Hi, good day everyone

I received my 55w Philips low pressure sodium lamp along with the ballast, capacitor and the ignitor, looking at the wiring diagram on the ballast I am not sure where should I hook up the capacitor..

I just simply hook up the capacitor and managed to fire it up and the third photo shows how I attach the wire and not sure if it is correct, I have had it run for 5 minutes, it starts with faint pink glow, and within 3 minutes it achieve full brightness, and thank god it didn't blow up . I like to make sure I did not under / overdriving the lamp, also does these thing will not work without a capacitor?

I am no electrician btw.

The capacitor in choke (Non autotransformer like autoleak transformers) based systems, is only used for power factor correction, and don't affect lamp performance, so you can give up with it, and connect the lamp without it and it still light at full performance even if the power factor would be low. LPS lamps takes a good 9-15 mins to reach full brightness.

The capacitor does not go there, it goes straight between points L and N, parallel to the line voltage. It would also work correctly without the capacitor at all, but draw higher current from the line

Placing the capacitor in series with the lamp would overdrive the lamp (it depends on the ballast impedance and capacitor value, but in this case i think it would). If the lamp did not have strange effects and the electrodes are not sputtered, i guess the overdriving from this one time was not too extreme and did not cause noticable damage to the lamp

Definitely the capacitor, if used, should be parallel to the mains, just as a power factor correction (as both Dor and Ash have already mentioned)
The current was likely limited by the coil saturation, so most likely the lamp had survived it.
But definitely the system was overloaded:
- Mainly the capacitor have seen large overvoltage (I would guess nearly about double the mains voltage)
- The rms current was for sure way higher than rated, so after longer time the ballast will overheat
- The current most likely had high crest factor, overloading the lamp electrodes. But hope the time was short enough to not really cause any extensive damage.

Thanks guys for giving me the ideas and being very specific about everything, silly me .. the capacitor is currently ignored, now the warm-up takes around 10 minutes, the pink glow seems to stay there longer, as mentioned by dor123: 9-15 minutes to reach full-brightness, seems about right. Now everything feels normal and I've let it run for 30 minutes and the ballast only getting a bit warm this time . Unlike yesterday where I feed the capacitor to the lamp, the ballast was burning LOL..

Damn, watching these thing warming-up up close is amazing, at full brightness the deep orange light sort of bring back old memories for me as I haven't seen these street lights for maybe 15 years as they were all replaced by hps..

One more question though, the Philips SX72 ignitor unit looks like a electronic version of fluorescent lamp starter, can I use a S10 T8 fluorescent lamp starter instead and will it make any different? because I've seen some mercury lamp with built in starter / ignitor in the middle of the bulb, and the starter / ignitor looked exactly like a starter from fluorescent lamp, hate electronic stuff

First I guess it were not mercury, but most likely HPS, where that is quite a common method for a selfstarting HPS lamp (MV's use an auxiliary small electrode connected via a resistor to the opposite electrode).

The starter is there in series with a resistor to limit the current and most important a bimetal switch, which connect it only when everything is cold enough. The reason for the bimetal is, when the arctube is too hot, it won§t be possible to restrike. If the starter will be connected at that time, it will constantly attempt to restriking the lamp and because that will take too long, it will get destroyed (it's life is limited in the total number of clicks). In such case the bimetal switch keeps the starter disconnected until it cools down. When it is cold, the lamp ignites for the first or second pulse, so the starter wear is then very limited (normally when striking the lamp for 5 hours, assuming 25khour life that means 10k clicks; when it will be clicking constantly, this amount of clickings will be reached after just few hours of cooling, so few 10's of hot restrike cycles).

But Im not sure, if that will work with SOX.
The point is, the lamp has to establish an arc with cold electrodes and it takes about a second or so for the electrodes heat up. And when they are cold, there is about 100..200V extra voltage drop on the border between the cathode and plasma (called cathode fall). Only the cathode heats up, this voltage drop gets reduced to just about 10..15V (the intention is to make it as low as possible for the normal operation, because it dissipates power, but does not generate any light)
The reason for the HPS vs SOX difference is, the HPS needs the higher voltage pulse really only for the first ignition, but when it is cold (when the starter is connected by the bimetal), the arc column voltage drop is low, so the mains voltage suffices to feed the discharge even with cold electrodes.
The SOX has nearly the same anode column voltage drop already when cold, so the extra cathode drop means the mains may not be sufficient to sustain the discharge for long enough so the electrodes warm up and so lower the cathode fall before the arc dies completely.
Some SOX have low enough drop (mainly the low wattages), so it will work.
But the longer ones need the ballast to provide higher voltage for that period. And that is, what the electronic ignitor provides: It generates extra HF voltage (few kHz or so) around the 230V, what means the peak voltage gets way higher, so my support the cold electrode discharge. When the electrodes warm up, the voltage drop disappears and so the ignitor shuts off the HF generation and becomes "inert".
And with the starter the problem is, it is able to generate just a single pulse after quite long time (no more than 2 or 3 pulses per second). That may provide the initial breakdown, but it won't provide the support needed for the cold electrode state (for the first second after the initial ignition).

So you may try it with the starter, but don't expect it to work reliably.

I want to add: Never operate low pressure sodium lamps with the base down. This can cause accumulation of the liquid sodium at the end seals of the inner tube, and distroy it by corrosion. Operate it when it mounted horizontally. You can also operate it with the base up, but never ever with the base down.

Do those electronic SOX ignitors really switch multiple times in a line half cycle ?

Do those electronic SOX ignitors really switch multiple times in a line half cycle ?

They start to oscillate at few kHz...

For what actually ? Breaking the circuit once in each half cycle not enough ?

Not with longer lamps.
The thing is, the lamp has to start with cold electrodes, so there is extra 100..200V drop across the lamp just when you first ionize it. And even with that the ballast should be able to feed enough current, so the cathode fall warms up the electrodes and so allow the lamp to transition into the normal operating levels. Without that the lamp will just be flashing on the ignitor pulses and that will be all, it will never turn into an arc discharge.
The same happens with longer fluorescents: Try to connect a HPS ignitor behind a standard 36W ballast and power the 36W lamp with it. You will see it will just dimly glow form the ignitor and nothing more. And once you connect e.g. a 9V battery to one of the filaments, the lamp starts - with the battery you heat up the electrode externally (at least one), so at least in one direction the current starts to flow, heat the other filament by just an anode dissipation and that then turns the lamp into  a normal mode.
But when you connect the same lamp on an electronic ballast so the filaments are shorted out (or one wire on each end not connected in case the ballast has a voltage mode heating, so it still can not warm them), the lamp will start. That is, the resonance generates way higher voltage than available with mains, that voltage would be enough to pass sufficient current even with the extra cold cathode fall and so then warm them up.


With HID's the simple pulser works, because when the lamp is cold, there is nearly no drop in the gas itself (it is short and of a low pressure), so the 230V mains is still sufficient to operate the discharge even with the 200V extra (at least from the ignition pulse till the end of the half cycle; when that repeats for each mains half cycle, there is quite a lot of heat for the electrodes to warm them up).
Once the electrodes warm, the arc voltage drops to the 30..40V typical for the cold HID's and the really high current arc starts (usually takes about 1 second or so). As the electrodes warm up a bit assymetrically, the lamp rectifies quite strongly over few cycles, that is then the "thud" noise from the ballast...

Are the igniters for HPS lamps like an autotransformer? Just in circuit for a few cycles also do HPS lamps on HF ballasts need a starter?

HPS is a high pressure lamp, so with rather short distance between the electrodes. That means when cold, its arc voltage is way lower than when running warmed up. So that means the 200V mains and above (so at least 280V peak) is way sufficient to maintain the cold cathode discharge, once the tube is ionized. That means you need just a single pulse at a right moment to successfully start the lamp. But that pulse should be capable to break down the gas, so needs high voltage or some other ionization aids (internal or external electrodes, UV flasher bulbs,...) for that.

On the other hand the low pressure bulbs are easier to ionize, but as the pressure does not change that much as they warm up, the voltage drop does not do so either. So to get reasonable voltage drop across the arc for the normal operation (so the 10..15V of non luminous cathode fall is not that much of losses), you need the arc to be long. And that means way higher voltage is needed for the glow discharge, where the 230V becomes insufficient.

So the HPS or MH ignitors ae usually a capacitor charged to about 200V and then discharged into a primary of a pulse transformer. The pulse transformer then steps up the 200V to the required final 2..5kV pulse. In the US it is common to use a tap on the ballast winding, so use the ballast coil in the role of the step up pulse transformer, but otherwise the function is the same as with the European superimposed ignitors.
When the lamp has some auxiliary electrode (an antenna close to the arc tube,...), voltages as low as 1kV may be enough and so then just the inductive kick from a fluorescent-like starter (inside of the selfstarting lamps) is enough to provide that initial ionization.
In both cases the pulse is quite narrow, does not carry that much of energy, but mainly there are no more than just a few pulses per mains cycle. That is too low to do anything more than just the initial ionization.

The electronic ballasts often use another technique to generate the high voltage: Because all components are already there, it is not that much difficult to operate the output stage in a such way, the HV is generated by means of a resonance on the output filter. As first this high frequency is capable to easily overcome all resistive leakages along the arctube (those tend to equalize the field, so increase the breakdown voltage; so with HF the breakdown voltage becomes lower), so that method is frequently used there.

But the resonance has another feature: It may deliver rather high power levels, so may be used as well to aid the glow phase, so suffice with way lower native ballast OCV. This is used in many modern lower power HID ballasts, allowing the intrinsic OCV of the ballast to be way lower, so just a two transistor halfbridge becomes sufficient (that is not able to generate more than about half of the DC voltage it gets, so about 200V, already includes the PFC stage voltage boosting).
These then operate the lamp at HF and so rely on the resonance till the electrodes warm up and the arc voltage drops, only then switch over to the LFAC operation (where the native OCV must be then higher than the arc voltage).


The ballasts for all the low pressure lamps then have to either provide separate electrode heating for the start (all fluorescent preheat, RS,...), provide sufficient OCV to support even the glow discharge (the traditional LPS ballasts without any ignitor). Or they have to feature similar means as the HID ballasts I've mentioned as the last one: Utilize some form of resonance to generate the higher voltage necessary for the cold cathode operation.
Because just plain capacitor parallel to the lamp is not possible on 50/60Hz (discharges can tolerate only small capacitances), the SOX ignitors use high frequency to generate the elevated voltage. It isn't anything complex, the whole circuit is just a spark-gap like oscillator, only made using triacs instead of a real spark gap. This then generate voltages around 1kVpk with output power in 10's of % from the rated lamp wattage, so providing quite substantial cathode fall dissipation to warm up the electrodes.

You were right, it is indeed a HPS with a starter inside, I'm going to stick with the BSX ignitor for the sox then. @dor123, thanks for the advise, I have a lot to learn, I'll be sure not to use the lamp base down from now on. lol

Also btw, I found some old parts, it's some unused 36w ballasts from t8 fluorescent fixtures, I am wondering can I use the 36w ballast to power 2x philips 18w PL 2-pin compact fluorescent tubes? the PL tubes has built in starter..

You don't need a ressistor for starting LPS lamp with a fluorescent starter. Also: 55W SOX lamp will restrike instantly when hot with the glow starter. These are usually the high wattage SOX lamps, when used on an autoleak transformer, which needs to cool before they can restrike. With starter or electronic ignitor, the pulse rate is high enough to instant restrike them as well.
Regarding to my experience with restrike of HPS lamps with internal starter: The generic ones don't have a bi-metal, as I've seen here at Qiryat Ata in Schreder Alura postops. With Osram NAV-E/I 70W, when the bi-metal of the starter is disconnected, these lamps often manages to restrike directly from plain 230V after 2 mins average, when the bi-metal is still open. With cycling Philips SON 70W, their bi-metal opens after 4:30 mins after the lamp extingushed, and the lamp glows a mercury color and flickering like they doing during cold starting, until they changes into an orange color.