A/C unit small leak to ground

[jdm1]

While rewiring the room I found that my Daikin aircon inverter (exact model on request) leaks some ac voltage in the power ground lead. It works fine otherwise. An old supercheap tester says it 30V and decreasing. A not much more expensive says is 112 V t-rms (note we have 230V), 50Hz, 20 mA and steady. I find the discrepancy strange and don't know which one to believe. I can try another needle type meter,  if I can find it.
For now, I left the ground lead disconnected.
Anyway,  i can surely replace the power cable but then what if isn't that? Is there a common failure mode for this, like a cap failed short? Or do I have to study and go with the service manual?

[fzabkar]

Are you saying that you measure 112VAC on the ground lead when you leave it floating? If so, then that would be normal for a 230VAC installation. You will probably find noise suppression capacitors between ground and each of the Active and Neutral terminals.

[TheMG]

In addition to interference suppression capacitors, there is likely going to be some capacitance between power devices mounted on the heatsink, and even some capacitive coupling from the motor windings as well.

The entire inverter circuitry and its outputs are not isolated, so they are mains referenced but through a bridge rectifier, which is probably why it's giving weird readings on some multimeters.

The total amount of leakage due to all this capacitive coupling is going to be non-negligible in a unit like this.

[wraper]

Only current is meaningful figure in this case. If it's really is 20mA, it's an order of magnitude too much. Measuring half of the mains voltage is normal due to leakage through noise suppressing Y type safety capacitors between live wires and ground.

[jdm1]

Thank you all for your answers. I've found and attached a document that on page 7 mentions a range of 0.25mA~3.5mA for Y class capacitors.
And I have an identical A/C unit which I've measured to 2.5mA on the ground lead when on, 0.06mA when off. Indeed 25mA on the other seems a tad too much to me.

Now, working on the external unit of the suspicious A/C is not that difficult but in this season I'd rather stay fresh in my renewed watch repair room.
Do you think this a repair that can postponed, even indefinitely? Beside obviously bulged caps, where should I start looking? I only have a DMM and an open source component / ESR tester.

[wraper]

It's not about bulged capacitors. It's about Y rated safety capacitors (usually blue ceramic disc), or insulation problems. Probably there is a liquid damage somewhere which compromises insulation.

[TheMG]

For now, I left the ground leader disconnected.

You've reconnected it, right? DO NOT leave the ground disconnected, it's there for a reason.

The worst that could happen, with the ground connected, is that it trips an RCD/GFI (if there is one), or if the leakage turns into a short to ground it will trip the breaker.

With the ground disconnected, there is a risk of electric shock.

[Monkeh]

For now, I left the ground leader disconnected.

Stop right here, call an electrician. You should not be messing with this.

[fzabkar]

For now, I left the ground leader disconnected.

Stop right here, call an electrician. You should not be messing with this.

If the problem is on a PCB, what will an electrician do? He will probably just replace the PCB, or tell you to call the AC serviceman who will then replace the PCB.

[Monkeh]

For now, I left the ground leader disconnected.

Stop right here, call an electrician. You should not be messing with this.

If the problem is on a PCB, what will an electrician do? He will probably just replace the PCB, or tell you to call the AC serviceman who will then replace the PCB.

Nothing to do with the AC. He is rewiring his house and just intentionally left a protective conductor disconnected on a piece of equipment he knows to be faulty. He has no business working on any of this wiring.

[fzabkar]

Are there two varistors at the mains terminals?

Here are some service manuals:

https://elektrotanya.com/showresult?what=Daikin&kategoria=All&kat2=All

[jdm1]

He is rewiring his house and just intentionally left a protective conductor disconnected on a piece of equipment he knows to be faulty. He has no business working on any of this wiring.

Yes I'm rewiring my house, as I have done since age 14 and for many decades now, as my knowledge and safety concerns plus the law in my country allows. Yes I have left a lead ground disconnected while testing, with the good knowledge that in the unlikely event that if live shorts to ground nobody will suffer, as it's an external unit that can't be reached, So, thank you for your paternalistic advice but I think it's you that have no business in this constructive discussion, so any further reply from you I will ignore.

[jdm1]

The worst that could happen, with the ground connected, is that it trips an RCD/GFI (if there is one), or if the leakage turns into a short to ground it will trip the breaker.

Of course there is one (actually more than one for the different circuits). It's from a leading brand, tests OK, on the label shows an IΔn 0.03A, so the question become, why it doesn't trip @ 25mA? I have no problem replacing but in the interest of safety I would like to fully understand what's going on.

[wraper]

label shows an IΔn 0.03A, so the question become, why it doesn't trip @ 25mA? I have no problem replacing it but I would like to fully understand what going on here.

Because trip current is not exceeded.

[Monkeh]

He is rewiring his house and just intentionally left a protective conductor disconnected on a piece of equipment he knows to be faulty. He has no business working on any of this wiring.

Yes I'm rewiring my house, as I have done since age 14 and for many decades now, as my knowledge and safety concerns plus the law in my country allows. Yes I have left a lead ground disconnected while testing, with the good knowledge that in the unlikely event that if live shorts to ground nobody will suffer, as it's an external unit that can't be reached, So, thank you for your paternalistic advice but I think it's you that have no business in this constructive discussion, so any further reply from you I will ignore.

The ground is then tyically passed to an indoor unit along with power. Your 'unlikely event' is occuring right now, as a potentially lethal amount of current is connected to now unprotected metalwork outside and possibly inside. My tolerance for DIY ends at the creation of a dangerous situation.

It's from a leading brand, tests OK, on the label shows an IΔn 0.03A, so the question become, why it doesn't trip @ 25mA?

What current does it actually trip at? If properly tested you would have a number for that - the test button is crude and does not prove the device works within specification. And, as mentioned, 0.025 < 0.03, so it may or may not operate on such a leak - assuming the leak is pure AC rather than rectified, which may not cause some types to trip at all, and may even prevent them from tripping in the presence of a true AC fault.

[themadhippy]

The worst that could happen, with the ground connected, is that it trips an RCD/GFI (if there is one), or if the leakage turns into a short to ground it will trip the breaker.

With the ground disconnected, there is a risk of electric shock

.
in this case id say theirs a greater risk of a shock with the earth connected as all points connected to earth will have the fault current on them,if no rcd is fitted and the fault current  is below the fuse protecting the circuit its just going to sit there waiting for its next victim.If the external unit is  insulated  and placed out of reach theirs no chance of the fault current getting to other earthed bits or people touching it accidentally . However its still an accident waiting to happen.

[wraper]

The worst that could happen, with the ground connected, is that it trips an RCD/GFI (if there is one), or if the leakage turns into a short to ground it will trip the breaker.

With the ground disconnected, there is a risk of electric shock

.
in this case id say theirs a greater risk of a shock with the earth connected as all points connected to earth will have the fault current on them,if no rcd is fitted and the fault current  is below the fuse protecting the circuit its just going to sit there waiting for its next victim.If the external unit is  insulated  and placed out of reach theirs no chance of the fault current getting to other earthed bits or people touching it accidentally . However its still an accident waiting to happen.

What? Earth shouldn't be floating, so there is no potential victim. Earth is protective wire which should nullify any voltages caused by current leakage on enclosures and other parts exposed to the user.

[NiHaoMike]

Might an insect or other small animal have got inside bridging a connection to ground? I have seen a locust under the compressor terminal cover on an A/C unit I opened up to troubleshoot, wasn't the cause of the problem but I did "debug" it while I was there.

[Circlotron]

Seeing it's an outdoor unit, why not just drive an earth stake into the ground right next to it and run a wire to it? And reconnect the original earth wire too.

[wraper]

Seeing it's an outdoor unit, why not just drive an earth stake into the ground right next to it and run a wire to it? And reconnect the original earth wire too.

[SeanB]

Common on AC units to get ground leakage, normally this is from condensate water in the indoor unit from a blocked drain, which needs to be cleaned out along with the tray and the coils, as they do collect a lot of dust and salts. Outdoor unit as well will have some leakage, normally with inverter from dust on the bare PCB in the unit, as they are not normally enclosed and tend to collect dust as well. Brush off the PCB with power disconnected ( not turned off, the outdoor unit board is powered all the time with AC power connected to keep the communications link alive) after waiting 15 minutes after disconnecting, as the power supply there typically has 2200uF 400V of capacitance on it, or more, and that will bite. Leakage also common at the compressor terminals, remove the cap and clean it with some spray oil, wipe off and put the cap back on.

[tkamiya]

Taken OP's statements as a whole, I wouldn't worry about it too much.

Cheapo meter shows 30V 20mA and decreasing.  That means source impedance is quite high and VOM is enough to discharge it rapidly.  I don't know what "much more expensive" meter he is using but that one probably have quite high input impedance.  I'm sure if he waited long enough, it will go to near zero shortly.

Heavy electrical equipment is not the same as small circuitboards we normally deal with.  Harsh outside element and its construction contributes to this. 

If you have multiple ground rods, they must be bonded before it enters the house.  Please do not drive one right next to it and connect to it.  (yes, I am an electrician).  If I were you, I'd just reconnect the way it was and leave it alone.

[Circlotron]

What's the issue with having an additional ground rod right next to the outside unit? Leave the normal ground wire connected of course.

[wraper]

What's the issue with having an additional ground rod right next to the outside unit? Leave the normal ground wire connected of course.

That it does not help and does not fix the faulty isolation issue. Fools errand in other words.