Thanks a lot for the research on maximum open circuit voltage for the H1-A. Opens up the possibility for nice fully regulated triples, PWM free too. Output power is about right, I'd reduce the output current to ≈2.5A, which I believe is more than enough for compact single cell flashlights with 3S emitters. Definitively a winner.
RBD, you are basically correct. I did not spend a whole lot of time looking through the datasheet on switch frequency, but my money would be on a larger inductor to keep efficiency up and heat down with this driver. That said, the stock hardware seems to perform fairly well, perfectly suited to smaller lights where you won’t be able to hold max output for long due to heat anyhow.
If using in a larger light, swapping the inductor for a larger unit of the same inductance, and ensuring good contact of the ground ring to the flashlight body, would probably allow the continuous use of high mode on a single cell given the 10A switch limit.
Barkuti, yes you are correct, I did mean KX70 driver. The new 4A boost driver for XHP70. Hoping we can push to 5 or 6 amps, will have to see.
Overall I am very happy with these results since it is fully regulated PWM free output with good 3.0V LVP. Sure we can probably hot rod better output with 3 parallel emitters and a FET driver, but really this is more output than most triple hosts can handle for very long anyhow heat wise.
Great write up in that thread, and thanks for that picture that clarifies which resistor to modify output voltage. I took a bunch of photos of both boards after I removed the inductors, just need to upload them somewhere since Google hosting is a pain these days.
No clue as to the ratings of the ceramic caps on the board per RBD’s concerns, but it is functioning as expected at least. Thankfully ceramic caps won’t go bang when over volted like electrolytic, they just lose capacitance.
I really want to try to get this driver into an S2+ triple setup. For under $11 shipped, this board is a big winner in my book.
I think it will fit in an S2+ if the driver pocket in the pill is opened up and the driver is soldered directly to the pill. I was looking this afternoon as I want to do this myself. The idea of a fully regulated triple is very appealing.
I’m playing around with modding the H2-C driver. Stock it was putting out right around 1.4A to an XHP35. Then I put a 0.007ohm resistor (length of wire I measured) in parallel with the 2 stacked R100 resistors.
Input at driver:
For about 74% efficiency. But it only outputs this for several seconds before it starts to blink the output. Next mode down was ~1.8A. The inductor gets hot, but it’s not burning in the seconds before the output starts to blink at around 2-3Hz. Any idea what the blinking indicates? This is all preliminary testing as I plan to get a lower resistance inductor.
Also, I turned up the input voltage to ~2s levels on my power supply, and saw momentarily 3.5A to the LED before I disconnected it. Any idea why this might be? I’m not sure if it would was just a transient high current and I don’t want to try this again for fear of breaking my LED. Thanks for any help.
Bad, verrry bad EasyB. Don't game that way with the sense resistor stack, been there done that (Sir Magical Smoke appeared ￼ onstage).
Around four millimeters of AWG37 (≈⅛ of an AWG28) thinner than hair wire has about those 7mΩ of resistance, and certainly nowhere near the amount of mass and cooling ability to handle the massive amount of electrical power who supposedly should cross through it. Of course, then you can have the “cool” idea of coiling a thicker wire, effectively setting an inductor in parallel with the sense resistors, and then you'll soon start smelling funny and hearing humming noises around the stuff… please stop now before you break your driver.
Using a proper sense resistor stack and 2S battery input for the H2-C should allow you to drive an XHP35 (probably) to 2.5A with little or no issues.
Originally posted on Fri, 06/02/2017 - 20:01; little fixup.
I replaced all the sense resistors with a piece of wire with 28mOhms, which should result in ~2.4A output. The wire is 0.025” diameter solid copper and 50cm long. It’s not coiled so shouldn’t have significant inductance. I get the same results as above, around 2.4A with 1s input but still ~3.5A with 2s input. Is this a symptom of a component being operated out of spec? like the inductor being saturated or something?
I switched back to the two R100 stacked resistors and the driver behaves as it should; 1.4A with 1s or 2s input.
Well, the odd behavior I’m seeing does indeed have something to do with the wire sense resistors I made. I made a 50mOhm section of wire and the driver did not behave the same as with the 2 stacked R100 resistors. I guess I’ll just wait until I get the proper sized resistors.
Yeah a wire that long, even not coiled up will present a fair bit of parasitic inductance to the circuit which could certainly cause the strange behaviors you see.
I will play around with the H2-C a little more when I get a chance, now that I know it is an external switch we can push past the 10A input limit of H1-A. The switch should be good for around 15A, though the stock inductor isn’t.
Using 2S input levels should prove to be very good on H2-C based on my initial test. I bet we can get around 2.5A output safely using 2S input and the stock inductor, as barkuti said. Which could make one heck of a light using XHP35 HI.
I will be sure to post KX70 results when I get the driver. Just got my tracking number a few days ago. I may make a new thread for all the KD boost drivers to document everything. I am also hoping for around 5A output with modifications, the inductor may need changing for that though, as it doesn’t look much bigger than the ones used on the H1-A and H2-C from the pictures.
I got some proper sense resistors from digikey to continue modding the H2-C. With a 0.030 Ohm resistor the output current was right around 2.37A to the XHP35, with 1s and 2s input voltages. But the output was still not stable in 1s mode; after a few seconds the output would start to blink.
Some efficiency measurements. Voltage measurements taken at driver.
Input: 3.46V, 13A, 45W
Output: 14.1V, 2.35A, 33W. Leads to 73.6% efficiency.
I also ordered some toroidal cores to make my own lower resistance inductor, but I got T38 core material which has a high relative permeability (10,000). This leads to the inductor having too high an inductance. Even with 1 turn the inductance is about 100x larger than 2.2microH. I’m using this calculator. With this inductor installed the output current was lower and the output was not stable (it would start to blink).
I took off the stock inductor and measured the DCIR, and it was only 7mOhms. Here is a similar inductor I found on digikey. It is rated for 12A and has a 27A saturation current. Any idea how high the current pulses are with this driver?
Look out for a triple build using H1-A from me soon. Just finished modifying one of the triple pills from kiriba-ru to accept the 20mm driver. It is do-able, but the driver will have to be soldered to the pill, not enough material left to do a retaining ring. Not a problem for me though as I usually solder the driver in the S series lights anyway.
Just need to decide what emitters to run. I wanted to go with 219C 3000K from clemence, but I may go with XPG3 3000K for now since the 219Cs are not available yet. Maybe XPL2.