I've some evidences that internal starter and penning start HPS lamps, often hot restrike directly from the mains, when the bi-metal that disconnects the starting coil (In penning start HPS lamps) and/or the glow starter (In internal ignitor HPS lamp), is actually still open.
1. One member on this site, have a youtube video , of an operation of an Iwasaki heat start HPS retrofit for MV lamp, that uses a thermal starter that rely on tungsten filaments instead of nichrome heater to open and close. The hot restrike part of the video, looks like the lamp strikes by itself, without any feedback from the starter/filaments (As if it strikes without any starting aid, directly from the mains). This video posted in LG, somewhere I don't remember, and some members said that HPS lamps can strike directly from the mains without an ignitor or a starting aid, when the lamp is in the LPS color mode.
2. When I played with a Gaash Mar 10 lantern in the electricians house of kibbutz Givat Hayim Icud almost tens years ago, I operated an Osram NAV-I 70W HPS lamp with an internal glow starter. During the first hot restrike I did, the starter operated all the time, and hot restriked the lamp. During the second hot restrike I did, the lamp was out, and the starter wasn't kicked in (I didn't knew at that time, that the glow starter is disconnected by a bi-metal). After several minutes of a total darkness, suddenly a ping was heard, and the lamp kicked on without any light from the glow starter first (As if the arctube hot restruck directly from the mains).
Is this possible that most penning start and internal starter HPS lamps, actually restrikes directly from the mains, when the bi-metal of their starting coils/glow starters, remains open?

The ignition voltage versus temperature is a kind of tricky business.
When too hot, the high density gas of evaporated sodium prevent the ignition.
When cooling down, the vapor density drops (as the sodium condense), so does the striking voltage. It may reach so low level, the 230V mains become sufficient for the lamp to ignite. So the question may arise "Why to bother with the flimsy starter?"
Well, when cooling further down, it reaches a point, when there is nothing to condense anymore (all the mercury and sodium is liquid and/or solid). At this point the ignition voltage is the lowest.
But when you cool down further, the atom's velocity in the gas drop, so the collisions are not as energetic. As these collisions help the electrons to jump out of the atoms and turn them into plasma ions, when their energy drop, it have to be replaced by higher electrical field, so higher voltage. And in this area, colder the lamp, higher the striking voltage.
Now as we want the system to properly start even in the middle of freezing winter, the ballast have to provide sufficient voltage to ionize gas even such cold. And hence the starter even in the Penning mixture lamps: It generate about 700V peaks, what then guarantee a reliable startup even at very cold winter nights (e.g. after the electricity was restored after few hour blackout).

There is no starter in penning HPS lamps. Only a starting spiral, which operates like a third external starting electrode.
The HPS lamp with the starter, have the usual xenon, and here the 700V are used to start the lamp.

There is no starter in penning HPS lamps. Only a starting spiral, which operates like a third external starting electrode.
The HPS lamp with the starter, have the usual xenon, and here the 700V are used to start the lamp.

The character of the starting voltage dependence vs the temperature is the same regardless on the gas fill.

The "starter ignited" lamps would use something more easier to ignite than the regular lamps do, as the starter is not able to generate as high voltages, so it would be some mixture, maybe Xenon based Penning...

Internal starter HPS lamps, have the same color as the regular HPS lamp (White). So I think that the only option is a small additional of argon or a different xenon pressure.
Also, you can start regular 50W and 70W HPS and MH lamps using starter for fluorescent lamps: Search videos in Youtube (Especially of the youtuber JendaLinda and George).

Deleted

IMJ: No. The incandescent lamp, is to limit current to the starter. If the starter connected directly to the ballast, it easily overheats and its contacts stay closed.

How is that ? The starter's glow current is less than the HID running. for 70w (1A HID arc) it is not so far from what it would be exposed to on a 80w choke (800mA HID arc, higher voltage). And most starters are rated for 65/80w wihtout any incandescent lamp.....

How is that ? The starter's glow current is less than the HID running. for 70w (1A HID arc) it is not so far from what it would be exposed to on a 80w choke (800mA HID arc, higher voltage). And most starters are rated for 65/80w wihtout any incandescent lamp.....

The incandescent is supposed to protect the starter's electrodes.
The diference towards fluorescent is, than with fluorescent, the starter is long time stressed only when the lamp fail. As the starter is supposed to be replaced with each lamp, it is not a problem.
But with the HPS the hot restrike attempts stress the starter for the whole time it take to cool down, so one stormy night may wear the starter completely.
The lamps with an integrated starter use a thermal cut outdevice, what disconnect the starter when the lamp warm up and connect it back only when it cool down, so when the arc is extinguished e.g. by the mains dip, (during the cool down period) the starter is disconnected, so does not wear out.



@dor123: The 3% of an argon is exactly the gas composition modification, what makes a huge difference in the ignition voltage (Penning mixture), but yet does not alter the color of the dominant Xenon fill.

And the regular HPS could be ignited by the glow bottle starter, but such method is less reliablethan the HV pulses from the regular ignitor (mainly at low ambient temperatures,...)

Theoretically, when using a fluorescent starter to start 50-70W HPS lamps, it is possible to connect the system to a surge protector with an 4 mins delay to manually disconnect the starter (And actually the whole system) by external means. But this is effective only to protect the starter from a hot restrike caused by a power interruption/outage or a voltage drop. When cycling, the cycling occurs after the surge protector, so the 4 mins delay willn't activate.

What you mean by a "surge protector"?
A device clamping the overvoltage can not help, when the actual cause is a (temporary) undervoltage...
And the lamp restart after the power dip (or some temporary disturbance yielding an arc instability and extinction) should be fully automatic, so no manual reset or so...

I think he means the "electronic protectors" i was talking about in a computer forum (another forum where i participate). Those have 3 protections :

Delay - If there is a momentary dip in voltage it will cut off the voltage output with a relay for few minutes. Beneficial alo for fridges and everything with compressor

High/low line - If the voltage is out of 180-280v limits, it will cut off with a relay and be reconnected only if voltages return to normal and stay normal for a predetermined time

MOV's for spikes

Then I think the regular HV ignitor become way simpler solution to the problem...

I'll try this later with the lamp in here, there's a 70W SON-E (I) lamp in tonight for a change.


BG

Deleted