[Solved] Need help on home lighting and possibly on life choice

[Random Model Maker]

So, before we get into this, let me have a quick side rant

+++side rant start+++

So, I'm an EE student, as many here probably are, and at the moment, I highly doubt the german universit system.
Today, I tried to change the kitchen light in my flat.
That's when I noticed, that by now I am able to do fourier transformations and taylor polynoms without a calculator, I can work with matrices, vectors and tensors and all that good stuff, though I learned to use matlab and maple i detail, which we're not allowed to use although we could (duh).
To this point, no-one has shown us how to use a scope or even a meter, we don't know how whatever electronics, say opamp or 555 work, and I am, three years in, seemingly still unable to hook a bloody lamp to my kitchen ceiling.

So, what am I doing wrong? Was I expecting the wrong things? Did I inscribe to the wrong university? Am I studying the right thing? And does it get better?

+++rant end+++

To my problem.
In my flat, there were these old lamps, sixties-gramdma-style, big ol' 40W bulbs, that weren't mega bright too, so I decided to take them off and go for some new ones, GU10, so 230V as well.
First of all, weirdly enough, the old lamp was connected unearthed, although it clearly provided an earth terminal. The connected wires were, to add to my surprise, brown and green/yellow. (In case you didn't know, in central europe, usually brown is line, blue is neutral and yellow/green is earth)
So, measuring with a multimeter, I really had 230V from brown to yellow, so smells already of drunk uncle Bob wiring, but the floating blue wire measured 30odd V to the brown and 80odd V to the yellow wire. No clue where that's from.
So, I connected the line and neutral of the lamp across 230V and connected the blue odd voltage wire to the earth terminal on the lamp.

What happens now, is that the lamp works fine, when I turn it on to test it, then I go on and close the encasing and adjust the spots, still works fine, mind my business and suddenly 20 minutes later the lamps turn off.
So, flip the breaker, up the ladder open the lamp, everything's fine, turn it on, works, close it, and 20 minutes later, darkness.
Did the whole routine about five times.

I do not have many clues to what that is, but I have a few hints on what it ain't:
-Does not trip the RCD, not that I ever met one in this house, but if there is one in the basement I've never seen, doesn't trip that.
-Does also not trip the breaker, so no short there
-I can't measure the voltage because the error only occurs if the thing is closed
-Maybe has it to do with the misterious blue wire, is it worth a try to disconnect it? (Don't wanna be electrocuted, and for all it seems, this is the earth)
-Could be thermal, as those halogens can become quite hot
-When I leave it dangling on the wire clamp, it works perfectly fine.
-Could be that the lamp is shaffordable, students are always on a budget, it's one of those hardware store lamps, didn't cost an arm and a dick
-Occam's razor suggests it's a wonky connection, but the wire clamp seems tight to me
-Finally Murphy's law suggests a fault in the house's wiring, which meant ripping off the whole ceiling plates and rewiring the house

If anyone could either support one of those possibilities or exclude one, or maybe think of another solution to the problem, that I haven't thought of, I would be really grateful for support.
I have a second exact same lamp, I will, just to have tried, swap them out as soon as I got daylight.

[danadak]

Here in US first labs with instruments and soldering irons ocured, for me
in the early 70s, year 3 of 4 year degree program.

I had the good fortune of having been doing stuff since age 11 and a 6
year stint in Navy as a technician. I would encourgae you take the ball and run
with it on your own, build a couple of kits, get a low cost scope.....

Here is a zero costs scope that might be of interest -

If you need a scope, here is a PC based freebee that handles AC coupling only, but very usefull. Note in the
manual you have to limit input to sound card on the order of a volt, but still application, with FFT and Sig Gen
makes a very usable freebee instrument. Build an interface for the signal gen, a simple opamp buffer would do,
to buffer sound card output, and an attenuator/clamp on sound card input and you should be good to go. A
couple diodes paralled should work as a clamp.

https://www.zeitnitz.eu/scope_en

http://www.daqarta.com/dw_0all.htm

http://homediyelectronics.com/projects/howtomakeafreesoundcardpcoscilloscope/?p=3

http://www.build-electronic-circuits.com/sound-card-oscilloscope/

After you build the protective interface to sound card check it first by placing a higher V
on input to protective circuit and look at output of circuit that would connect to sound
card input.

Note this scope is only good for audio type frequencies.

Regards, Dana.

[steverino]

I'm assuming you only have visual of the wires from the lamps junction box?   I own rentals, some dating to the 1920's,  and you'd be surprised  at what's hidden behind walls, etc.  I had one unit that had an electrical outlet with a lamp connected to it and it started having periodic outages.  The voltage with the lamp off read 120V (US Mains voltage).  Plug the lamp in and the voltage would read maybe 50v.  Went up into the attic to chase the wire and found that the neutral was connected directly to a steel gas line!  Apparently, the current return path followed the pipe into the ground.  The wiring looked original.  (Added) Needless to say, I completely rewired the unit.

[jeroen79]

If you suspect that the colors of the wires do not match their function then I would try to sort out what is what first.
Or ask an electrician.

Do you have a picture of how the socket is wired now?

Why did you connect the blue wire to the yellow wire in the lamphousing?
What was your thoughtproces that led to this decision?
What happens if you undo the blue-yellow connection and just use the original wiring with the new lamps?

Earth should not be connected to neutral inside the socket.
It should be connected to the metal casing of a device incase that gets shorted to live or neutral.

Are you sure that the power is not interrupted, ie can you measure voltage on the lamp socket or another outlet?

Does the problem occur if you use a oldfashioned incandescent lightbulb?
Does the problem occur if you use the GU10 lights in a known good socket?

[helius]

You can be tricked by your multimeter measuring mains wiring because of its high impedance. It detects "ghost voltages" that are coupled from nearby live wires without any direct connection. The solution to this is either: 1. use a DMM with a Lo-Z mode, or 2. use a test load and measure the voltage across it. A 50KOhm, 1W resistor would be safe to use.

[sleemanj]

Uncertain wiring aside, this doesn't sound electrical to me, I think it's thermal.

It only happens closed.
It only happens after a period of time.
It doesn't happen if the lamp is free-hanging.

All says to me that when it heats up, it shuts off.

Now why that should be is another matter, could be as simple as heat expansion causing a broken connection, through to active electronics sensing a thermal overload and shutting down.

[Random Model Maker]

Alright, problem solved, looks like mr Murphy and mr Occam worked hand in hand here, it actually was a wobbly connection, but in the last place you would look for.
The lamps came with these terminal strip things, which were a bit dodgy.
Effectively, the assemblers in beautiful Shenzen had shoved the wires in the lamp that far into the strip that it couldn't properly clamp the live wire. The lamp just hung on the earth when I left it hang loose.
Why the lamp took several minutes to loose contact, still no clue.
What I suppose is that the contact point must have been so small that the copper oxydised away as the current ran through and then got contact back when I moved it.

As for the wire, the blue one was really earth, at least at that spot, I measured continuity to a wall outlet.
What's line and neutral, no clue, I have no power resistor on hand at the moment.

Anyways, thanks for the advice guys.
You're always a help here on the blog (in opposition to my professors...)

[Delta]

Are you on a TN supply or TT?

If you are on a TN-CS, I would *guess* that the blue really is the neutral, and has at some point gone high-resistance (loose terminal or fractured conductor), and some cowboy has simply connected the lamp across phase and earth to get it working again.

Good advice by Helius there, measuring mains wiring with a 1+ Mohm input multimeter is futile, you really need a relatively low impedance to load down any ghost voltages present.  An 10W incandescent lamp is good.