I recently set-up a benchmarking rig to test a Nanjg-092 I finally got around to building several months after buying the bits from Richard. It pulls 3.8 amps off a Samsung –28A charged to 4.15v. It works.
Out of curiosity I tried testing some current draws on my Small Sun ZY T620 bought last year and promptly trimpot-modded to suck 2.5 amps off its 2s stack. [starts drawing 3 amps then settles to 2.5] I unsoldered a lead from the LED and put the DMM (14ga leads) in series and got 3.8 amps. Doing the math that’s a 73% efficiency. I removed the trimpot and repeated the current draws. Stock, the driver is 84% efficient.
So…Boost drivers aren’t that efficient I guess, at least not when driven above their designed output. This kind of result is similar to that thread discussing boosting the 1/2 AA 800mA driver that Fasttech sells.
The T620 is sort of my single-XM-L throw-cannon light. I need what amounts to a 2s 1XML direct-drive driver. I’d like to put 5 amps through the emitter without approaching incandescent-like waste through the driver.
Now that I think about it you’re right, the presence of a toroid always makes me think boost, because I’ve seen buck drivers that did not have one.
When IOS is back online I’ll check it out. Unless the LD-2C can be boosted to 5 amps, or would that have the same efficiency drop as this stock driver?
You're also doing your efficiency calcs with a usually very lossy item in the circuit that's not normally there, that being your ammeter...
If you have a second DVOM, measure the voltage drop between the two ammeter probes (that will show how much is being lost inside the ammeter & leads) while it's running.
I’m using a HF ‘free’ DMM with 14ga PVC wire leads. The DMM is seeing .4 ohms in the leads, but that .4 ohm would apply to both sides of the calculation, no? I do have a second [original] DMM, but first I blew something other than the ammeter fuse, and then it started reading high for voltage and I’m just waiting for a good time to pitch it.
I never said I wasn’t confident in the HF DMM. When I first got it it was to have as a backup to the original and it matched the measurement values of that one, and its been consistent ever since. I wouldn’t be bringing this up if I thought the DMM was to blame.
That 0.01 Ohm shunt resistor in the meter and the test leads, even 14 ga leads, can affect your measurements and calculations. I had come across this before and it is the reason I posted this about constructing an external shunt resistor. I was measuring 15A and the voltage drop across the test leads and the .01 ohm shunt in the meter really took a toll. I made an external shunt of only 0.001 Ohm.
At 5A you are probably dropping 0.1V across you meter and leads. That would be 0.5W of lost energy.
If you don’t want to bother with making one, Cereal-Killer sells one already made.
I don’t doubt it does, but as long as that ohm resistance is consistent, then it will take proportionally from both the battery and the LED sides of the driver, both at 18 watts and at 7 watts. Thus when calculating efficiency of the driver, and not actual current draw, it isn’t an issue.
Power loss is proportional to the square of the current and since with a buck driver input current is always lower than output current your efficiency calcs will be off. That said, building an efficient buck driver seems to be more difficult than one might assume.
So the DMM ’s resistance is reading artificially lower on the LED side than the battery side. That does make sense as you explain it. I still don’t think it’s really changing the fact that this driver is inefficient when asked to swallow 19 watts of power.
But how many of those 19 watts are being lost in the meter? Nobody knows.
I suppose it would be possible to build a buck driver that's 98-99% efficient (like the DD/single cell versions), but it wouldn't ever fit in anything even vaguely flashlight shaped. And it would cost a small fortune.
Every design has to deal with compromises in one area or another. In the hot rod world there's an old saying, "fast, cheap, & reliable: pick any two."
And a driver designed* to do a target efficiency at a target output may do perfectly fine as long as it's run within the intended parameters, but then one of us yahoos comes along and stacks a bunch of resistors with no other changes to the critical components responsible for efficiency, and, no surprise, efficiency goes to shit. Most often the size/cost/performance balancing act affects the inductor & diode, as long as we're talking about switch mode buck drivers. Ask an inductor rated for 2 amps to carry 6 amps and it will get hot, that's where the inefficiency comes from. On the other hand, a driver designed to properly handle 6 amps and running at 6 amps can have the same 90-95% efficiency as the 2 amp driver running at its intended 2 amps. Ask it to do something it was never designed for and something is likely to suffer.
*'designed'... there's no guarantee all these drivers were designed, much less competently designed. That doesn't tell you anything about the suitability of the driver type (buck, linear, etc.). Just because a switch mode buck driver is capable of 95% or better efficiency doesn't mean there aren't any switch mode buck drivers out there that will only do 70% efficiency even completely unmodified.
In a buck driver it's used as a buffer between the input & output, to smooth out the switching frequency (this is not the same as PWM used to get lower modes).
