Have got myself an 1800LM XM-L T6 LED 3 Mode Headlamp, and am wondering about the circuitry. I didn't buy an adaptor, and am holding off buying the 18650 batteries til I know more about the reliability. So I have hooked up 3 1.2V rechargeable C cells via the recharge connector and marvelled at the power.
But... the light does not maintain a steady illumination. It switches off and on at an odd interval. The SOS setting, perhaps? I am supposing there's a bit of circuitry responsible - or that I've already fried the thing somehow!
Yes, that is the SOS mode. Press the power button to cycle through the modes. To get the best out of that kind of lamp, you should use 18650 cells. Buy good ones. 2 of them will cost you around $10 from china or £10 from the UK. Don’t buy crap cells.
While you are at it buy a charger. The ‘i4 intellicharger’ is pretty good for the money.
Finally, DO NOT use the charger that came with that thing. Chuck it out of the window as hard as you can. Just block up the ‘charging’ port with some tape/silicone/blue tack and forget about it.
There is no active current regulator as such on these headlamps. There is a current limiting resistor and a mosfet. The ‘modes’ are just varying degrees of PWM.
Would be glad to get more advice about batteries - wondering if there are better brands among what I see online.
As for the lamp circuit, I should have said that no matter what I do in terms of cycling through the power states, it maintains that pattern - even holding the button down for some longer time. In the video, I turn on the light then cycle through 5 button presses. It maintains the SOS pattern throughout. It won’t even shut off until I pull the power connection!
So I wonder if I have inadvertently forced it into a permanent SOS state.
If the batteries needed are in series.
Using 3*1.2v C cells is only 3.6 volts, assume fully charged then at most 4v, the light needs 2*3.7 so 7.4v.
Looks like the driver has low voltage detection and when it goes under a certain point sticks on S.O.S instead of low power, and with only 4v under no load, you don’t even close to the required power
If you can arrange at least 6~7v you can then test that the light is ok
If they are in parallel then the C cells could be suffering from voltage drop, 3.6v to under 3v under load is not much.
:bigsmile: I’m already baffled, so more doesn’t hurt! Bring on the waffles!
The lamp wants parallel 3.7v batteries. So I have put a chain of 1.2v C cells together, which my trusty El Cheapo multimeter says produces 3.5v under load. And while I haven’t got a variable DC power supply, I can swap 1.2v for 1.5v to get exactly 3.7, or above. Regardless, the SOS pattern persists.
What I’d be interested in - rather than simply returning it - is to mod the circuitry so as to bypass the problem bits. I have no interest in strobes, and would be happy to retrofit the lamp so that it is simply on/off.
So is there anything fussy/delicate about the lamp power source? Is it feasible to wire directly into the LED capsule? That would presumably bypass any fancy signalling patterns.
3.7v is nominal, most 18650s charge up to 4.2v typically ( a few are spec’d to 4.35v but they come with special charger need so I would skip them for now). Panasonic or Sanyo are good cells; most cells made in Japan should be. The come with or without protection circuits and button tops; though careful length is OK if you get either.
I agree with bypassing the onboard charging port. You can replace the driver with a different driver and bypass that feature. Just got one of the Dr. Jones drivers and I think the UI for a headlamp for be great. Pick a driver based on whether you have a mechanical or electronic switch. Keep your cells parallel and pick driver for up to 4.2 volts. You can PM him here at BLF if you can’t find a link to his web info from his posts or a search.
Driver will wire from cell + & - to led + & - poles.
Now that I have time to look at Dr. Jones’ website (http://lux.yi.org/), it’s down. :~
So I’ve just gone to Cree and had a look at the XM-L spec sheets for info about handling, soldering and circuitry. (I’m not entirely sure I’ve got an XM-L, but that’s what the vendor says it is.)
So far, everything I’ve seen is about performance and soldering of the LED itself, though I read something that seems to say the LED needs to have a voltage ramp up and down.
I reckon I’ve got a module built by the retailer, and as the LED seems to be held in place by the white nylon insert, I’m not about to try digging it out.
So it’s a mystery as to whether there are any more electronics inside the metal case. It seems that the closest I can get is to the leads coming from the battery case.
The metal case itself is one piece, with an O-ring in a channel as a moisture barrier. I suspect the case is mainly a heat sink.
All told, I’m still at my starting position of wondering whether I can wire directly from the battery to the LED leads, bypassing the circuit board entirely.
I would guess, given the variety of comments on the modding post, that I could bypass the buck and step-down circuitry, but am thinking it still needs something to provide current control.
The LED pics don’t help much since nothing is exposed, but that may not matter. Surprised not much response here. Did you PM Dr. Jones?
Is that an electronic switch? It looks like one, but it didn’t sound like on on the video.
Aassuming 4.2 parallel cells and that your switch is electronic - If I were tackling it, I’d replace the existing PCB circuit you’ve already exposed and try the lumodrv driver . I just got one and the default setup is on, click to high, click to off - ramping option is easy to use if you don’t want the default high/low. It’s 3A standard, but I assume you could remove 7135 chips if you need less current. The pic on his site show you how to wire it up (assume you’d remove spring).
First, I’d eliminate the charge port wires. What remains look to be wires from cell holder to driver, wires from driver to switch, and wires that I assume go to the LED module.
The first link you posted is the one I was thinking of, but I don’t see buying info there. The other one looks older and links to standard drivers from DX and not the newer ones. [Note: the 2013 page says: “If you want such a driver, you can contact me at budgetlightforum.com (BLF) …”]
Really suggest you PM him to confirm what you need/want. He can send you a programmed driver as shown on that first page.
Otherwise I’d look below for something with the level of current you want to allow - again your choices depends on the way your switch works. If the switch is a more conventional one as commonly used in most flashlights, you could probably just wire it inline with the battery negative.
I would think yes, but that could simply be undersized wire or other resistance if you don’t want mode/level control. You could also burn out your LED and if that’s hard to take apart and get to then wiring in a different driver with your desired level of current control makes more sense to me.
If I wanted to test direct drive, I’d use a pair of 1.2/1.5 cells that total around 3.0v at max.
Haven’t sent a PM to Dr. J - I reckon he’s probably got a website to deal with at the moment. Possibly tomorrow. Lumodrv sounds good. Having high/low sounds good to me.
I think the switch is manual - it’s a button press, as seen in the centre of the photo.
You’re right about the wiring.
Here’s the wiring for the lumodrv - http://drjones.dyndns.info/lumodrv-connect.jpg - and it looks like it would do. Oh, I beleive it’s a standard 17mm so you can check for fit. Don’t know the thickness, but that will depend on whether you take chips off the spring side (along with the spring).
I’d probably take them off for a headlight and for heat management since you don’t know what is backing the led. The ramping feature is so easy to use it would get used when you want to set a temporary level.
His spec says lumodrv is for electronic (momentary connection, I think) and that switch still looks like one to me (though I can’t see side view).
Standard switches are mechanical latch on/off with a half press that is a momentary that cuts the connection (reverse) or makes a connection without latching (forward clicky).
I didn’t look through the IO driver list to see if they have something suitable for the switch. I gravitated toward the Dr. J version for my own replacement where a flashlight I just got had the same problem with SOS instead of steady (likely bad programming in the MCU chip).
Ive got that headlamp. The aluminium pill is hollow, the star is on top, sitting on a ledge, and the plastic retainer is not doing anything.
I had a dedomed xml t6 in it for a year, worked great, so decided to put in a dedomed xpg2-r5 on sinkpad in it the other day. Filled up the entire pill with JB weld too for heat sinking.
It throws less than the xml for some reason and its half the lumens. Next ill chuck an xml2-u2 dedomed in there and see how it goes.
Switch: I think I understand, but not having seen the mechanical type, I’m not entirely sure…
So I’ll just take it that mine is regarded as an electronic type.
Driver diameter: here’s where I get to ask some really newbie questions! Such as…
is that PCB meant to fit inside the metal pill?
(as I don’t see how I’d get it to fit in the battery case.)
Is there a handy assembly tutorial online here?
I can probably get the thing assembled without one, but it would be nice to have for reference.
It looks like the driver has a circular battery contact in the centre - is that also a spring?
1. If you have an inexpensive flashlight with a tail switch you can compare to that. If you can use a DMM for continuity testing, you can check to see how the switch functions. I originally used my DMM instructions but I’m sure there is a Youtube video for that. My electronic switch has no continuity until fully pressed, and then none when I let go - that makes it a momentary. There are mechanical momentarily switches on some lights, but they aren’t typically used by themselves - something needs to latch power for continuous on. If a full press latches the continuity on, then it doesn’t seem likely to be electronic.
2. PCB for you would need to replace the existing PCB, preferably in the same location. 17mm is pretty small - I’m guessing your existing board is about that long, but it is hard to tell with pics. If not you’d have to determine whether you could widen it out. The PCB can also probably be sanded at the edge to narrow it a little (ask Dr. J.). Just a guess, but a single sided driver (4 chips instead of 8 + without spring would probably be ~ 5mm thick) - again ask Dr. J.
Your pill appears to just have the 2 led wires, and the driver needs all 6. In a flashlight the driver PCB normally goes under the pill and above the positive pole of a cell so it uses direct contact for 2 of the 6 wires where there is an electronic switch.
3. Assembly would just be soldering to the points in Dr. J.s picture. You’ll have to figure whether you can get it to fit the available space. I’d be sure to insulate the hot wires & traces with Kapton tape or similar if you can’t fix the driver in place.
4. Yes, I’m guess the center has a small spring that I would remove and use the center pad for positive from the cell pack (make sure cells are wired parallel –4.2v and not 8.4 with 18650s).
If you’re not comfortable with soldering, look for some youtube vids and practice using some scrap wire and contacts first. It’s only hard until you’ve got some experience.
Well, I have couple crappy irons & tips so it’s hard every time even with a little experience. My smallest is a 3/8 inch chisel tip that is older than my adult kids and has been eaten away over the years so that is neither a chisel or a point. A replacement tip appears to be almost as much as a new iron with cheap replaceable tips.
Could just be the switch. Appears to be an electronic switch. Meaning, it makes contact when pressed and then breaks contact when released. Basically, it just sends a signal to the driver when pressed. Disconnect the wires from it and check for continuity with the the button depressed and for lack of continuity when the button is released. If it behaves differently, it probably is malfunctioning.
As a double check, you can touch the 2 switch wires together to see if it changes modes.
Of course! Should have thought of that when I was looking at IC pin voltages.
Once upon a time I did this sort of thing for a living, so have some idea of working my way round a board.
However, the only residual from that era is the notion that I can probably make sense of how a thing works. No DMMs, no fancy soldering/desoldering tools, no scopes or supplies.
**Update - shorting the switch makes no difference to the pattern. I used a wire to mimic the switch, and got the lamp to cycle through the various states, but each time it returns to the SOS pattern after about 10 sec.
One other thing I’ve been wondering is whether there’s enough current draw - given that I’m still using 3 C cells.
When the lamp is on full it draws about 135mA. So I need to find out whether that’s low relative to the spec.
