I’m getting 7 pcs of 90w sox lamps and i dont have ignitor for these. I only need one but i was thinking that would there be anything else way to start those lamps?  Here in Finland all sox light has been removed 10-15 years ago so it very hard to get gear for those.

Some have had success with the short-it-out method. Using a switch (or a very durable fluorescent starter) just short out the ballast momentarily to get the inductive flyback pulse, that should strike it. With the proper ballast of course.

Would 80w mercury ballast work? (230v) if doesnt then i might be bit screwd.

Ehh, I would say good enough for short term. My semi-accurate calculations predict ~16% underdrive. I am not familiar with how LPS tolerates underdrive so I have no idea if this is acceptable or not. Also the lamp might not stay struck, as 230VOC is kinda low for a 112V arc lamp, especially with something as high power factor as LPS.

If the ballast has different voltage taps on the input (220, 230, 240), use a lower one than normal to maybe get a little more current out of it. I don't think you will severely damage the lamp, but I don't guarantee it will actually work.

Edit: Nevermind, I just looked it up and it appears that ~230V choke ballasts do exist for LPS. It should stay struck just fine, just probably underdriven.

I made Some research and discovered that 90w lps runs on 112v and 80w me runs in 115v but a bit lower current than 90w Sox. 18w Sox start up very well with s10 starter and 18w t8 ballast.

Yup, so you should be good on from the OCV perspective, I am sure someone else can chime in later about the potential 16% underdrive and whether or not that is gonna cause a problem.

Thank you! I hope that somebody else would know something more.

I am sure someone else can chime in later about the potential 16% underdrive and whether or not that is gonna cause a problem.

A 16% underdrive will work for a short test, the lamp, run up fine, but if you plan to use the lamp, long term you, should stick to a less than 10% under or overdrive for your SOX lamp, or really any SOX lamp for that matter.

A useful way to finely adjust lamp current in 230V realm is to hook some low power fluorescent chokes (like 8W T5 or 11W PL which are almost equal) in parallel to main choke. If still not enough current, hook another one or upgrade say, to 30W choke

As for ignitor, 230V starter may work alone (or better in series with 100W lightbulb to make short circuit current reasonable). Or may not work with some hesitant LPS lamps which really need a dense series if ignition pulses to do glow-to-arc. In that case, assuming you have a pile of like LPS lamps and do not especially care for one of them, I'd just whack the lamp with some superimposed HID (MH/HPS) ignitor. Ignitor voltage will be somewhat high but outer volume of the lamp won't break down as it is hard vacuum. There is a remote chance of base flash-over, you have to look, but I believe ignitor voltage will be immediately clamped by arctube breakdown and lamp will light-up just OK.

Also, there are really heavy duty electronic ballasts having circa 800mA current, one probably will take a good care of such lamp, but these are quite rare, intended for equally rare high power CFLs.

I would try SON ignitor to a wire wound externally around the lamp and not connected to anything

The wire can be in insulation

Superimposed ignitor can be wired from the same ballast which is powering the lamp, just connect the lamp before the ignitor so it does not get the pulse to the main terminals. This will still let the ignitor sense the lamp arc voltage, and stop igniting once the lamp fires

Ballast dependent ignitor (and possibly a small X capacitor) may be required if you figure you need a more powerful pulse. In this case use whatever ballast that provides the correct tap for the ignitor. I think you might still try to connect the entire 2nd ballast setup from the output of the 1st ballast (parallel to the lamp), to get the ignitor to see the lamp arc voltage, but there may be more things that might or might not work correctly

The lamp is internally coated with a conductive layer (ITO), which would act as a large capacitor plate, standing between the wire and the arctube. This is ok as long as you can couple enough pulse to this layer, so the entire layer gets to HV potential relative to the arctube

The environment (including metal luminaire parts, and in an open bench test - just the more distant general environment) is also a capacitor plate - a grounded one, which is undesirable, as it makes a 2nd capacitor from the lamp to ground, forming a divider with the capacitor made by the wire. You can increase the divider output by winding more wire

@Ash
That is a very interesting idea! Is it possible that (long term) the wire around the lamp could attract sodium to the surface of the glass and limit lifespan? Or is it not enough surface area to do that?

The arctube sees the conductive ITO layer inside the lamp, not the wire directly. The ITO is coupled to the wire by capacitance, ie. it is coupled for AC (including the pulse) but not for DC

The wire will be at the same potential as the L cathode of the lamp (except the ignition pulse). So AC voltage of 0 between this cathode and the ITO, and AC voltage of the full lamp arc voltage (some 120V) between the Neutral cathode and ITO

The ITO may be charged by DC from any effects already taking place within the lamp. I do recall reading something about it here (James ?). The AC coupling would at one half cycle add and in the other half cycle subtract from DC effects, the overall effect would be either zero (expectations...) or maybe something else depending on the mechanism of some quirky effect

Finally, at any moment the ITO will either attract or repel Sodium ions, depending on the polarity of the momentary voltage (sum of DC + this moment in the AC) between it and the region in the arctube from which the sodium would be drawn. The DC voltage (and how it compares to the peak of the AC) will determine if the sodium would be repelled all the time (best case), attracted all the time (worst case), or attracted and repelled alternatingly throughout the cycle

I do recall reading here about a failure mechanism in SOX lamps, where the getter flashing was a bit out of size and place, and shorted one of the main terminals (where they come out of the stem) to the ITO, which have led to fast destruction of the lamp. In this case, there is no capacitive coupling to the ITO, and no DC effects can take place. It would be at hard 120VAC relative to the other (not shorted) terminal, so half the time pulling sodium out and half of the time repelling it (which does not undo the damage from the 1st half)