'scope normally.

I used an isolation transformer extensively when repairing the power supplies of Television Picture Monitors & Receivers, but if there was no power supply fault they were unnecessary.

If you are going to get yourself so tangled up that you somehow zap yourself using an isolation transformer, I suggest you need to take a break.
,

[Zero999]

My main concern about an isolation transformer is that it can give you a false sense of safety.

The same can be said about an RCD, which could also fail to activate, especially if you don't test it every time.

If you use one, you will be able to make one external connection without damaging anything or getting any indication that you have just made one point live that you would normally expect to be innocuous - such as the chassis.  Then, when you move to another point in the circuit and forget what potential the chassis is at, you find out the hard way.

An RCD can't protect you, because there isn't an independent fault path and if the chassis was properly earthed in the first place, you would likely have found your attempted connection blowing a fuse or some other dramatic indication.

As I mentioned previously, some kind of indicator to warn the user that one side of the secondary is earthed would help a lot.

I can see the potential dangers. The key is understanding the pros and cons of isolation transformers.

By the way, isolation transformers are mandated in the UK for electric razor sockets, where they provide effective protection against electrical shock.  I believe they were chosen because, they don't have contacts which can weld together like RCDs do, which makes them more reliable.

[Vtile]

RCDs doesn't give any sort of protection in certain cases. Their working principle is to measure the current difference between "hot" and "neutral" leads (I'm not sure if there is models that also have separate protective earth sense, but it doesn't make the difference here ), so when you get shock with chest on neutral and hand in "hot" the RCD is happy considering that the appliance is working as intented (in case of electric chair it is true) and not do anything. They "only" sense a leakage to the ground wire (IF ANY!!!).

RCDs are truely a good ones in normal use, but in repair style work their protective effect is questionable.

The request to ban "all isolation transformers" is a bit hastily written and silly, if someone do read what in example CENELEC HD60364-7-7704 or equivalent IEC parallel publication  (60364 section 706) says about them.

PS. There is no magical blue pill to give you safety. The safety in this kind of work is a system consisting of many layers of protection, including the knowledge and proper training, mission-critical protection devices and gear, environmental considerations ... etc.

[oldway]

As for security, there is a lot to say ... I unfortunately do not have time to develop the subject now.

But the most important thing is that you have to know the dangers of electricity, understand the schematics to know what is dangerous and what is not and above all .... be very careful, never be distracted, not responding to his GSM while we repair or other thoughts or activity.

The advantage of using an isolation transformer is that it is possible to choose the circuit that is earthed ... for example the negative rail on the primary side of a switched-mode power supply.

If an isolating transformer is used, the circuit must be earthed. The advantage is that you can choose what is earthed.

[Brumby]

Please understand - my responses are made with one very significant factor in mind....

... this question is posted in the beginners section - by someone who is clearly a beginner.

Using an isolation transformer adds an element of "abstraction", if you will, where safe and proper use requires an understanding and an advanced awareness that you can not expect a beginner to have.

A live wire and an earthed chassis is a much clearer scenario.

[madires]

Isolation transformer with built in variac, voltmeter and ammeter plus differential probes for your scope. A RCD is great for your work bench but not for repairing mains powered stuff. Relying on a single layer of safety is a bad idea.

[oldway]

Please understand - my responses are made with one very significant factor in mind....

... this question is posted in the beginners section - by someone who is clearly a beginner.

Using an isolation transformer adds an element of "abstraction", if you will, where safe and proper use requires an understanding and an advanced awareness that you can not expect a beginner to have.

A live wire and an earthed chassis is a much clearer scenario.

If he is really a beginner who is not able to deal with an element of "abstraction", there is only ONE ANSWER POSSIBLE....

You are not qualified to work with mains and high voltage related devices.....start with low voltage circuits (<50V), wait to have enough knowledge and experience before trying to deal with electricity and power electronics.

[Brumby]

I would certainly endorse that.

Going into mains gear without a fundamental understanding and some experience is a bad idea.

Adding an isolation transformer into the mix is complicating an already dangerous scenario.

[Vtile]

For me the red-flag is always the black and white trashing of one method, while glorifying another. Indeed something like differential probe is easy way to get one piece sorted on the safety puzzle, but it only solves the ground referenced measuring instrument exposed conducting body issue (ie. in most scopes). It doesn't solve any other hazards.

The trashing will eventually lead to situation where the parroting leads to self-regulated obsolescence of valid solution for certain safety issues.

Oldways answer is good.

If he is really a beginner who is not able to deal with an element of "abstraction", there is only ONE ANSWER POSSIBLE....

You are not qualified to work with mains and high voltage related devices.....start with low voltage circuits (<50V), wait to have enough knowledge and experience before trying to deal with electricity and power electronics.

[rstofer]

RCDs doesn't give any sort of protection in certain cases. Their working principle is to measure the current difference between "hot" and "neutral" leads (I'm not sure if there is models that also have separate protective earth sense, but it doesn't make the difference here ), so when you get shock with chest on neutral and hand in "hot" the RCD is happy considering that the appliance is working as intented (in case of electric chair it is true) and not do anything. They "only" sense a leakage to the ground wire (IF ANY!!!).

I don't know anything about RCDs but I do know a little about the US style GFCI and this device measures the difference in current going out the hot wire and returning on the neutral wire.  The ground wire is irrelevant to the operation.

The request to ban "all isolation transformers" is a bit hastily written and silly, if someone do read what in example CENELEC HD60364-7-7704 or equivalent IEC parallel publication  (60364 section 706) says about them.

PS. There is no magical blue pill to give you safety. The safety in this kind of work is a system consisting of many layers of protection, including the knowledge and proper training, mission-critical protection devices and gear, environmental considerations ... etc.

Perhaps a bit hasty but it's not due to a lack of understanding the concept.  It's the idea that these are discussed by professionals in an area of the forum related to beginners.  It's true, we were all beginners once.  But I would be willing to bet that every single one of us has gotten 'bit' a time or two. Come on, fess up!

Surely you remember the hot chassis radios.  They were great until you lost a plastic knob and you tried to set the volume in your bare feet on a concrete floor.  This predates polarized plugs and receptacles. Ouch!  That stuff is probably still around and it probably still doesn't have a polarized plug to match the modern receptacle.

This was not the only time I got 'bit' but it was a learning experience.  Note to self:  Wear shoes!  And get a knob!

I would rather spend the time to come up with a plan to troubleshoot that didn't require isolation but I don't play in the part of the sand box where such a thing would come up.  Clearly, others do!

[Zero999]

Using an isolation transformer adds an element of "abstraction", if you will, where safe and proper use requires an understanding and an advanced awareness that you can not expect a beginner to have.

The same could be said about an RCD. If one doesn't fully understand the risks, they shouldn't be working with mains voltage in the first place.

If an isolating transformer is used, the circuit must be earthed. The advantage is that you can choose what is earthed.

That's an interesting statement. I thought one of the main advantages with an isolation transformer, is isolating everything from earth. One thing I know is that the IET doesn't recommend earthing anything connected to an isolation transformer, probably because it defeats the purpose of it being separated from earth.
http://electrical.theiet.org/wiring-matters/52/shock-protection/index.cfm

The greatest issue I can is, is if a non-isolated, high voltage (real HV, above 1kV or so) step up converter, is connected to the secondary of an isolation transformer, there's a risk of insulation failure, if one side isn't earthed. The answer of course is not to do this. If one delves further into the safety standards, they'll discover this is also prohibited. Fortunately someone working with isolated switched mode power supplies won't do this anyway. The main hazards are things such as old valve powered equipment, neon sign transformers and microwave ovens.

[oldway]

An isolation transformer without an earthing of the circuit that it supplies is intrinsically dangerous because if there is an insulation fault putting a part of the circuit to the earth, you are not not warned of it and
no protection can protect you, nor RCD (30mA differential switch) nor circuit breaker.

Furthermore, you do not know which part of the circuit is at a high potential in relation to the earth ... it is extremely dangerous.

On the other hand, if you put yourself part of the circuit you choose to earth, you know what is under high potential and what is not.

This allows safe measurements to be made on the primary circuits of the SMPS.

Of course, there are voltages up to 400V related to earth, but you know this and you measure it with high voltage X100 probes...

NEVER USE AN ISOLATION TRANSFORMER WITHOUT EARTHING THE SECONDARY CIRCUIT, THAT'S EXTREMELY DANGEROUS !!!!!

[Brumby]

RCDs doesn't give any sort of protection in certain cases. Their working principle is to measure the current difference between "hot" and "neutral" leads (I'm not sure if there is models that also have separate protective earth sense, but it doesn't make the difference here ), so when you get shock with chest on neutral and hand in "hot" the RCD is happy considering that the appliance is working as intented (in case of electric chair it is true) and not do anything. They "only" sense a leakage to the ground wire (IF ANY!!!).

I don't know anything about RCDs but I do know a little about the US style GFCI and this device measures the difference in current going out the hot wire and returning on the neutral wire.  The ground wire is irrelevant to the operation.

What Vtile says is true.  RCDs and GFCIs operate on the same principle and calling them the same thing is fine by me. 

You have taken the reference to the "ground wire" too literally - which is why I tend to refer to the "fault current path".  This can pass through the ground wire of a device, if you happen to bridge the live conductor and an earthed chassis - or not, if you happen to contact the live conductor and some earthed plumbing.

The point is, if the current passing through an RCD or GFCI is balanced, they will not trip.  If the current goes up your left arm and down your right - you could easily end up dead.

[Brumby]

Using an isolation transformer adds an element of "abstraction", if you will, where safe and proper use requires an understanding and an advanced awareness that you can not expect a beginner to have.

The same could be said about an RCD.

I think it more likely a beginner will be able to understand (if they don't already) the role, function and issues of an RCD which is a device commonly found in many homes than the use of an isolation transformer - which is not a common sight ... even in EE circles!

If one doesn't fully understand the risks, they shouldn't be working with mains voltage in the first place.

That goes for any environment - whether RCD, isolation transformer or straight off the grid.

[jastreb]

An RCD ordinarily trips at 30 mA and this is a LOT.  In the US, the GFCI trips at 15 mA and this can still do some damage.  It all depends on the path.  If it goes through the heart, you are probably screwed either way.  That's why the "One Hand In Pocket" rule.  It prevents current flowing through the heart.
y choppy pattern which is very laggy. Perhaps I'm doing it wrong.

I been reading up a lot about safety while working with mains. In regards to RCD my current understanding is that is is not just 30mA vs 15mA but also the time it takes it to detect that leakage. So while you may find that Australian RCDs may trip at 30mA they are faster at responding. I saw a nice graph somewhere which showed mA vs time and risks to human body at different levels.