When I was studying North Ameican HID lamps, I have found that only the low wattage high pressure sodium lamps and 6.6a series system mercury vapor lamps were designed with a simple series choke ballast in mind if it was to be operated off 120v mains without an autotransformer. Why were all other HID lamps for the North American market not designed for a 120v OCV in mind considering that 120v is the most common mains supply in North American countries and that HID lighting systems would be more efficient and less lossy when an autotransformer is eliminated from the ballast circuit?

There are several reasons, a few of the most important you find below.  Firstly, it is because the first HID lamp was developed in Germany and introduced in England (the MA mercury lamp, 1932), where the mains voltage at the time was 220-240V.  In 1934 GE copied it and introduced thAt identical HID technology to USA.

Experiments had already been made in UK and EU to adjust the design for low voltage countries, and it was found that this led to a severe reduction in lamp performance.  For any HID lamp when you halve the volt drop, you double the current.  Since many of the failure mechanisms in HID lamps are related to current density, this leads to shorter life.  This is principally due to electrode life.  The electrodes must be made larger to handle the higher current, and they then take longer to heat up after switch-on, which leads to increased damage during the phase of run-up.  As you know, the life of an HID lamp is very strongly dependent on the number of switchings, and that becomes progressively more severe as arc voltage is reduced.

There was also a second problem relating to arc tube design.  All other things being equal, if you halve the arc voltage while keeping constant power then you must approximately halve the arc tube length.  If that is possible, then it also results in a major drop in efficacy since the relative length of the light-generating positive column discharge is reduced, vs the length of the cathode fall area where relatively little light is generated but a lot of energy is dissipated.  In practice, even this was not possible because an arc of same power over half the length makes the arc tube wall run far hotter, beyond what the arc tube materials could withstand.  This could be partly offset by enlarging the diameter of the arc tube, but since the arc is arc-shaped and bows upwards this doesn’t help in case of horizontal burning lamps because the arc just bows up even more and still overheats the glass.

Over the past century it has been found many times that for almost all HID and fluorescent lamps, the optimum conditions to maximise lamp life, performance and efficacy all exist when the mains voltage is around 200-250V.  The advantages for this design are so significant that the in Americas it was more efficient to copy exactly the European lamps and ballasts and to run them on a step-up autotransformer to first boost the mains voltage - that device later being integrated into the ballast to form the autotransformer ballast.  The efficacy losses caused by such ballasts are quite large, but less than the losses that would be incurred by redesigning mercury and metal halide lamps for lower arc voltage, for the types where that is possible.

For HPS lamps the situation is different because they have wall-stabilised arcs, and the electrode degradation effects are somewhat less significant due to the xenon fill.  The relative losses by the ballast are always higher for the lower power lamps.  So exclusively for the small HPS lamps it was decided to make the 52-55V ranges for Americas, so as to simplify the cost and size of the ballasts.  However the performance and life of those types is always inferior to the high voltage types.

@ James

I also wonder why North American high pressure sodium lamps rated to run beyond 150w were not designed with 55v arc tubes fot operation on 120v chokes?




<Admin edit: Unnecessary quoting of long previous post removed>

@ James

I also wonder why North American high pressure sodium lamps rated to run beyond 150w were not designed with 55v arc tubes fot operation on 120v chokes?




<Admin edit: Unnecessary quoting of long previous post removed>

The factors described above still do play their roles with HPS, although it is not as bad as with other chemistries, it is still there.
In fact with every chemistry, there is an optimum arc voltage for a given light output (and/or power rating), generally higher power having its optimum at a higher arc voltage. And with every one, the performance degrades if you depart from that optimum point. With MH the degradation would be very steep, so therefore you would not find lamps with arc voltages below 70..80V or so.
The HPS is not that steep (because of the low operating temperature of the arc), plus for the lower wattages the optimum voltage is not that far from the 55V or so, the maximum what the 120V OCV would allow. Because of the low power, the equipment cost is mainly dictated by its complexity (so there is a significant cost difference between transformer vs simple choke ballast), yet the energy cost related to the lower efficacy is not that large due to the low power involved, so it made sense to sacrifice the efficacy in order to make the system simpler and cheaper.
But with higher wattages the energy cost becomes the biggest contributor to the total cost of the light, so it becames way cheaper to pay for a more complex (transformer) ballast upfront so the lamp could be designed close to its optimum, so savings on the energy costs over the system life become way greater then the purchase cost difference. Because (mainly) the HPS use was driven exclusively by the desire to get the light as cheap as possible (when even the light quality did not matter), there was simply no market for anything except getting the amount of light required over the target area the cheapest way possible.
So for the 120V market, with the low power packages that mean using the 55V arcs and just a series choke, 150W is about the same either way, for higher power it means to optimize the arc voltages to the last point and use dedicated CWA ballasts.
For 230V areas that means designing all for the 70..90V arc voltage range, as the extra losses related with the transformer would never offset the gain from the 10..20V "distance" from where the optimum would be (for tge higher power versions), so all settled to work with just a series choke.