Undertsanding Instaflash Limits

[Chodes]

Undertsanding Instaflash Limits

I've already got a small collection of instaflashed bulbs. Would appreciate some suggestions on potential setups:

Osram 64440 - I was using 5 x Sony 18650V cells and as expected , no problems with the bulb. Wanting to push it a bit further , I tried 6 cells. Started at around 3.9v to be safe , gradually increased charge on the pack until the obvious - around 4.1v / cell is about the limit. 4.15v / cell flashed the bulb. (no softstart)
Lux's charts show instaflash at 22.4v. I'm using a mag switch with kui kit. I assume I'd lose around 0.5v with resistance of the light , so at 3.9v/cell I would have had around 23.9v pack volts , around 23.5V bulb. Under 5.5A load cells drop quickly to 3.5v cell , 21v pack , 20.5 bulb.
At 4.1v , I had 24.6v pack , 2.2v above flash limit.

So I'm trying to understand ballpark estimations - when chosing bulb , can my pack volts be above bulb flash limit (even without soft start) relying on the instant volts sag under load? Is nominal volts better to use for deciding?
Would I be likely to be losing more volts to bulb due to resistance?

If I use an AW incan driver , I assume I'll have less overall resistance. Ignoring that , would the softart allow 6 cells to be fully charged or will the bulb still be damaged in the first few seconds until the pack volts drops?
If using the AW driver , will using 30% setting for the first 30 seconds or so save the bulb , or should it be considered the bulb is still getting a too high a voltage , even though it's only on 30% of the time?

I understand there are no definite anwers , just trying to save a few popped bulbs.

There's an AW driver waiting at the PO , wondering if I'll have problems using 4 of my Sony cells with the driver and 64625. Flashes at 14.2v, bulb volts should only be above 12v briefly with almost 10A load.

 

[mdocod]

At the ~5 amps (give or take) that you are running through this bulb, you can expect voltage sag from both the cells and from the resistance in all the contacts and whatever switch you are using.

at ~5A, a LiMn charged to ~4.1V (they tend to settle close to 4.1V if you let them sit awhile, so this is technically a "full" charge IMO) cells will sag to ~3.75V within a few seconds, and then you have whatever resistance in all the contacts between the cells and the switch to get to the bulb.

Point being, with cells charged to ~4.1V, when you are using 5 of them, you are going to be starting around 19V@batt before taking into account other components of resistance in the system. With 6 cells, ~22.5V- which is definitely where these bulbs will flash, but there is more to it than just this of course....

Before going any further, I MUST stress that you can not use the flash point in LuxLuthors destructive testing to make direct comparisons to battery driven applications. Batteries have a capacitive behavior when the load is first applied, and it takes a few milliseconds for the voltage to settle down to the levels described above. In LuxLuthors testing, you are seeing where the bulb flashes when voltage is ramped up steadily, no sudden spikes involved here. Another point to make is that, within a few tenths of that flash point, there is VERY little bulb life available, and it could have popped sooner or later as bulbs gets very unpredictable when pushed to the bleeding edge. In other words, any configuration that actually pushes the bulb around that flash voltage after the cells have "settled" into the load of the bulb is likely to flash the bulb every time.

With that in mind, if the bulb flashed at 22.4V with a steady ramping up of the voltage from a power supply, then there is a good chance that you would have the same problem on 100% with the AW soft-start driver. Draining the cells down on a lower setting some first may help, but it's still not going to be a perfect build. Ideally speaking, it would be nice to regulate this lamp right at around 20V (give or take) to maintain some useful bulb life.

My personal feeling is that, the open circuit voltage of the pack at full charge can be at or around the flash point of the bulb, provided you know that the cells will drop in voltage quickly to a more reasonable drive level.

As an example, I am currently using an osram 64275 in my M6, driven by 2xEmoli 18650s. The results seem quite good so far. The bulb is handling it very well. The open circuit voltage of the fully charged and rested pack is usually ~8.22V, the instaflash point for this bulb in LuxLuthors testing is 8.1V, but these cells drop quickly to around 3.7V per cell at this load and continue tapering off as they run of course. (very comparable to the 1185 output, but with a lower voltage higher current setup). It will be interesting to see how this plays out as far as real bulb life goes (haven't used it enough yet to say anything for sure).

-----------------

At 10 amps, those sony V cells will still be starting off around 3.6V after those first few milliseconds of capacitive effect.
http://www.candlepowerforums.com/vb/showthread.php?t=205814
(see link for verification, another VERY useful thread by luxluthor!)

That's 14.4V starting. I think you're going to fry that 64625 on 4 of those cells. The 62138 may be more appropriate, just guessing though.

Eric

 

[lctorana]

...My personal feeling is that, the open circuit voltage of the pack at full charge can be at or around the flash point of the bulb, provided you know that the cells will drop in voltage quickly to a more reasonable drive level.

I think that's worth adopting as a rule of thumb.

 

[Chodes]

Thanks Eric.
My Sony cells are a little "underperforming". They sag to 3.5v under 5A load (capacity seems OK though) so I'd guess they'd sag a little more under 10A load.

I tried 4 of them with a 62138 tonight. I'll post some beamshots.

 

[Chodes]

Osram 62138 with 4 x SonyV 18650s was short lived. Was using Kui modded mag switch. About 20 mins run, around 20-30 startups. Now trying AW Incan driver. I'd like to run the 62138 on 4 x Emoli 26700s, but if it proves unreliable on the 4 Sonys, 4 x A123 cells should provide a safer voltage.