The bulbs "shoot" when the light in the room is turned off with the sw

I have a lamp on the ceiling for 4 E27 bulbs. I have a double switch by the door. Bulbs are connected 2x2, i.e. two ba one switch. I have such a strange problem.
Sometimes when turning off the lights. No matter which light bulbs, the bulbs actually explode. It makes a flash like a strobe, a bang like a firecracker. The bulb is intact. For once it happened that the bubble burst. It strikes an S immediately in the box, and once even knocks out all 3 in front of the counter, the main ones that are connected together into 3 phases.
I usually have light bulbs because I am worried about leds. I had a LED bedside lamp turned on once and it was a hell of a shit with this incident. Today, when disconnecting the electric heater from the electric strip, the same happened. But I suspect that it has nothing to do with it.
I recently took off the lamp to clean it. The contacts in the frames are ok. I have read it prophylactically with fine dirt paper. The switch also looks ok.
Csrgo could be the reason? This has never happened before with the same connection. It has been happening notoriously for 2 years.
Thanks in advance for the hints.

The matter is quite serious, because there is a probability of a break in the N line of the installation.
Perform a check, including the measurement of the fault loop impedance.

So that's it. It is all the more strange that it only happens in my room, and I have the entire level connected to the same S.

What is the manufacturer of these bulbs?

I changed producers with dirt. Maybe I already had 3 companies. I changed because I thought the bulbs were of bad quality.

Are they still producers from across the eastern border or, for example, Osram?

I don't remember to be honest. do you think it may be the fault of these bulbs?

Measurement of the short circuit loop impedance and voltage on the light bulbs under load will allow you to diagnose a possible fault in the installation. In the case of a three-phase installation, there is a risk of a failure described by the col. Elpapiotr and damage to other receivers (and even electric shock or fire). Especially that the burnout of the light bulbs (possibly related to the voltage increase) occurs when turning off other receivers. However, such a test can only be performed on site by an electrician with the appropriate measuring instruments, it cannot be done remotely. The matter is urgent, until then I will be able to disconnect the other receivers from the installation.

On the other hand, my experience shows that some manufacturers' bulbs (from the eastern direction, regardless of the country) tend to burn out after 2 weeks and to effects also related to the bubble explosion. Bulbs from reputable manufacturers (eg the aforementioned Osram) cost a bit more, they have a halogen bulb embedded in the bulb and in my opinion they are much more durable and do not "explode".

Hello, of course, that loose "N" and when the load or its removal increases the voltage from other receivers supplied from the 3-phase switchgear, that is, they get too high voltage, look up to the farthest available terminal "N".

In that case, I will check carefully all the connections in cans, etc. I will add that this happens every 1-3 months. But it's still too much. Compared to the fact that nothing like this had happened before.

There is no need to check in boxes only in the circuit board where there are fuses for individual circuits. But two years it would have to be something to grill (burn out).

If you remove any errors in the installation and the problem with the bulbs does not disappear, you need to replace the wall switch with which you turn the bulbs on and off.
There are very poor quality switches available on the market and there is a strong sparking when switching off, which has known consequences.
A similar situation may occur with a badly worn circuit breaker.
If you replace the bulbs with halogen ones, the sparking could be even more intensified.

Everything may point to N / PEN problems, but for just about two years now, all bulbs are of dire quality. I can see this in the company where I work, where unfortunately there are still many light bulbs. We do not buy Osram bulbs, of course. Now we want to replace everything with LEDs (maybe half is already here) - bulbs last up to two months (they shine for 2-6 hours a day). As you wrote, the bulbs shoot and cause overcurrent protection to work. And a curiosity ... in our case it is caused (after a quick analysis) "FOR A GOOD" SHORT-CIRCUIT LOOP! despite the old communist installation
Added after 55 [minutes]:
The effect is that when the "filaments" in the bulbs break and there is a short circuit loop at the level of eg 500A to the socket and such a filament falls on the lead wires, the overcurrent protection 10A is activated. With a short-circuit loop, e.g. 100A, tripping of the overcurrent protection is less likely. From experience about 4 years ago such situations were rare. It is everyday life now. The "EU bulbs" are also causing much more short circuits on the stems. In the company I work for, there are several thousand light sources, a few transformer stations, a few buildings, so it's not like that all the N / PENs loosened right away. Such a crap just appeared at some point and the short circuit loop is of colossal importance in the operation of security, even in the event of a stupid bulb burnout. The same is of great importance in the case of switching overvoltages - which we have experienced more than once.

Buddy, the short circuit loop has nothing to do with it, because the reason it works is the quality of what you call a light bulb.

From what you write about "too good short circuit loop" it appears that designers are doing a mole job trying to make the short circuit loop as low as possible. Because only then does it guarantee the fulfillment of the SWZ condition in circuits of a considerable length and in circuits with protection of a greater value or with higher characteristics.
And that's why your analysis is not holding up, because it suggests that security can be compromised so that the blows that burn through do not cause the security to work.

If you wrote that the voltage is too high, okay, but do not mix the short circuit loop with it.

By the way, the short circuit loop impedance value is given in ohms, not in amps. Amperes are short circuit current.

Yes, the fault loop is given in ohms. I am dealing with the current state, i.e. the projects were created 30 years ago and the transformer stations are up to 50m from the buildings. It is good that the conditions of the SWZ are met. I just note that a few years ago there was no such problem with light bulbs and now it is, so I am going to replace all of them with other light sources as soon as possible.
There are no too high voltages. We have a network parameters analyzer at the connection with the highest ordered power and never (even for a minute) the voltage on the phase does not reach 230V. Intervention of protections is a direct result of too "good" fault loop. The wire / cable impedances are simply too low to "dampen" short circuits.

Quote:

... a few years ago there was no such problem with light bulbs ...

And stick to that. Because they used to be light bulbs, and today it is a bulb-like product.
So do not blame the short circuit loop for what is happening (surely its value does not decrease with time) because such a diagnosis is heresy.
And the protection is just to work if a faulty receiver causes a short circuit.