Led Driver test
I bought a couple of cheap drivers from FastTech to play with, running them on my testbench got the following results.
This is not meant to be a review, bust just a check on how some drivers work.
The drivers are a buck, a boost and a linear driver.
All my tests where done with a XML led, using another type of led with higher Vf would probably have changed some of the result.
3 to 4.2 volt 800mAh driver (FastTech 1114501 East 060A)
The driver is a single mode driver. Diameter of this driver is 17.5mm.
The driver has a efficiency of 80%.
The driver is buck only (i.e. will only reduce voltage), this means that the led will not reach full brightness before the voltage is above Vf of the led, with the used led it happens at around 3.7 volt input voltage (The led only needs 2.9 volt, the rest is lost in the circuit).
AA/AAA battery 550mA driver (FastTech 1124802 Nanjg 102)
The driver is a single mode driver, but does also exist in a 5 mode version. Diameter of this driver is 12.2 mm.
I started my first test from a low voltage and slowly increased the voltage, the driver started working at about 0.9 volt.
At about 1.25 volt the driver increased the current consumption drastically and then the led dropped out.
The efficiency is not very good for this type of driver, due to the low voltage and high current.
The led gets up to 375mA.
Starting at the highest voltage, where the driver works and then reducing the voltage, shows that it works down to about 0.3 volt, before the led goes completly out.
This time I did get higher led current (At least for a very short time), probably because the drive was cold when it was at the high voltage.
It looks like this drive will work fine on alkaline, but will have trouble with lithium, NiMH may or may not work, depending on circuit resistance and age.
3 to 4.5 volt 2800mAh driver (FastTech 1122302 Nanjg 105c)
This driver is a linear driver based on the 7135 chip or rather 8 of them and a processor to control them. Diameter of this driver is 17 mm.
The number of modes in this driver is configurable, from two to five, it does also have a low voltage warning and cutout.
The modes are Lo (5%), Mid (30%), Hi (100%), Strobe, SOS according to the documentation, I have not checked it, my test is only of high mode.
Being a linear driver, it need a voltage above the led Vf, before it will work and the efficiency will depend on how much the voltage is above Vf.
The driver will reduce the current when it gets to hot (8 linear drivers on a small PCB do get hot very easily), as can be seen at 4 volt and above.
I had to start this curve at 3 volt, to avoid the low voltage warning mode.
This time I started from 4 volt and reduced the voltage, the low voltage warning activates at about 2.85 volt and turns off soon after. With a battery this there will be a long time before it turns off, when the led current is reduced the battery voltage will increase.
My test setup
The DUT is the device under test, i.e. the led driver.
I have tried to make the test setup with very short wires and low resistance.
As input to the capacitor I uses a computer controlled power supply, but the voltage I am using when drawing the curves are measured at the driver with a DMM.
Total I am using 3 DMM’s, the power supply measures the input current.
I did not see any reason to measure brightess of the led, the led current is just as good for these measurements.
Led Driver test
If I decide to test some more drivers, what tests would be nice to have?
Is it confusing with two sets of charts, one with voltage going up and one going down? I could cheat and combine them in one chart.
Some ideas from the thread:
- Test at all brightness settings.
- Include measurement of pwm frequency.
- Build XM-L2 test rig that can be used with 1, 2 or 3 leds.
- Change old test rig to XP-G2
I have talked with Hank from intl-outdoor and he gave me some discount on the drivers. I have ordered these:
I will probably start on the review at the start of September, depending on shipping time and how fast I get the new test rig build.
2AA boost driver from fast tech would be nice. Also 2x 18650 high amp drivers would be awesome.
Charts are great the way they are.
I know it would be a lot of extra work, but I’d love to see some data of the lower modes too, to know how efficient they are compared to the highest mode.
Thanks for these tests and great work as always HKJ!
I actually wouldn’t mind you using affiliate links, as your tests means a whole lot to us.
I think you can disregard completely charts with increasing voltage because that is not real life situation, process is always identical: charge the battery in/with the charger and discharge it in the flashlight…
Please supply some SKU's, it makes it much easier for me to find the driver.
Depending on the driver there can be a couple of problems, mostly with selecting mode, on a driver like the 105c I would have to do two curves for each mode, one from low to high voltage and the other from high to low voltage, to get around the low voltage warning and the high temperature reduction.
But generally I would not expect it to be that much of a problem to do.
Strobe and sos modes is another story, the jumping up and down in current and voltage will affect the measurements.
As long as I try to simulate a specific battery it would work, but I like to test the full specified range and there it does not always work.
Look at this curve from the 105c driver:
It is supposed to work up to 4.5 volt and I started a bit above 4.4 volt, due to heat the driver shut down and has first cooled sufficient to use full power at 3.7 volt (I do only use a few minutes to make this curve).
When I start from low voltage and goes up it looks this way:
When mounted in a flashlight it will depend on how much cooling it get.
What is the value and wattage of the resistor in series for current measurement?
Really nice test…
I did some small personal testings with light meter only and 2,8A nanjg 105C drivers and it is really not efficient as I thought it would be and your test confirms that.
I changed from brass/aluminum/copper pill… The thing is that heat will hit led sooner than you think and aluminium/copper pill can really prolong efficiency of light but not much… I will have to do more testings on this.
So far the most efficient driver I have is Dereelight 1SM Buck driver at 1,5A… It can run continuously for hours without light drop but 2×18650 is necessary for this and it will not impress with light output like nanjg 105c.
Hope one day some guy will invent better 1×18650 drivers, and some kind of current regulators/boosters so we can enjoy single 18650 usage without efficiency drop for longer periods of time.
It is a 0.01 ohm resistor, i.e. with 3A current I looses 0.03 volt over it, it is 5 watt.
Clearly to make the driver more efficient, we need to add more 7135’s until the light is basically direct drive on high
Once again HKJ, great stuff!
The graphs you are making makes your driver review top notch!
I hope to see some info on PWM too, and how audible it is.
I would not mind seeing proper reviews on these drivers:
Testing those on tripe XM-L, single MT-G2, and also single XM-L (on those drivers that are suited/capable of single XM-L).
Maybe you are qualified to comment of how mod-friendy they are too? Im thinking about increasing or decreasing current by changing resistor values.
I know some of those drivers are used by several people, and various info can be found in several places, I have shared some results too, but it would be nice to be able to easily compare them using your graphs and info!
Just the thread I were looking for. Thank you so much for your efforts.
A question about 3 to 4.2 volt 800mAh driver (FastTech 1114501), do you have any further info if this driver can be used with multiple liIon setups? Does it support 8.4V or higher like there are some 18V drivers?
Thanks for the testing HKJ. Maybe you can find a decent 3 mode NiMH driver. I’ve given up on a driver that operates correctly from .9 volts to 4.2 volts.
I recently bought that Nanjg 102, i now know its not worth its weight in salt
This thread and willingness to test other drivers fills in a huge void in flashlighting, thank you HKJ
These are the drivers i would be interested getting tested, but the cost would be kinda high for all of them
A single AA constant brightness driver is what i am looking for, and the LD29 looks interesting (looks like it has an inductor), i’m kind of confused with it
I just realized…….using actual batteries will yield higher efficiency, especially with additional 7135’s raising the vF and sinking the input voltage.
Take for example the Sanyo 2600 cells, at a 3A load they are at 3.8v within a minute. I think the 4.2v results are misleading since we equate that with a fully charged battery when in actuality the driver will not see 4.2v on high with a battery.
Awesome, very interesting for someone new to modding.
you are corect, voltage sag kicks in immediately, but for completeness having 4.2V results is good