I was wondering if the modern lamps are made to be more tolerant of power fluctuations.  Reason I ask is because tonight while working, there were a couple of power surges due to a storm.  All of the lamps that are only about a year old stayed on while the rest went out. I just wondered if they were made to be more tolerant to these variances than the older lamps are.

I think maybe old lamps are made to tolerate more. Nowadays eletric networks is pretty stable compared to 50-30 years ago(at least in Finland) many Overhead power lines(voltage under 20,5kV) has been replaced with normal cable that has been burried(sorry dont know the right word). So storms doesnt be so big problem. Ofcourse countryside is different thing. In Rovaniemi Radio- and TV-station they had huge variacs(If I remember right total RF output power of tv transmitters was 35KW so variacs had to be huge) that were used to stabilize electric to transmitter equipment. these were removed some years ago when there were no use to them anymore. electric power was stable enought. And maybe ballast/control gears has nowadays better manufacturing tolerance than before(e.g. lamp voltage/current) But this is just guessing....

The main reason is that lamps with fewer burning hours hold up better during the voltage dips, as they require less voltage to maintain an arc. (It has nothing to do with current-production lamps, which in many cases are more poorly-made than ones manufactured decades ago).

This is particularly an issue with HPS lamps, but also to some degree MV & MH - as the electrodes burn away, the arc tube gets blacker & retains more heat, which in turn increases the operating voltage. When the ballast's supply voltage drops, so does the voltage to the lamp, and if there's not enough to keep the arc going, the light goes out.

I've noticed that with my HPS outdoor light circuit (5 lamps), which is on the same circuit as the microwave (for some reason, that's the way my house was wired), an HPS lamp near the end of its life will cycle if I turn on the microwave (which uses a lot of energy, thus dropping the line voltage on that circuit). Normally, that doesn't happen.

The other reason is the ballast. CWA ballasts have better regulation and thus if the supply voltage drops a little, the ballast compensates for that and keeps the lamp lit at the original level (thanks to the capacitor). My OV-20 is operated off a CWA ballast and during wind storms when the lights blink momentarily, it stays lit while all my HPS & mercury fixtures (on HX ballasts) cycle.

Id like to add just for info that the "regulated" ballast performs the best for any HID during voltage dips. They are usually larger and have three coils instead of one or two. The third coil is wired to the capacitor.

There are some older GE CW ballasts with four coils but they are still wired like a normal constant wattage (CWI or CWA) but still perform better than reactance (HX, no cap) type as Alain said.

I've noticed that my 175W MV fixture is super sensitive to voltage fluctuations ( I live on a generator and off the grid) and so are magnetic PL preheat adapters, I used to have an ABCO one that was 13w and went out and restarted every time my chest freezer cycled in, as does my MV light that I mentioned above. Magnetic rapid start f40t12 ballasts seem sensivitve too, CFLS re somewhat (but way less) sensitive, and the elecrronic IS ballasts in my kit  hen and bathroom lights are pretty much immune to small voltage fluctuations.

The lamp wear is indeed the major difference.

The complete "mag-reg" ballast have quite high energy stored in the regulating resonator (the coil with the cap).

The US PLS's are quite sensitive, because they have quite high arc voltage compare to the mains (above the 1/2), what make it "just on the edge" to maintain the arc with normal supply, but unable to hold with even small dip.
The same are series reactor ballasts on lower wattage HPS (HPS need normally even higher ratio OCV/Varc to keep the arc stable during lamp aging when the Varc increase).
Other problem is, then the turn ON transient of the microwave transformer could cause rectification for few mains periods (when the transformer saturate as a result to not optimal turn ON point phase on the mains sinewave) - cause low frequency shift of the mains for few cycles. I'have seen this problem being caused by burned out "soft start" resistor in the microwave. The high impedance of the mains caused the fuses didn't burn out, but the mains collapsed. Replacing the resistor and the relay (it was the sometimes weak relay contact of the bypass relay, what burned the resistor) in the microwave helped a lot.

I seen, as a kid, system with HPS poles installed in a row (assume 3-phase 240/415V system, standard 100-150W HPS with plain inductor ballast, superimposed ignitor, and PFC capacitor)

The system is fed from a long underground cable that goes from each pole to the next. The main cable is not interrupted in the poles by a breaker etc - only the branch of each pole is

When a blackout (power failure) happened, the fixtures clearly appeared to go out in a fast "rush" from the point of supply and along the line (as in with delay between each fixture and the next) rather than all at the same time



I dont think i can see the electrical field disappearing at the speed of light, and i dont think that parasitic inductance or capacitance of the main cable is suffcient to cause such visible delay

What could explain that ?

When a blackout (power failure) happened, the fixtures clearly appeared to go out in a fast "rush" from the point of supply and along the line (as in with delay between each fixture and the next) rather than all at the same time

I dont think i can see the electrical field disappearing at the speed of light, and i dont think that parasitic inductance or capacitance of the main cable is suffcient to cause such visible delay

What could explain that ?

I think it is only an illusion - the human brain is process the most urgent potential thread (so the closest event) as first, what mean what happen closer get the priority, so is "registered" faster.

A CWA ballast, is less sensitive to voltage dips (when a computer is turned on, the lamp doesn't blink), but on a HX, it would.

What is the qality of power supply in that computer ? Most power supplies have at least an NTC in series with the input to lower this inrush current

@ Ash:
It is a standard laptop adaptor.

Those are the worst in my experience (also the inline power supply for my battery charger!) - when you plug them in you get quite a spark unless you shove the plug in quick!


BG

With some sockets this makes a problem - the contacts of the live and neutral in some power strips are too close inside. This spark, when it happens, sometimes jumps to be across the live and neutral and trip the breaker

I do try to insert the plug straight with them. But the ends of the pins on my notebook's power supply are all arced

I've not had that happen but the metal strips are indeed close in some of them, probably more so in the straight strips rather than the angled ones, although the straight one (4xCEE-7) I am currently using in the kitchen is better in that you can use a combination of plugs/wall warts, usually a problem with the angled ones but partly no doubt as a result of proper spacing between the sockets. Many strips hae their sockets too close together, especially BS1363 ones. It also has a switch, as do several of my other strips, which avoids the arcing problem.

My vacuum cleaner however would probably result in a near EOL lamp on the same circuit cycling as that makes the lights noticeably dip at switch-on!


BG

Some old wall sockets here (the better ones) had a tube enclosing the live and neutral contacts, so that each of the arcs is enclosed from the other strips