I had a multimiter years ago, but I don't remember its model and its functions, I only remember that it was yellow. 

The "volt meter" that I use now is a bare neon screwdriver, and it's one of my favorite tools. I enjoy to detect the dangerous mains leaks of the cheapo power supplies.  Next time I'll a light bulb with two probes, one to neutral and the other to the exposed metal of many cheapo things.

With this little friend I will say to you that NEVER touch an exposed metal of many famous brands flat TV's like coaxial cable shielding or video ports while you grounded, because for shure you will get a surprice!

I practically wired all my house antenna network, and I got few zaps. 

The Neon will happily light up at some 10's microamps and quite brightly at 0.5mA, but you begin to feel some tingling only above 0.5mA...1mA or so at 50Hz. Many of the leakages of such appliances are through small capacitances (Y capaciitors or parasitic) that can't provide more than 0.5..1 mA

I'll connect my subwoofer to a 10v transformer, way safer than connecting it directly to the mains, what I was thinking. 

If I connect my 8 ohm speaker to the 10V AC, I should get 13W of discipation in the speaker, I don't know if my 4.5" sub will be overdrived.

Be aware the speaker rating assume the power load is only for short time, not a constant loading.
For constant loading it will overheat and/or break mechanically.

4.5inch sounds to me as very small for 50Hz, it will very likely reach its cone displacement limits (hope the magnetic and not the mechanic ones, as the later would mean it will disintegrate after very short while) for anything above 0.5W or so. Plus I have strong doubts the 4.5inch is even able to radiate anything at 50Hz, these small "subwofers" usually just cover the range from 300 down to 100Hz and then they fall down very steep (the 100Hz is already resonance boosted).

I'll ad a 10w 220v bulb in series for limit the current. The subwoofer that I want to use perfectly produces 50Hz sounds in my tests (I did't tested it in the mains LOL). Subwoofes produces sounds below 100Hz, while woofers produces sounds over 100Hz and below 6KHz in most cases. My speaker have more expansion capacity than other speakers of the same size, so it is a subwoofer.

I got a multimiter yesterday, and I did some mesurements. The speaker that I have is a 4Ω speaker. I also measured a 220V 40W incandent lamp in cold state, and it is 92Ω. So if I connect all in series and power it with 10VAC all will have a discipation of 1W, a safe power that can't blow the speaker. Anyway if speaker start to overload the bulb will warm up and will change its resistance.

But would you hear anything (clearly enough to understand what the sound is) at 0.04W power input to the speaker ? Even with connecting this directly to 230V, you will only get 0.12W ...

I did't considerated the power that the speaker will dissipate, only the whole system, and I'm not good with formulas. Maybe using a 70 or 300w lamp will make the speaker dissipate more power.

But would you hear anything (clearly enough to understand what the sound is) at 0.04W power input to the speaker ? Even with connecting this directly to 230V, you will only get 0.12W ...

But 0.12W could well be way enough, if used with speakers that cover that range...
Normally home audio running as a "background" does not operate at more than few 10's of mW of the average power, so 120mW is not that small power. It is the sensitivity of many speaker systems being very low at 50Hz, which may require to be compensated by higher power, but there the differences among different models are in the excess of 30dB, so for the same accustic pressure level the required power could easily vary from mW up to W range...

The sensitivity of our hearing at 50Hz is some 5..10dB lower than at frequencies that appear the most in home audio..

The sensitivity of our hearing at 50Hz is some 5..10dB lower than at frequencies that appear the most in home audio..

But I dont think anyone would tolerate any monothonic hum to be as loud as normal audio uses to be in nirmal use...

The Neon will happily light up at some 10's microamps and quite brightly at 0.5mA, but you begin to feel some tingling only above 0.5mA...1mA or so at 50Hz. Many of the leakages of such appliances are through small capacitances (Y capaciitors or parasitic) that can't provide more than 0.5..1 mA

The neon can glow below 1 microamp.

The neon can glow below 1 microamp.

And the leakage up to 1mA on exposed metal parts is normally allowed. Provided the element responsible for that current is of a proper safety type (designed so it does not reduce its impedance up to few kV, observes proper isolation distances and there is no single point failure possible leading to dangerous currents; so Y rated).
Commonly are used capacitors of few nF to provide RF bypass (in power supplies), in some devices few MOhm resistors to provide static electricity leakage path (CRT outer frame connection to the internal instrument ground) or provide the main functionality (the resistor within your glow lamp voltage indicator). All these have one thing common: Certain current there is perfectly normal, even if it lights that indicator lamp (otherwise the indicator itself would be a death trap).

You have to always take into account what you are dealing with when evaluating any test, mainly the safety related ones otherwise you may very easily make big error of pointing onto something which is perfectly fine, but then overlook the real dangerous fault. The thing is, the first would become dismissed very soon, but as the main fault would be still uncovered,the device becomes hazard again. All because of that judgement error you made by mislabeling a normal designed feature as a dangerous fault.

My neon screwdriver is rated to glow to voltages from 100 to 300v, but it is glowing to a 10V AC leak of the TVs in my house, at 15 microamps, that's weird. Also if I touch the coaxial cables shield and my fridge with a mains shorting fault with a tiny red LED it barely glows, both with the same brightness.

My neon screwdriver is rated to glow to voltages from 100 to 300v, but it is glowing to a 10V AC leak of the TVs in my house, at 15 microamps, that's weird. Also if I touch the coaxial cables shield and my fridge with a mains shorting fault with a tiny red LED it barely glows, both with the same brightness.

"... a 10V AC leak..." How do you know the leak was limited to 10V? Leak means current source, so it passes that current and the voltage is then dictated by the load resistance. So it could well be, that TV leak was actually from way higher voltage, but the 10uA meant it caused just 10V drop across the V-meter internal resistance, so it read "10V". But in realitry your meter was not measuring the open circuit voltage, but the current of that leak. When the glow lamp tester was connected instead, it has higher resistance (easily >100 GOhm when there is no discharge in the bulb), so allowed the voltage to go up, till it strikes the discharge in that bulb. And 10uA is high enough to make the glow discharge visible there. It is in fact about the current normally flowing into your body when you are isolated from ground and just touching the phase (e.g. via the tester - compare to the reactance of your body when isolated from ground the tester is nearly a direct connection). Your body capacitance is about 150pF, which leads to about 11uA at 230V 50Hz. And about 68pF (of course Y-rated) were the common antenna coupling capacitors in "hot chassis" TV designs, allowing the RF to pass, yet provide isolation for the mains voltage.

I measured the leaks between the leak source and a big metal surface. So for correctly measure the darger of a leaking I'll need to measure between the leak source and neutral or ground.