In looking at Lithonia's Metal Halide 480 VAC Ballast Data the starting current ranges from .4 to .85 amps depending on the ballast with an operating current of 1 amp. I will have to look at the above paper mentioned by Joe to see if there is any difference with an in-rush figure as in-rush could be significantly different and you probably already know the figures I gave above.
Re: H.I.D.lighting start up current inrush#128450 03/21/0305:01 AM03/21/0305:01 AM
Transient Inrush on most Reactive devices like these Ballasts would range from 400% of rated high-end current, to as much as 1000% of high-end current. Time frame would be maybe for only 10-20 cycles for highest draw, tapering down to like 200-300% by the time a full 60 cycles has passed. I need to check into this and verify data, but it "seems" this would be accurate.
The use of Inverse Time Trip Breakers or Time Delay Fuses is important here. MCCB with the notation of "HID" usage have this characteristic, plus offer "Switch" duty and designed to deal with quenching a more stable arc, from opening HID Ballasts' circuits during operation. I'll paste some data regarding HID duty MCCBs later.
Basically, the inrush should not normally trip a properly designed circuit - unless there is component failure or some Ballasts are Resonating the circuit big time (Capacitors mix-matched or incorrectly connected to Reactor Core/Coils, causing hard starts), or there are simply way too many Ballasts on a circuit.
Keep circuits from being max'd out, plus compensate for long circuit runs by upsizing conductors.
Another concern is bouncing harmonics around a 4 wire wye multiwire circuit, where ØA, ØB and ØC have an odd number of different type HID Ballasts. Could result in very wierd initial inrush levels.
Let me verify all this stuff first, then get back with conclusions.
If you are looking for Highest Starting Current / Highest Load Current for the Ballast, this is totally different than what has been listed above!
Ballast for a normal "Probe-Start" M59 Lamp (400 watt Metal Halide Probe Start Lamp) would be a common Constant Wattage Autotransformer configuration (CWA). These have an almost unchanged input line current level from when the arc is first struck across the tube (starting), to driving a stable lamp's full output (running).
Input KVA is adjusted to fit per available voltage levels (within a small tolerance, usually within 10% +/-).
For an ANSI M59 driven by a properly connected and functioning CWA Ballast, the high line current at 240 VAC would be 2.0 Amperes - unless the circuit's power factor is offset (shorted capacitor), then would be more in the 4.0 or higher range!
Properly functioning Ballast would have an apx. power factor of 0.8334 (83.34%), or simply one Ampere per 100 watts, per 120 VAC