I wrote a fairly long post about this, but lost it after the forum went down for that unexpected maintenance.
I don’t have much time now to re-write it, but here we go while it’s still semi-fresh in my mind. The PWM / Modes chip does not seem to be the common CX2829, who’s pinouts appear to be:
As we can see from ohaya’s pictures, pins 3 and 4 appear to be “not connected” on the unmarked SOT23-6 package. If that’s correct it can’t be CX2829.
The pinout we appear to have on this driver is :
Vin or PWM (probably Vin.)
GND (sort of…)
NC
NC
offtime cap?
PWM or Vin (probalby PWM)
Assuming any of that is correct, pins 3 and 4 probably need to be pulled to GND or Vin in order to change mode groups, if it’s possible at all. If ohaya will set the driver on “medium” and probe those pins for voltage with a DMM we can get a better idea what’s going on and then recommend an action.
So, I guess I have a stupid question (sorry, being a little self-deprecating today :laughing:… Why NOT short the resistor? Would current go to infinity or something?
Or is the concern that current would go high such that it might overheat the driver (which kind of seems like a good thing :laughing:?
I’m not ignoring you all - I’ll try to get those physical measurements you all have asked for, today, after work.
Jim
EDIT: Also, BTW, I only have one working driver (I KNEW I should’ve ordered more of them). I used a heatgun to pull the R200 off last night, and a bunch of the components came off at the same time. I’m planning to try to re-construct the driver when the new resistors come in, but not now since it won’t work anyway - that was part of the reason I asked about shorting the resistor…
Also killed one XM-L2 so far on this “project”… kind of amazing that it’s not more :)…
Tough luck on the components that came off. Putting them back on would be good practice for you…
The buck circuit rapidly switches a FET which puts a current through an inductor. The inductor smooths out the voltage to an intermediate voltage between the full “on” voltage and the 0v off state. The duty cycle of the FET is determined by the voltage drop across the sense resistor. Duty cycle means what percentage of the time it’s turned on. Shorting the resistor will produce negligible voltage drop and push the circuit to 100% duty cycle. At 100% duty cycle Vin = Vout and Iin = Iout (more or less). In other words the buck circuit will stop stepping the voltage down and go into DD (direct drive).
Since the purpose of using buck drivers in flashlights is generally to run a much higher input voltage than output voltage (MT-G2 on 3s/4s, XP- or XM- on 2s or 3s), DD is catastrophic for the LED.
Taking those measurements isn’t directly useful to me either, as I don’t plan to run this driver for anything, but I think it would be a nice thing to do for other people. Maybe bridging some of those pins enables even more blinky modes ;-).
Don’t worry about it! Despite the time we find to chat online, I know most of us have limited time for experiments. It sounded like a joke with a grain of truth and I took it as such, but if you never find time to do the test I couldn’t blame you either.
Did you lose any of the stuff for the driver that fell apart? I was at least half serious about putting it back together. It’s definitely doable if you want to, there are orientation marks for all the components with critical orientation. I feel a little bad about that one since I pushed you to steal the resistor!
Based on his post above, I think Slim Pickens may have been specifically interested in the size of the toroid itself (the donut shaped piece) without the wire wrappings.
I see what you mean. Maybe an accurate figure for wire thickness would be useful instead?
I think 4 amps could be a bad idea. The freewheeling diode (SS34) is not OK for that current, and I think the little shrimpy FET is also not going to be a fan of the higher current. You may be able to run this thing that way, but your losses will be high and if you get a failure it won’t be a surprise. Also, pushing a buck driver too far can zap LEDs due to spikes in the current. In short: no, do not do that.
I wouldn’t sweat ~3A, whether it’s 2.8A or 3.2A or whatever. Not everyone agrees with my decisions though and I certainly don’t know everything. There is a lot of unknown stuff happening in that driver.
Those tiny parts just don’t look like 4A would be their friend. For that matter they don’t really look like >1A would be their friend though, so who knows.
Yes, 3 & 4 are the pins farthest away from the dimple.
Unfortunately I’m not 100% sure what’s going on there. We’d better start by measuring everything against the outer ground ring, but we may end up measuring against that big crescent shaped trace that everything* seems to hook up to. I don’t really understand the purpose of the separation between the “true” GND ring and the other one unless it has to do with implementing the Zener diode.
*This is what I mean by everything:
303 resistor
104 resistor
cap
W8 diode S4 diode (err, I guess it’s not hooked up to the crescent)
cap / white thing
R200
AOEC #2
Toroid OD (NOT incl wire, I think): 8.05 mm
Wire diameter: 0.30 - 0.40 mm (I took several measurements, I think some parts of the wire was near where it was soldered)
Voltages to (emitter side) ground ring with driver in medium mode:
Pin 3: 4.18V
Pin 4: 0.115V
EDIT: My resistors shipped from Digikey (25 0.1 and 25 0.2 ohms)
