@James

Wow, that is a lot of good info! I think that is very likely the case, since there is still some extent of red color on the edges of the seal. Very very interesting, thanks!

But I still can not believe magnetic ballast is that bad to consume 78W when driving 35W lamp. That means efficiency is below 50% which is crazy!

Something is wrong with the first photo. 18W lamp + ballast combo can not consume 78W of active power!

that's because i was using one of my 35W SOX lamps with a 35W Glow Object ballast as i noted. i don't have the 18W Glow Object ballast yet.

Something is wrong with the first photo. 18W lamp + ballast combo can not consume 78W of active power!

update: i received a 35W Glow Object ballast in the mail yesterday, and promptly hooked it up to a Signal Transformer 300VA step-up transformer i bought off Digi-Key, and one of my 35W SOX lamps.

and... it works GREAT!   

some details i noticed:

• the ballast is dead quiet. even with my ear right next to it, i couldn't hear any 60Hz hum
• the initial strike of the lamp is much "cleaner" in a sense... there isn't a noticeable blue flash as the strike occurs like with my magnetic SOX ballasts
• the efficiency and current draw are significantly improved. with my 35W SOX magnetic ballasts, i would never get below about 77W total (0.83-or-so power factor) due to ballast losses, but this Glow Object ballast draws just 39W total and has an average 0.95 power factor
• this doesn't really affect the functionality, but just an added bonus - the Glow Object ballast is slim and lightweight as compared to the typical magnetic ballast

so far, just from my initial test run, it's a winner. i'm very pleased with its performance.

i've attached some power meter readings from a magnetic ballast and the Glow Object ballast for everyone to see.

i'll definitely be performing more tests in the future, and i have an 18W ballast on the way from Glow Object as well to run my SOX-E 18W lamps.

This is a very good observation and question.  I do not know the exact answer but have a theory.

In the Chinese lamps I have analysed their seals still make use of the same 42% niron core with copper sheath and top coating of sodium borate.  The glass types made in China do however tend to be slightly different than the long-established compositions produced by the Western lampmakers, during the era when pretty much each lamp company also made its own glass.  That era has almost finished, and Chinese lamps seem to be made with standard commercial glass types rather than the unique compositions developed by the lampmakers for optimum compatibility with their high speed production machinery.  One of the main differences of Chinese glasses is their elimination of heavy metals such as lead and barium oxides, for environmental reasons.  That results in slightly harder glasses with higher softening and annealing temperatures.

When a Dumet seal is first made, the colour is of black copper oxide.  During prolonged heating of the seal-making process that gradually dissolves into the glass, followed by the red coating of sodium borate, and if heating is maintained long enough (or hot enough), that will eventually also disappear into the glass to leave a brilliant metallic copper interface.  According to old GE and Philips lamp training manuals, the optimum seals are made just before the last of the red borate dissolves into the glass.  However, working so close to that limit makes the manufacturing process controls very critical.  Too much heating, and the complete loss of an oxide interface between the metal and the wire results in increased risk of leakage.  Especially since the Western companies typically ran their production on high speed lampmaking machines, it was too difficult and critical to make seals in which the oxide coating had been dissolved as much as would be ideal.  The risk was too high that some seals may become overheated, which then very quickly begins to impair seal quality and increases the risk of gas leakage.  Hence the compromise was to aim for a salmon colour seal, in which some red borate was still present but the brassy colour metal just began to be visible.

In China, due to low labour costs there was so far no need to invest in high speed machinery : the high capital investment could not be recovered by reduced production costs.  So the typical Chinese lampmaking lines run very slowly.  With such a process it's easier to better regulate the heating of the glass, and dare to keep the seals hotter for longer.  The slightly higher working temperature of their glass also means that during the annealing phase, there is naturally more dissolution of the borate into the glass.  I believe these two factors, when combined, account for the fact that Chinese dumet seals show significantly different colours than in the West.

I've noticed that in some HPS lamps like my Osram NAV-TS 70W Super 4Y, the sodium vaporizes from the middle of the arctube instead from the ends, where the sodium amalgam usually condenses.
See my video
Why this is happening?

@James, @RRK, @Medved?

In Canada, it depends on the province. Some provinces have fairly mild summers and so don't have a lot of homes and apartments equipped with AC while in other provinces with fairly hot summers, AC is pretty common. In my city, the summers are pretty hot and humid so AC is fairly prevalent.

To be honest, if I lived in a home or apartment without AC, I'd install one, even a small window unit that cools only the bedroom helps a lot in the summer.

Older homes and apartments were typically heated with hot water radiators or electric baseboard heaters. Those typically use window units or mini splits for AC retrofits. In newer homes with forced air ducts, central AC systems are what's mostly used.

For newer apartments, what's typically fitted is a system with a central chiller and boiler and a single fan-coil unit in every apartment feeding some ductwork to distribute the air throughout the apartment, similar to what's used in larger office buildings and malls. Mini-splits aren't that common in apartments.

Older apartments usually use window units, either retrofitted into the window or placed in a cutout under the window in older apartments that were built with AC. A similar setup is also used in hotels, although larger hotels might use fan-coils.

Smaller office buildings and warehouses typically use rooftop mounted AC units that are fed with ductwork into the building. AC in a warehouse or factory isn't that common unless required for the goods being stored or the production process due to costs.

I agree central A/C is almost of no use in Europe (and Asia, and Russia too) for private homes or flats. More of American thing. I presume because of huge cost and complexity of installation. Also the service like cleaning the ducts looks complicated. And almost impossible to add it as a retrofit to an existing home, until one wants to go into some really serious construction. Simplified variants are practiced sometimes, like at my friend's flat, where indoor unit is placed under the false ceiling and connected with ducts but that is rare. Central A/C is often (but not always) practiced in some newly built commercial buildings though.

Overall, almost all A/C here is split systems with individual indoor units fit to the walls or in the ceiling. Most usually driven by own outdoor boxes, but sometimes in multi-split configuration. Window boxes type A/C is all gone out of style.

It is absolutely not true that A/C is overall rare in Europe.