Very light weight Mce running headlight.

[ifor powell]

This is my latest home made. It is what I would use for an ultra run if my body was up to it.

As it is it will be used for winter training runs and as a spare for Night Orienteering and MTB outings where I have heavier and more powerful lights.


Weight 160g
High bin cool white Mce
Ledil CMC-D Optic
1 x 18650. Optional 2 x 18650 external pack.
Home made driver based off this post

The head is made from a heatsink and cut and glued aluminium. The weight is low but there is good surface area so when running the cooling is good. In stationary air it heats up fairly quickly e.g. 5 minutes on high would get it too hot to touch.

Headband is from a £5 headlamp I saw and bought specifically for the headband its nice and comfy without any excess weight. Battery and driver are attached with Velcro.

Battery holders are glued sheet plastic I have 1 cell and 2 cell versions. They are not immersion proof but could easily be taped shut in the event of heavy rain being expected, light rain should not be an issue as they are.

Driver is based off this tutorial with some tweaking and using an 8 pin tiny85. Code is C rather than assembler and at the moment would fit in the smaller tiny25. Regulated levels in milliamps are 50, 125, 250, 500, 1000, 1500, 2000. There is also an option to go direct drive, with 1 cell I get 2700ma as measured off the charger. Only real smarts in the driver at the moment is a flash on dropping out of regulation.


Driver case is folded and glued sheet aluminium. The FET is bolted to the case with thermal compound but I have not noticed it getting hot. No cover on the case yet I just tape it up for now and thus have easy access for program enhancement. The case is not immersion proof but the button mounted underneath is ip67 compliant so I expect no trouble in the rain.

The battery can be connected directly to the led without the driver and you just get direct drive and save 25g


In use having the battery driver and head distributed around the headband means that the weight is hardly noticeable and there is no wobble up and down or side to side. for just walking the 125ma level is plenty the 50ma would be suitable for in tent use. When moving quicker the 500 or 1000ma level is probably sufficient most of the time. With the higher levels available for enhanced route finding.

For a more extended outings the 2 cell pack with extension cable to put it in the rucksack top pocket should be sufficient for all night use with some care taken on the level selected. The optic has a wide spot with no sharp transition which I can find distraction when running.

Here is a link for more pictures.

 

[JHJHJH]

Excellent lamp!

 

[SilentK]

Looks very nice! It make me jeoulus when i see that the only mc-e headlights are custom made.:thumbsdow someone needs to mas produce these.

 

[BillyNoMates]

Hi Ifor

I like your headlamp concept.

You have certainly reduced the lamp unit to the bear minimum, I'm not surpirsed that it gets really hot when stationary - I've tried various heatsinks on my lamps, but cooling by natural convection really relies on the hot surface being HOT (ie 50C above ambient) to get a decent heat transfer. As you point out even the gentle air-flow from walking can provide much more cooling.

Although it is an interesting idea to put the driver separate for weight balance, if the driver was placed in the same unit as the LED, then the home brew driver (BTW, I think it is great that you've been able to build the driver by following instructions posted on this forum - absolutely brilliant) could be used to monitor the temperature and reduce the drive level to help protect the LED in those scenarios where the temperature creeps up.

Also, you may be able to squeeze a bit more run time out of the battery by using a buck-boost driver, that way the regulation can be maintained when the battery voltage drops below the LED Vf.

I have a spare part-built version of the buck-boost driver I did for my 5xRebel Bikelight/headtorch project. This should be capable of driving the MCE if the die are independently wired. I'm willing to let you borrow it to play with, PM me if you are interested.

 

[ifor powell]

Hi Ifor

I like your headlamp concept.

You have certainly reduced the lamp unit to the bear minimum, I'm not surpirsed that it gets really hot when stationary - I've tried various heatsinks on my lamps, but cooling by natural convection really relies on the hot surface being HOT (ie 50C above ambient) to get a decent heat transfer. As you point out even the gentle air-flow from walking can provide much more cooling.

Although it is an interesting idea to put the driver separate for weight balance, if the driver was placed in the same unit as the LED, then the home brew driver (BTW, I think it is great that you've been able to build the driver by following instructions posted on this forum - absolutely brilliant) could be used to monitor the temperature and reduce the drive level to help protect the LED in those scenarios where the temperature creeps up.

Also, you may be able to squeeze a bit more run time out of the battery by using a buck-boost driver, that way the regulation can be maintained when the battery voltage drops below the LED Vf.

I have a spare part-built version of the buck-boost driver I did for my 5xRebel Bikelight/headtorch project. This should be capable of driving the MCE if the die are independently wired. I'm willing to let you borrow it to play with, PM me if you are interested.

The light this was designed to replace was a 2 XR-E with an intergrated nflex driver so it has the thermal cutoff advantage.

With just a single Mce the space needed for the driver is considerable compared to the mce and optic and I just did not think I was up to making a small integrated unit.

I have a spair input avalible on the microcontroller so I have been toying with using it with an external thermistor in the head unit this is one of the enhancments I am considering. I want to use it for a bit for real though befor doing much more with it.

Driver wise I had experimented with a 2s2p and a maxflex acting as driver but reprogramed to limit the current dynamical based off an external tempreture reading and the input batter voltage. The maxflex as driver did not work well though the battery sag with the high curent just gets very messy as there is positive feedback more sag leads to more curent and more sag. I tried using an 18650 + a Ni-MH AA cell together with the maxflex, this worked better but still was not too good and was heavier battery wise than I wanted.

I am happy with the linear driver. The % of the nominal batter capacity that gets used still in regulation is 97.5% at 0.5A 91.3% at 1A and 87.6% at 1.5A which are all very good numbers and you get a nice 45 minutes of out of regulation light before getting down to the 50ma level. The 2A level is not realy worth it the line follows the direct drive one too closly you only gain 10 minutes at the 2A level and have a lot of the usfull batter life out of regulation. I will see what things are like with the 2cell pack at 2A the reduction in internal resistance may make the in regulation time better.

I am somewhat sceptical about a buck boost driver setup they generaly do not have good efficentcy numbers what are you getting? After seeing the behavure of the boost driver when the battery gets low I think I will stick with what I have and the gracefull out of regulation tail off is very nice and a major improvment on the old 2 xrc light where I have 3 or more 14500 cells. This has a nasty habit of the battery protection circute cutting in and you just go from full output to nothing. Even with the nflex giving a flash warning this is not very desirable.

Ifor

 

[BillyNoMates]

Hello Ifor

I can understand your concerns about CC drivers placing more load on the battery as the voltage sags - it is important to ensure that the battery can supply the necessary current at the specified cut-off voltage otherwise you end up in trouble. There is nothing worse than the light just turning off. To help get around this issue, I force the lamp into the lowest drive level once a specified supply voltage is reached (at the moment I'm using 3,0V).

There is nothing wrong with the linear driver approach - the efficiency may be worse at the start, but improves as it starts to come out of regulation which gives the long run-out you mention.

I'll try to do some efficiency measurements over the next couple of days - that way comparisons can be made. It will be interesting if nothing else.

 

[Jagge]

Well done, it must be excellent light. Weight-output-running time-beam pattern combination looks superb for night running purposes. I don't think there is anything like this commercially available. This may be the best running headlamp there is at the moment.

I still sort of hesitate to attach li-ion cells to my head. How much weight you think you saved or got more running time by using li-ion pack instead of NIMH (like 3 x sanyo 2700 AA cells)?

 

[ifor powell]

The 18650 cells I have are 45g and the single cell holder I made is 25g so 70g all up. I do not have any modern NIMH cells only some cheap 4 or 5 year old ones rated at 1800mah and only good for 1500mah now. I don't thinck that 3 of these will be capable of driving over 1A they have a lot of sag, 4 should work I will do a test. Three of these are 85g and a 3 cell open to the elements holder is 10g so 95g so not that much heavier if modern cells could sustain the output. I will go and find some of the runtime test graphs and see what the voltage looks like at 1.5 or 2A for modern 2700mah cells.

The different shape of the battery pack plus the extra weight would mean having to put them on the very back of the strap so as to get three points of contact to stop any wobble. I have to do this with the 2 18650 pack I have, it will not sit still without wobbeling on the side unlike the single cell pack which has no movment of it's own at all.

 

[ifor powell]

I tested with my old AA NIMH cells. I have two different brands both rated at 1800mah. Fresh off the charger they could only manage 1480ma and 1600ma respectivly so they would not be much use. 4 cells works. A bit too well direct drive max I get 3400ma so well over spec for the MCE. With 4 cells the fet is liable to start getting hot fairly quickly at say 1500ma but the metal case would probobly cope. 4 cells weigh 110g in the open holder I have.

Looking at the battery shootout thread there is quite a bit of differance in the sag at the 2A-3A levels between the various cells. The Sanyo 2700 ones look to be good on the sag front as well so I would guess they may perform similarly to the my 2500mah 18650 setup. I think the modern hybrids are also better at I high curent but I only have a couple so can not test them.

 

[Jagge]

Does this driver setup any advantages over simple AMC7135 based setup I have been using - other than lots of levels? If it works as a "smart resistor" as I believe it has about same efficiency and behaves same way as AMC7135, drops off regulation to direct drive when battery voltage get low, it has thermal protection, V in must be 0.12V over needed voltage to stay on regulation.

 

[ifor powell]

Does this driver setup any advantages over simple AMC7135 based setup I have been using - other than lots of levels? If it works as a "smart resistor" as I believe it has about same efficiency and behaves same way as AMC7135, drops off regulation to direct drive when battery voltage get low, it has thermal protection, V in must be 0.12V over needed voltage to stay on regulation.

No real advantages over your 7135 setup from an efficentcy point of view. It is a linear regulator (smart resisrtor) just with a microcontroller in charge so I set my own levels and have the out of regulation warning feature. It is only any good when the batery voltage is just over the required led vf. I probobly need less than 0.12V to stay in regulation but down at these levels you need to be worried about the wiring losses just as much.

I have a rear bike light setup with two sets of the same circute from one microcontroller driving a pair of red XREs. I have various different flashing modes avalible that you would not normaly have and it's potentialy very bright compared to a normal bike rear light. Good for daylight fog. That is a 2 NIMH setup and works well the vf of red leds is only 1.9-2.0V

 

[Jagge]

OK. I can see in ultrarunning use one might find this driver usefull and it is better than AMC driver for sure , because you can set brightess just right for the needed running time & battery capacity.

My 3 x Sanyo 2700 pack with connector weights 93g. I guess 3 x eneloop pack would be slightly less than 90 and it would suit you light quite well if you are not happy with your li-ion cell for some reason.

 

[TorchBoy]

Great amount of light output on high you've got there - makes the modified Apex look quite dim. Nice.