Cheap buck drivers for red LEDs?

[Nil Einne]

First, yes I'm aware of fellow Kiwi torchboy's driver list, which has been very helpful in my research

I'm looking for cheap drivers for red LEDs for 3x or 4xNiMH and 1xli-ion, something under US$5 including shipping. With a modes particularly a slow flashing mode would be good too although not essential. I live in NZ so anything in the US, UK etc, even probably Australia is likely to be expensive when shipped so realisticly I'm looking at DX/KD/eBay etc.

The only things I've really came across are the AMC7135 linear regulator boards (likely crappy efficiency given the low Vf), AX2002 boards (sadly don't have modes, at least not cheaply and not blinky/flashy, my main concern is what happens below 3.6V) or the LEDA 0712 boards (have modes but likely less efficient then AX2002 and limited info exists on these, I haven't even seen a datasheet). There a bunch of shot in the dark stuff like KD product ID 9908, 9910 and 9912, but I'm not particularly interested in being a guinea pig and they and exactly cheap and don't claim to have modes even.

Are there any other drivers people would suggest I'm missing? Or any other thoughts, suggestions etc? As you may have guessed, things like the microbuck etc are a bit too expensive for my needs.

I'm currently leaning to the AX2002 drivers probably the 4 pack from DX. This doesn't have modes which is unfortunate I'll probably add them myself although I may not even bother with modes at first, get it working or at least order stuff while I decide how to handle the modes

 

[Nil Einne]

Some background info/thoughts

I haven't quite decided which LEDs but the Vf will be typical for a red LED i.e. around 2-2.6V. I haven't quite decided on the current but I may be driving 3 or so at ~350mA.

Since this is for a rear tail light of a bicycle I would like a flashing/blinking mode (most likely some will be flashing and at least one constant). However most of the cheap drivers with modes seem to use the AMC7135 linear regulator. These are nice and cheap but while they work okay for white LEDs for a red LED given the low Vf with a li-ion or 3xNiMH (let alone 4x) they're gonna be fairly inefficient. Also I think the fast strobe they tend to have is a bit faster then I would like although many do have a slow strobe option. Finally they don't give so much versatility in the current selection.

So a buck driver would be better.

I've looked at the AX2002. This seems like a decent option, still fairly cheap particularly for the 4 (although I'm still not sure if the 4 boards are always crappier, e.g. with a 1A diode) but cheap AX2002 drivers with modes don't seem to be widely sold. Even download's AX2002 which is a bit too expensive for what I want to doesn't have a flashing mode. The AX2002 does have an enable pin and there a variety of ways I could add modes myself so that sounds like a decent option. Does require a bit more soldering and work of course although it likely means I have more versatility in what modes I add. One question I'm still uncertain of is what happens to the AX2002 when the input voltage falls below 3.6V, I asked http://www.candlepowerforums.com/vb/showthread.php?t=284983 but no response yet.

Finally there's the LEDA 0712? based boards which Torchboy calls TR-0022b (from the PCB). These do have modes. A bit on the expensive side although potentially cheaper then adding modes to my AX2002 depending on how I do it. However from the testing I've seen, these seem likely to be less efficient then the AX2002. Also I haven't yet been able to find a datasheet so who knows what it's supposed to be able to do (I did at least find out the sense voltage appears to be the same as the AX2002). It does seem to be limited to 8.4V and while I don't think I'll ever want to go higher particularly given the likely drop in efficiency is obviously a limitation compared to the AX2002. Also I have to take the two layers apart to change the sense resistor.

P.S. Yes I'm aware of MTBR.com and may ask there too.

 

[Linger]

I'm just gonna through some things out there.

Direct drive - does 2s Nimh give decent power to a red led?

Resistors - the older an less efficient way, but instead of buying a board that doesn't suit your requirements (e.g. no flashing) a simple resistor may drop the current nicely for you. With a 3md switch, you could add a 2nd resistor for a when pack voltage declines.

Run emmiters 2s (in series), your effective vf is doubled and your with-in advertised spec for a number of drivers.
Infact, Dx has a (big) boost driver advertised for 5 emitters BUT if you check the posts on the bottom, user EPROM lists results for multiple emitters (from 2s up) and multiple batteries. I think said driver does flashing too...
http://www.dealextreme.com/details.dx/sku.26106

 

[Nil Einne]

Thanks for your reply/suggestions.

The LED in series and use boost is an interesting idea I haven't considered. Efficiency would potentially be lower and cost probably a bit higher (but it has modes obviously) but definitely worth considering particularly with the modes and already tested so know should work, so thanks for that.

About the resistor idea, I was initially thinking I might be able to use an AMC7135 board with 2xNiMH so that's somewhat similar. Efficiency would be not better then the resistor and it will require more headroom on the voltage and cost slightly more but I can get one with flashing and also regulated output.

Edit: Ooops the biggest issue I just remembered is they only work down to 2.7V (with protection diode removed) or well at least that's what they're speced for (as with the AX2002, I don't know what happens when they run out of spec, it's not something relevant to most people) so not really suitable for 2xNiMH. Completely forgot about when replying/thinking about it this time I also realised the AMC7135 multimode are a bit more expensive then I thought. I think I got confused by the Torchboy list into thinking that you could get 3 for about $4 when you get a single board with 3 of them for about $4 ;-)

Another big issue is that I'm stuck with 2x NiMH (well could add another 2 in parallel of course), or poor efficiency and I would like the li-ion compatibility. I admit weight isn't that big a concern and realisticly runtime isn't going to be much of a concern either especially with 2x2NiMH and I have to worry less about exploding cells with NiMH so li-ion isn't that important. But on the other hand the torch I'm using as my handlebar light is 18650 and I think I worked out how to fix my waterproofing concerns so I'm also thinking why not stick to 18650 (well I was thinking a removable battery pack or something so I can use AA too if I want).

Another issue that has occured to me is I'm not really sure what the Vf will be, it may be in the iffy range for AMC7135 with 2xNiMH. However I'm currently leaning towards the cheap Chinese '3'W LEDs and I'll probably get more then I need so could bin by Vf myself to some extent. I'm and also considering running several at a lowish current so Vf also doesn't have to be that high. I'm actually not really sure how much of an issue this would be in any case. Cree specify 2.3V as the typical Vf at 700mA for XP-Es red/red-orange but also do bin their red and red-orange by Vf, but even Cutter (who I'm not that actively considering because I'm not planning to order enough to make up for the shipping seem to specify the bin) don't seem to specify the Vf bin although I didn't ask if they can. In any case little chance of getting a specific Vf bin with DX or whatever. Perhaps most of them are the low ranges so they don't bother.

As for the DD, well if I do go with the cheap Chinese led option and bin them myself then I can use that to help. And actually one thing I haven't considered before is get the leds, and then work out what to do when I've tested them. Well I was/am thinking I could add the modes later but had been mainly thinking I should order the drivers perhaps even before the leds since otherwise what am I going to do with them (and if they're broken I won't realise until too late). But really whether DD or with resistors I can test them fine. (I don't have any constant current power source.)

Edit: Actually both the DD & resistors is more appealing then I was thinking when first replying. Well the main thing is I was thinking the AMC7135 would be better alternatives, since they had modes for 2xAA. Except of course I now remember the problem is they don't support a low enough voltage. So anyway thanks again for your suggestions

BTW the idea of adding modes myself is two part. Number one is as I've said, many seem a fairly fast strobe and I'd prefer something slower. Well there are the 20/16 mode boards, usually AMC7135 ones which tend to have some slow strobe but then I'm going to have to cycle through many modes if I ever want to change or do it by accident. The other one is the idea intrigues me particularly since other then having greater control over the modes I can potentially do some interesting things. It's not exactly great from a cost viewpoint although since it's also a fun thing, the cost should be seen in that light I guess. And something like the PICAXE shouldn't be too expensive (fortunately it looks like I should be able to get them here at a not too bad price).

Of course that's the fancy option. I think I could do something simpler which should be cheaper like with the 555 chip.

However the idea of adding modes is not something I've really looked in to enough yet to know if that's definitely where I want to go.

Cheers

 

[Nil Einne]

It occured to me another option is one of those boost drivers with linear regulator. Perhaps would have no problem with 2xAA and (single series) red led

 

[steveo_mcg]

I have tested running red CREE's from the DX driver 03256 it runs fine if pushing a little too much current through the emitter. The emitter needs good heat sinking or else i don't think it would last long.

I was testing from a super cap bank so the voltage varied too much to give accurate efficiency but at around the 4aa range it wasn't bad if not great.

 

[Nil Einne]

I have tested running red CREE's from the DX driver 03256 it runs fine if pushing a little too much current through the emitter. The emitter needs good heat sinking or else i don't think it would last long.

I was testing from a super cap bank so the voltage varied too much to give accurate efficiency but at around the 4aa range it wasn't bad if not great.

Thanks. This is with recentish drivers with the AX2002 chips (look something like http://www.dealextreme.com/feedbacks/browseCustomerPhotos.dx/sku.3256~id.52518) not one of the very old ones right?

Did you ever test it at low voltage? Like say 3.3V which is below what the driver is speced for (3.6V) but still should be well above the Vf of the red led?

 

[steveo_mcg]

Thanks. This is with recentish drivers with the AX2002 chips (look something like http://www.dealextreme.com/feedbacks/browseCustomerPhotos.dx/sku.3256~id.52518) not one of the very old ones right?

Did you ever test it at low voltage? Like say 3.3V which is below what the driver is speced for (3.6V) but still should be well above the Vf of the red led?

Yeah I got the driver about 6 months ago and it looks the same.

I was worried about it blowing the emitter when i dropped from specified ranger but i rekon that 3.6v minimum is more for white emitters, i think it was still in regulation until about 3v when it went to direct drive but nothing went

Just to add, if you still wondering about using a resistor for direct drive that's what i did for my rear it lasts my full weeks commute and its pretty much daylight visible.

 

[Nil Einne]

Yeah I got the driver about 6 months ago and it looks the same.

I was worried about it blowing the emitter when i dropped from specified ranger but i rekon that 3.6v minimum is more for white emitters, i think it was still in regulation until about 3v when it went to direct drive but nothing went

Just to add, if you still wondering about using a resistor for direct drive that's what i did for my rear it lasts my full weeks commute and its pretty much daylight visible.

Thanks for that. I was thinking the same thing as you i.e. it may drop out of regulation below 3.6V and actually increase current potentially killing the led but was hoping the minimum was more because of the Vf of white led means it was unlikely to have enough headroom so they didn't bother to spec it below that. Was planning to buy and hope for the best but good to get some confirmation from testing it seems okay

 

[Nil Einne]

A long time later but I found http://club.dealextreme.com/forums/Forums.dx/threadid.1049928 where someone found with a P60 dropin that uses a AX2002 with a red Cree LED, it just doesn't work below 3.5V and barely works at 3.5V (low current). At least there's still no sign it kills the LED even though it would obviously be a bit more useful if it still worked below 3.6V as in steveo_mcg's case. My conclusion of all this is don't count on it to work below 3.6V, but it probably won't be as bad as completely dropping out of regulation and thereby killing the LED.