Help me before I potentially kill myself

[jbomba]

Hey guys I work in the Air-conditioning industry and would like to have a go at repairing inverter Pcb's. Obviously there are no schematics for these and it's trial and error. I've got a board at the moment which I presume has a short on it. Main fuse on PCB blows as soon as mains voltage is applied .I'm just trying to get my my head around safety equipment working on these. So far the list goes

RCD protected outlet
Isolation transformer
Variac

My plan with this short is to input a lower AC voltage into the board and bring the current to a lower than fuse rating level and look for hotspots with a thermal imaging camera. These boards invert to 380vdc so I just wanna make sure I've got safety sorted before I proceed

[ejeffrey]

Most important safety device is on top of your shoulders.  If you aren't confident you know how to do this safely that is a good sign.  A high power 380V inverter with a known fault and no schematics or documentation is not the greatest place to start working with high voltage / high power electronics.

A couple of other thoughts:

First off, on your equipment: an isolation transformer and a RCD don't play well together.  The RCD trips when it detects current returning to ground rather than the neutral wire.  An isolation transformer breaks the connection to ground so the RCD will not trip.  Use an RCD and not an isolation transformer.  RCDs also come in different ratings -- make sure you have one designed to protect life.  Some are only intended to prevent fire or protect equipment and have trip thresholds too high to be useful as a safety device.

Second, if you get to the point of measuring an energized circuit you need to think about how to do measurements.  High quality multimeters rated for high power operations and oscilloscopes with high voltage differential probes are the best way to probe a live system like this.

But the most important things to do come before you apply any power.  First off, make sure any high voltage capacitors are discharged and then physically inspect the board.  If you have a known-good board, compare them.  Really look in detail.  You may be able to see a short, or a capacitor that is bulging or a MOSFET with scorch marks.  Even with relatively big components it can be helpful to look with an inspection microscope -- that can help you notice wire fragments or solder balls that have gotten in the wrong place.

While you are at it, try to figure out the basic circuit design and layout.  You won't get everything necessarily, but you should be able to get a good idea of how the circuit operates just from looking at it.  This will give you some idea what to expect while probing it.

Next measure the board with an ohmmeter before applying power.  You may be able to identify a short with no power at all, for instance if you have a capacitor with a short circuit fault.

Only once you have done a thorough inspection, familiarized yourself with the board, and attempted diagnosis with no applied power should you even consider probing it live.  And if you are not comfortable at this point, or if you don't have a good idea of what things you want to measure, then don't try.

That said, if you have followed all the above steps, have the device connected with a correctly rated RCD, applying a small voltage and examining with a thermal camera is a pretty good technique.

[james_s]

Learn how an inverter works and then start testing components, don't just apply power and see what gets hot, that's a good way to completely destroy something that may otherwise be fixable.

[AndyC_772]

The one thing you know for sure is that any equipment you're asked to fix will already have been plugged into the mains in its non-working state.

I wouldn't worry too much about causing further damage by plugging it in again. Whatever was going to go bang, will have done so already.

[cowasaki]

First off, on your equipment: an isolation transformer and a RCD don't play well together.  The RCD trips when it detects current returning to ground rather than the neutral wire.  An isolation transformer breaks the connection to ground so the RCD will not trip.  Use an RCD and not an isolation transformer. 

I have a new house with everything fully RCDd.  I've just bought an isolator transformer but not yet fitted it.  So what you are saying is that an RCD is better than an isolation transformer and I've wasted my money?  My isolation transformer would be plugged into the mains ring in the workshop which is itself protected by an RCD.

[CM800]

Number one thing with Inveters... the DC bus capacitors. possibly upwards of several kJ in some units, at 450V or 800V+
Literally the most lethal form of energy you can find in electronics, short of in Labs or some specialist fields.

Something like a capacitor discharge tool like this would be invaluable:

http://jestineyong.com/build-your-own-capacitor-discharge-probe/

This is a really nice design as it has a small LED designed into the circuit so that it lights up until it is safe. I would still probe with a multi-meter after just to make double-sure....

[BurningTantalum]

RCDs don't need, or use, a ground connection to operate. The GFCI devices used in the US socket outlets may operate on a different principle- I have never seen one in the flesh.
BT

[hermit]

How many things can you test on this board for shorts before resulting to this setup?

[madires]

There are several threads with RCD vs. isolation transformer discussions in this forum. Both got their use cases. For working on mains voltage stuff I'd prefer the isolation transformer (for powering the DUT), differential probes for the scope and one hand in my pocket. Rubber gloves would be a nice extra.

[Ian.M]

RCDs and GFCIs dont work with isolating transformers.

Neglecting the tiny leakage current due to inter-winding capacitance  KCL tells us that the current into one end of any single transformer winding must be equal to the current out of the other end.   Therefore there can never be an imbalance on the primary side of an isolating transformer due to a fault on its secondary side, so a RCD/GFCI feeding it wont trip.   Equally, if you attempt to put a RCD/GFCI in its secondary circuit, there cant be any imbalance, so it wont trip on any fault, unless you ground the secondary neutral before the RCD/GFCI, which renders the isolation ineffective.

If you choose to use an isolating transformer for safety, it is essential that nothing on your bench introduces a ground to the D.U.T.     Connecting *ANY* test equipment which has a signal ground that isn't isolated from Earth ground (and rated for a differential greater than your  DC bus voltage + peak mains voltage) would make single point contact with your D.U.T potentially lethal. 

For working on mains voltage stuff I'd prefer the isolation transformer (for powering the DUT), differential probes for the scope and one hand in my pocket. Rubber gloves would be a nice extra.

+ a thick rubber mat on the floor, and eliminating any other grounds in the immediate workbench area.   

You've also got to worry about the energy stored in the DC bus capacitors - one slip-up when probing and a component could explode.   Your PPE should be resistant to arc flash and impact of high velocity fragments.

IMHO, Don't take on this sort of repair work without a full risk assessment signed off by yourself + a health and safety officer with experience of best practices for live working in your industry.

[Mr. Scram]

No worries, potentially killing yourself is not a problem. Actually killing yourself is where things go awry.

[Ian.M]

Just make sure your funeral plan is up to date, and tell your friends and relatives not to expect an open coffin viewing.

[wkb]

Just make sure your funeral plan is up to date, and tell your friends and relatives not to expect an open coffin viewing.

Yes, that. And remember that in case you plan on getting cremated you might be able to negotiate a discount.

[cowasaki]

I’m not going to risk anything till I have the full information.  At £45 the isolator was a bargain. It can sit for months while I learn the best for to do things and if it doesn’t involve the isolator I can always sell it on for what I paid.

[Ian.M]

Used correctly, an isolating transformer makes any single point contact with the D.U.T. (apart from EHT rails greater than its isolation rating) minimally hazardous.   However it strips away any extra safety offered by RCD/GFCI protected supply, so used incorrectly, it makes the D.U.T. far more hazardous.    The difference between reasonably safe and a deathtrap waiting to kill you can be as simple and slight as moving a scope ground clip 1" across a PSU board.

You need appropriate equipment, specific knowledge, experience, and state of mind to stay safe, and IMHO its *NOT* practical to teach it in a forum.  It should be developed in a controlled environment (i.e a classroom + lab or by one-to-one mentoring), where the hazards you are exposed to are limited and the instructor is fully aware of all the specific risks of the D.U.T you are training on.

[IanMacdonald]

The need for the isolating transformer is a practical rather than safety one, in that you can't make measurements on a circuit fed via a bridge rectifier with grounded test equipment otherwise. Since, all of the DUT will have a half-sine pulsating voltage WRT ground.

As said, the safety aspects of this kind of work are considerable and getting professional training would be the best approach. I was self taught in CRT TV repair, mind you, which had similar or worse hazards. The most important principle is the 'swordsman's stance' of one hand behind back when doing live tests. If you need to use two hands, always power off and discharge caps first.

I've once come close to death, and that was when a metal bench lamp touched a heatsink with 1200v on it. Which underlines that the type of equipment and fitting on your bench influences safety too. the less earthed metal, the better.

[alm]

The need for the isolating transformer is a practical rather than safety one, in that you can't make measurements on a circuit fed via a bridge rectifier with grounded test equipment otherwise. Since, all of the DUT will have a half-sine pulsating voltage WRT ground.

I would argue against the use of test equipment with ground-referenced inputs in this case. It's risky to use and easy to fry on these high-energy circuits, and I don't see the need.

For scopes, you either need some high voltage differential probes, or maybe better a scope with isolated inputs designed for these high-energy circuits (e.g. Fluke ScopeMeter) with appropriate IsoPower probes. A ground clip of a regular scope probe coming off could easily result in fireworks.

What other grounded test equipment would you want to use? A spectrum analyzer? Frequency counter? Signal generator?

[cowasaki]

I had training in a classroom 25 years ago but need a bit of a reminder and some things have changed. I am possibly looking at the isolation covering the test equipment with the RCDs covering the equipment I am working on. I rarely work on live equipment and I am extremely careful when doing so but anything that helps is good. I was always taught to repair anything plugged in with one hand in my back pocket. :-)

[Electro Detective]

RCDs don't need, or use, a ground connection to operate. The GFCI devices used in the US socket outlets may operate on a different principle- I have never seen one in the flesh.
BT

Sorry mate, but AFAIK 99.999% of the RCDs here in he Australia region (and UK) WILL NOT TRIP unless they are part of the Earth/ground chain 

otherwise how would a 'fault to earth between LIVE or Neutral' cutoff device possibly work? 


Tip for young and old mains electricity players still alive, thanks to russian roulette LUCK: 

If you are going to blindly trust an RCD device exclusively to bail you out of OD (Oops and Dumbass)
then use TWO portable RCDs in series (one plugged into the other) to lower Murphys Law effects,
and test them beforehand with an external RCD/Polarity tester, not just the on board feelgood -TEST- button   

plus a fast acting separate double pole MCB for faults and shorts between LIVE and neutral, because in this BIG BANG! non theory scenario the RCD will NEVER work 

A dim bulb tester wouldn't hurt the pocket too much either,
but it like the isolating transformer (if it is TRULY isolated) and variac (which is never isolated), aren't really going to do you any favors unless thoroughly familiar with those items and their limitations,
and how to use them for particular purposes

[jm_araujo]

Sorry mate, but AFAIK 99.999% of the RCDs here in he Australia region (and UK) WILL NOT TRIP unless they are part of the Earth/ground chain 

The ones I know trip when there is a current difference between the Neutral and Live wires, so no Earth/ground connection to the RCD is needed. They will trip even if the current leak is to a earthed element not directly connected to the mains panel earth wire, like a water pipe (edit)... or a human!

[madires]

RCDs don't need, or use, a ground connection to operate. The GFCI devices used in the US socket outlets may operate on a different principle- I have never seen one in the flesh.
BT

Sorry mate, but AFAIK 99.999% of the RCDs here in he Australia region (and UK) WILL NOT TRIP unless they are part of the Earth/ground chain 

otherwise how would a 'fault to earth between LIVE or Neutral' cutoff device possibly work? 

... by comparing the currents passing through L and N. They should be the same because N is the return path of L.

[LA7SJA]

Check out Mr. Carlson's Lab's videos on the subject.




Johan-Fredrik

[Electro Detective]

"RCDs operate by measuring the current balance between two conductors using a differential current transformer.

This measures the difference between current flowing through the live conductor and that returning through the neutral conductor.

If these do not sum to zero, there is a leakage of current to somewhere else (to earth/ground or to another circuit), and the device will open its contacts.

Operation does not require a fault current to return through the earth wire in the installation; the trip will operate just as well if the return path is through plumbing, contact with the ground or any other current path. Automatic disconnection and a measure of shock protection is therefore still provided even if the earth wiring of the installation is damaged or incomplete.

Residual-current detection is complementary to over-current detection. Residual-current detection cannot provide protection for overload or short-circuit currents, except for the special case of a short circuit from live to ground (not live to neutral).

For a RCD used with three-phase power, all three live conductors and the neutral (if fitted) must pass through the current transformer."

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

I don't totally buy into some of the copy paste wiki info above, 
if anyone is game to try an RCD sold in Australia on a TRUE isolation transformer, they may be surprised that it won't work,
because there is no leakage to earth/ground or anywhere else in most cases to create an imbalance and make it trip

and any RCD sure as hell won't trip on a LIVE to Neutral BANG!  unless it's an RCBO or combined RCD/MCB type

extra surprise surprise: if an RCD does trip when used with an isolation transformer, then that particular transformer is NOT truly/fully isolated

i.e. even if you THINK you know what you're doing, and have all the right gear, some combination of the above may bite you 
 

[madires]

I don't totally buy into some of the copy paste wiki info above, 
if anyone is game to try an RCD sold in Australia on a TRUE isolation transformer, they may be surprised that it won't work,
because there is no leakage to earth/ground or anywhere else in most cases to create an imbalance and make it trip

That was mentioned before. Since the secondary of the isolation transformer is floating, i.e. not referenced to earth/PE, there can't be any current to earth/PE. Therefore an RCD can't be triggered.