IR test sets off gas hob igniters - WTF?

[NickD]

As per the title really, IR test @250V DC, (L+N)-E, whole-installation test of domestic property, main isolator switch open (obviously), testing caused the piezo igniters on the kitchen gas hob to spark even though no-one was pressing the knob to ignite them, and failure of IR test at about 150kohm. Switching off the appliance FCUs in the kitchen made it stop, IR test then passed at 2.2Mohm. What the hell's going on there?

 

[Marvo]

There are potential snags (pun intended) if you can't be sure of the test conditions. For example, an appliance that has been overlooked plugged into a socket outlet with SP switch that is off, will be subjected to test voltage without its L+N connected together.

Depending on the circumstances borrowed neutrals could also be a problem as well as single pole switched sockets and lighting circuits. Guess we'll agree to differ, I'm not predicting wholesale damage and destruction if you know the risks and fully understand all the possible test paths but given what I've seen on the forum there's a lot of guys who don't and I think it's an unnecessary risk to take especially in this day and age where the LED lamping in a premises could cost a small fortune.

 

[davesparks]

And if they do manage to damage something with an IR test then they will hopefully learn a valuable lesson about their own competence and maybe the client will think twice before choosing teh cheapest price next time!

 

[Guest111]

there is less chance of damaging something by going L+n to earth,notice I said less chance! If something is left connected,lets be honest how many threads have we had regarding strange readings on an IR test with "everything off or disconnected" only to find the "fault" is a neon on an FCU hidden away in a cupboard somewhere.

 

[NickD]

Cor, I didn't expect the Spanish Inquisition.


Job was additional lighting point on upstairs lighting circuit, and a spur to new double socket (protected by RCD FCU) on upstairs sockets. Problem was that the CU internals was a proper stuffed-in spaghetti job - no rhyme or reason to the conductor positions in the earth and neutral bars, and a complete pain in the backside to get any of them out and back in again even if there had been, and risk of doing more harm than good in trying. This will not be that unfamiliar a scenario to anyone who does a sigificant amount of domestic work, I suspect.


So I elected to do whole-installation (L+N combined)-E testing at 250Vdc, to allow for connected equipment, followed by live test for polarity (including switching being of the line conductors) and Zs at the new lights and sockets, plus a Ze test at the CU (earthing conductor out) to check the source of earthing.All of my new wiring I tested at 500V between all conductors with CPCs connected to earth prior to final wire-in to the existing circuits.


Yes obviously if the result is a big fail then you can end up chasing your tail, and functional earths (e.g. on RCD protected sockets/FCUs) and filter networks can cause said big fail, but if it isn't then it's happy days...?


At risk of stating the obvious, testing L+N combined to E at 250Vdc does not inject any voltage between L and N and so nothing is getting supplied L-N with 250Vdc. This is the same as an IR soft PAT test. I wasn't inclined to spend forever running round the house in advance unplugging, switching off FCUs, taking out bulbs, removing neons and dimmers etc. etc. and then missing something anyway. Didn't test L to N as connected loads would have caused test to fail and exposed items to the test voltage L to N.

 

[NickD]

PS I have yet to blow anything up testing L+N combined to E at 250Vdc. In fact the only thing I've ever blown up to my knowledge was an RCD FCU accidentally tested at 500Vdc L to N.

 

[Lucien Nunes]

testing L+N combined to E at 250Vdc does not inject any voltage between L and N and so nothing is getting supplied L-N with 250Vdc.

Not directly, however as per my previous post there can be internal paths via suppression components often from both lives to CPC. If these do not conduct symmetrically, for example when an SP switch defeats the L-N connection at the tester, then DC voltage can develop across L & N.

It's unlikely to cause damage, hence there's no reason not to do IR in a PAT test. The DC current through the appliance is limited by the tester's source resistance, also by one VDR in series with the rogue path to CPC plus there's the opposite VDR shunting the tester from the connected live (probably N if it's an SP switch that's off causing the assymetry) directly to CPC. So it's most unlikely that any damaging voltage would occur, even if the load is vanishingly small as is the ignition generator.

Obviously I can't see the ignition unit in question but I would be surprised if it deliberately passes line current to earth, in which case it could only have been powered by test voltage appearing L-N for a reason such as the above. The fact that the spark gaps use the cooker frame as one terminal is immaterial, because it is a transformer stepup arrangement where the primary is driven by impulses when a cap discharges through a GDT. I would expect the primary to return its current to N, but of course it rectifies the mains and will run on tester DC. Battery powered versions will operate from one AA cell for years of regular use, so you can imagine how little load they present to the mains.

If anyone knows of ignition circuits that do pass primary current to earth, I would be interested to hear how the circuit is arranged.