Driver Efficiency? This is bad...

edit: this is sort of a double post from my thread for this mod - not much interest in that (it is kind of a frankenstein), but I'm hoping someone can shed some 'light' on this situation...

Today I potted this driver (# kaidomain sst-50 driver) with devcon 2 ton and alumina.

http://kaidomain.com/product/details.S009442

I connected a 5AH 12V sla and an xm-l to the driver and the driver pulled 1.6A from the battery on high!

I didn't hold that test lead on there very long

so, I tested a little more thoroughly w/ a pt-54.

pretest battery voltage 12.8V

battery voltage under load 12.5V

current from battery to driver 1.47A

current to pt-54 4.93A

voltage across leads to emitter 2.2V

WTH?! <60% efficiency?! Is this somehow related to the low Vf of the pt-54?

I just tested again to make sure, and I got the same readings. The driver is potted inside a copper pipe fitting and after about a minute on high it was uncomfortable to hold my finger on it for more than a few seconds.

Meanwhile, the host isn't even warm...even w/ my finger right next to the emitter.

edit2: 2x cgr18650ch, 8.05V at rest, 5A to emitter, 2.56A to driver - depending on voltage sag, that I didn't measure, ~ same efficiency (or worse)

Ouch. At $10 this looks extremely bad.

I've been worrying about driver efficiency for a few years. We had the Shoppe building a variety of good drivers for the old M@gAA dropins -- still for sale -- with real attention to how much efficiency could be squeezed out with careful design.

But as LEDs got brighter and we started seeing cheap drivers, the efficiency information disappeared.

I'd love to see this given as much attention as it deserves.

Waste Makes Heat.

yeah, bad at $2, but ouch at $10

luckily, there will be two ways to power this light, the other will be nanjg 105c w/ 3 or 4 nimh

this driver will usually run off a 125ah deep cycle, so it doesn't really matter as long as I heatsink it - but it is crap, don't buy!

Did you happen to test the driver before potting? Devcon really has no heat transfer value and might actually be grounding the driver to the copper. Not a direct, full on ground, but more like a small bleed. I would (if you have another one), try it long enough to get a reading on it, without potting it.

Just my thoughts here. I'm no master whne it comes to drivers.

So, You measured a Vf of 2.2V? The minimal specs at Kaidomain state an output voltage of 5V. If the Vf is really 2.2V, You are operating the driver out of its specifications and in this case, a low efficiency is not the worst thing that can happen.

Old lumens -

Normally I would, but I didn't because of reports that the driver dies almost instantly if not heatsinked. It came with a spongy thermal pad. That I didn't use.

The mixture I'm potting with is (theoretically, based on the 2 components physical properties) thermally superior to the most inexpensive purpose made thermal potting compound for electronics (mg chemicals).

It 'seems' to be conducting the heat fairly well, since those MM readings are fairly stable over the time it takes for the thing to get piping hot...

but, I don't know much about this stuff....or I wouldn't be posting...asking for ideas, so thanks :)

erzengel -

I think that is the input voltage, as it was also written on the little slip of paper that came w/ the driver, and says it is intended for sst-50.

The pt-54 can take a higher voltage (I think spec says 3.5V) but is intended to take ~ what it was drawing. I find it fascinating that it was working that way - I think there have been several questions about how different Vf (usually non-white) emitters will behave when driven w/ circuits intended for standard white emitters.

Well, apparently this one says keep that extra voltage - but maybe burning that off is exactly why its so inefficient? Ok, I have no idea what I'm talking about here lol...

I was afraid to buy this driver for 10 bucks. Thanks for biting the bullet on this one, sorry for the horrible performance.

I have successfully run SST-50s with a take out driver from a Manafont drop in also the same as the 980L or the C-88 driver. This driver will deliver well over 5 amps with a decent 18650 to an SST-50 so it will equal a well driven XML in output.

I bit it so hard I swallowed half.

Good info on the other driver. I think I'll have a use for that.

Unfortunately, in this case I really want to use 12V source. Its really too bad DX can't get their GeminiDeal stuff together and re-stock their sst-90 driver. It is the cheapest 12V >5A option I've found...

Driver hell with high amp low voltage emitters. Still waiting on mt KD 9A.

http://kaidomain.com/product/details.S020079

Let us know when it arrives, I was looking at that one too.

There must be Chinese designers and manufacturers who want to build better drivers (and want to connect with potential customers around the world who'd see the value in better drivers).

Will do. Working with the PT54 red. Have some IMRs laying around and want to put them to work.