While looking at a document from Universal Lighting Technologies that had some listings for some probe start metal halide ballasts, I have apparently discovered that some of their F-Can ballasts require a pair of them to be connected in parallel to be able to properly operate a lamp.

For example, I have found out that you would need a pair of 1111-246C-TC ballasts connected in parallel to be able to run a 250W M58 probe start metal halide lamp or a pair of 1111-247SC-TC ballasts connected in parallel to be able to run a 400W M59 probe start metal halide lamp.

You can find the document here:

https://www.candelacorp.com/spec/?prod_num=M1650MLHAC5M215T

Additionally, I just found a couple of listings for the 1111-247SC-TC ballasts on eBay:

https://www.ebay.com/itm/232754466335

https://www.ebay.com/itm/113763485009

Wit this in mind, it seems like a single 1111-247SC-TC ballast may be a close match for a Japanese 200W mercury vapor lamp (120V 1.9A), but the ballast is probably best for a 135V 1.65A discharge lamp such as a Thorn 200W SLI/H HO low pressure sodium lamp (145V 1.55A) since a 400W H33 mercury vapor lamp or 400W M59 probe start metal halide lamp runs at roughly 135V 3.3A.

Additionally, it also seems like a single 1111-246C-TC ballast might be a close match for a European 125W mercury vapor lamp or a North American 125W H42 mercury vapor lamp (125V 1.15A), but the ideal discharge lamp for this ballast would run at 135V 1.05A because a 250W H37 mercury vapor lamp or 250W M58 probe start metal halide lamp would run at 135V 2.1A.

I wonder why Universal Lighting would go out of their way to design these half power F-Can ballasts meant to be connected in parallel even though 1 ballast is normally enough.

Why two ballasts?
Don't they also mention a "dimming" possibility with the arrangement? Like to allow "Full and Reduced output" modes by powering just one section or both?
Other motivation is to better fit around the main lamp or pole attachment assembly into the fixture in a way the whole thing still remains physically balanced (so it won't tend to twist the lantern around the pole arm tube).
Or to maintain slim format, so the ballast can fit into a given diameter pole/arm tube (two ballasts are occupying more length, but that is not that much of a problem, ballast not able to fit into a frequently used tube diameter would be a marketing problem)...

I just find it interesting because I have seen that many members here had mixed opinions regarding the safety of connecting multiple autotransformer ballasts in parallel. I recall some saying that it is safe to connect multiple autotransformer ballasts in parallel as long as they are identical while others said that it is dangerous due to phasing issues stemming from differences in OCV between individual ballasts.

 2 ballasts make no sense in that all it does is make the fixture bigger and add more weight ! Why make such a thing when F-can ballasts are readily available that will run the same lamps with only one ballast ?

I was also having the same impression about that interesting ballast oddity I found.

 I get needing to combine multiple ballasts to replace a ballast that doesn't exist anymore but the ones you listed just don't make sense to me .

I wonder if was made in order to have a 400w F-can format but keep the ballast a reasonable length. Single ballast 400w MH F-can units do exist but are comically long due compared to other F-can HID ballasts. These meanwhile are closer to the size of a normal F-can HID ballast. They look to be around the same size as a 250W MH F-can ballast. Wouldn't be surprised if it was just one of those with the capacitor value slightly modified and the difference label.

I'd imagine these would work better in something like a 2X2 recessed MH fixture, maybe with one on each side.

I just find it interesting because I have seen that many members here had mixed opinions regarding the safety of connecting multiple autotransformer ballasts in parallel. I recall some saying that it is safe to connect multiple autotransformer ballasts in parallel as long as they are identical while others said that it is dangerous due to phasing issues stemming from differences in OCV between individual ballasts.

If the ballasts are made using the same "blueprint", there can not be any OV differences - OCV is ratio of turns and turns are whole numbers, so if the same, they are really exactly the same.
And even when there are some small differences (e.g. when mixing different model years, when the exact design parameters were evolving), it has no impact unless the phase and/or OCV's are really far off, so practically only when you are mixing ballasts of completely different kind.
So once you are sure what is the exact design of the ballasts to be combined, and I would expect the manufacturer will have thios info, there is no issue at all.