3x CREE driving options. (build my own?) Update: Done building, pics and specs within

[VanIsleDSM]

Well, it's finally time for me to build a torch.. I've been a member here for a while and I'm mostly into fixed LED lighting.. but I have a lathe and the experience building LED drivers so I figure I'd better have a torch if I'm gonna be any kind of member here

I just ordered 3 CREE R2s and 3 reflectors.

Looking around for drivers I don't really see any good options though.. The only good switching driver I see on KD/DX is the kennan.. but it's got a fixed output.. do people mod these for an adjustable output possibly?

I've also heard of the shark if I wanted to go with a boost topology and a fewer number of larger cells.. but I don't know where to buy it? I think I saw someone say it was $20? if so I wouldn't buy it.. as I can make my own custom driver for less than that... which is pretty much what I've led myself to....

I'm thinking a buck driver with 6 AA sized 3.6v lithium cells (I don't know much about batteries either) would be a good candidate.. I designed the driver for input voltage of 16.5-25.2 with a pot to adjust the current up to 1.2A.. current dimming is much more efficient, and more desirable for me than PWM dimming in a battery operated situation.. Although the chip I am using has an optional PWM input, I won't use it.

So what I've designed so far is a 90% efficient driver that can do just that.. here are some graphs to show simulated input and output:

First graph is output at 4 different levels.. although it will be able to dim further than 350mA.. I just used standard values plus the 1.2A overdrive.. low ripple current of 15% is good for thermal management. The next 4 graphs show the current being drawn from the cells at the respective outputs.. 1 hour or so runtime at max power from 600mAh cells I suppose? ..and maybe somewhere up around 3 hours @ 350mA power.

Last graph shows approximate Vf output at 4 different levels.

Does this sound like a good setup? Is there something better out there?.. I somehow doubt it, since this isn't made for profit and it uses all high quality parts... but I want to check.. I'd hate to spend my time etching the boards and soldering everything together to find out there was a commercial model that did the same thing for half what I spent to make my own...

 

[TorchBoy]

Re: 3x CREE driving options. (Build my own?)

The only good switching driver I see on KD/DX is the kennan.. but it's got a fixed output.. do people mod these for an adjustable output possibly?

Sure, just replace the set resistor with a multi-position switch connected to a range of set resistors.

Your first graph looks like you should have a bigger capacitor across your output for the higher current setttings.

 

[VanIsleDSM]

Re: 3x CREE driving options. (build my own?)

Yes.. that's definitely the upper limit.. I'm not going to change the cap, but I may bring the max current down just slightly to smooth it out.. although the output is still stable.

How well does the kennan driver handle higher power? What value Isns resistor does it use as standard?.. I'd prefer a pot for adjusting brightness instead of a switch.. usually a 25 turn 10ohm pot will have enough resolution for sns resistors..

Edit: updated with slightly lower max drive current and Vf graph/If graph to match.. couldn't get that specific sns resistor anyway.. wasn't a stocked item.. so I went to the closest one lower that is.. no larger cap needed now... slightly less overdrive, probably not even noticeable though.

 

[TorchBoy]

Re: 3x CREE driving options. (Build my own?)

How well does the kennan driver handle higher power?

My understanding based on the spec sheet and some testing Drewfus2101 did on a similar board is that up to maximum recommended output of 750 mA the efficiency is very good - 85% or so - but at 1 amp the efficiency drops to around 66%.

What value Isns resistor does it use as standard?

The latest version Kennan is said to have a 0.2 ohm, which means 1 amp out. (I gave up on my order after three weeks and went with something else, so I haven't seen any of the new ones myself.)

 

[VanIsleDSM]

Re: 3x CREE driving options. (build my own?)

Well.. that settles it then.. I'll be going with my own design.. Mine will be 90% efficient @ 1amp.. and should be even more efficient at lower drive currents.

I'm surprised there isn't a good buck regulator like this available.. because of minimum order quantities I should have enough parts for at least one extra driver if anyone wants it.. should it all work out.

I wouldn't mind a little more run time.. but I don't want the unit to be too large either.. something slightly larger than a AA in diameter but not quite so big as a C with more mAh would be great.. or a bit longer would be ok too..I see all these 18650 and CR123 stuff.. I dunno what any of these are

I know AAA, AA, C, and D.. guess I better brush up on my battery knowledge.

 

[TorchBoy]

Re: 3x CREE driving options. (Build my own?)

I'm surprised there isn't a good buck regulator like this available.

There was a thread by a couple of CPFers who were planning on making one, but they may have been put off by the economics.

because of minimum order quantities I should have enough parts for at least one extra driver if anyone wants it.. should it all work out.

Ah, but at what cost?

And if you don't mind me asking, what's up with all the double dots? (A full stop or period is one, an ellipsis is three... and arguably shouldn't be overused.)

 

[VanIsleDSM]

Re: 3x CREE driving options. (build my own?)

Well, assuming I don't screw one of them up and I end up with both working I could part with the other for about $16.. not for profit obviously.. just to cover my costs and help out another CPFer.

I don't think cost is that prohibitive. Sure they aren't as cheap as the DX and KD stuff.. but you get a better driver.. and in volume you could probably make them for ~$6-7 or so. I think it's worth it, a quality high efficiency adjustable driver that doesn't fall short on any expectations.

As for my double dots.. I like 'em. Sort of like a comma but with a little more personality.. and no need to mess with the shift key for that pesky capital of another sentence so often

 

[OddOne]

Re: 3x CREE driving options. (build my own?)

As for my double dots.. I like 'em. Sort of like a comma but with a little more personality.. and no need to mess with the shift key for that pesky capital of another sentence so often

Perhaps, but that drives the grammar and punctuation Nazis here (like me) nuts. :laughing: (Yes, I noticed them immediately.)

oO

 

[TorchBoy]

Re: 3x CREE driving options. (Build my own?)

There was a thread by a couple of CPFers who were planning on making one, but they may have been put off by the economics.

I've remembered it was a 1 amp boost driver for one or two AA cells.

Well, assuming I don't screw one of them up and I end up with both working I could part with the other for about $16.
...
in volume you could probably make them for ~$6-7 or so.

I think you'd have lots of interest at the higher price, and LOTS of interest at the lower price.

 

[VanIsleDSM]

Re: 3x CREE driving options. (build my own?)

Almost ready to order the parts.. trying to amalgamate everything from other projects into one order.. just designing the PCB right now...

I have one small issue though.. using the 25 turn 10ohm pot will give me ~1.3 turns from ~1100mA to 350mA.. the switching frequency at 1100mA is about 550kHz, down at 350mA we're up just over 800kHz.. The chip has a limit of about 1MHz.. With my DMM I'll measure the frequency when I build it to see at what current level the frequency will reach is max.. may only be able to go down to 200mA or so.. which I'm fine with.. the issue that I get to is that I don't know what will happen if that frequency is exceeded, damage? I'd think not.. and hope not... but how would I go about physically limiting the 25turn pot from adding too much resistance after it has already turned around one time?

I could go to a 20ohm pot so it won't do a full turn.. and I could physically limit it somehow to stop the frequency from getting too high.. but I'd rather keep the higher resolution for adjustment.. if only I knew it wouldn't be harmful to the circuit if the frequency was exceeded.. maybe it just turns off? Then I wouldn't have to worry...

As for batteries.. it appears there's something exactly like what I was looking for.. those 18650 batteries look just like a AA but longer (and maybe slightly larger diameter too?) @2400mA each I could have just shy of 4 hours runtime at full power with 6 of these! That's perfect.. now do I have to devise something myself or is there a battery holder to hold 3 of these together to fit in a tube?.. maybe that's a question for another thread.

As for you punctuation nazzis.. you have much more to worry about than I, At least I put in a breathing space of some sort, and a little white space to make for a fairly easy to read post. I think the ".." more reflect real type of speech.. which is more akin to what posting is.. I'm not drafting a contract



Edit: After a bit of tweaking with the on time I got the frequency a little lower without becoming cost prohibative.. anything lower than I have it and there's no way other than using larger capacitors with a low ESR to keep the efficiency up.. they get really expensive really fast. This on time tweak also gives a slightly higher max current and happens to work out almost perfectly with a 20ohm 25 turn pot for Rsns adjustment. 1 turn on the pot will bring the output current from 1100mA @ ~500kHz to 181mA @ 900kHz. The first half of the pot turn will bring it from 1100mA to 350mA, then the second half the adjustment slows as it reaches 181mA.

 

[TorchBoy]

Re: 3x CREE driving options. (Build my own?)

As for you punctuation nazzis.. you have much more to worry about than I, At least I put in a breathing space of some sort, and a little white space to make for a fairly easy to read post. I think the ".." more reflect real type of speech.. which is more akin to what posting is.. I'm not drafting a contract

I don't have a problem with you being happy with it or you using it as some sort of signature posting style, but please don't be under any illusions that it makes it easier to read - it doesn't - or that your writing of posts should conform to your personal speaking style in any way. Where does that sort of thing stop? Should I mumble my posts by typing with my fists instead of my fingertips? :sigh:

I'm glad there are people are concerned about correct punctuation. Keep it up, guys. :twothumbs

 

[VanIsleDSM]

Re: 3x CREE driving options. (build my own?)

I would have to disagree. How does .. make it any harder to read than , ?

Rules bend if you're writing a book and you're between quotations. Posts are quotations.. therefore correct punctuation isn't needed, better in fact, as I see it.. is a more personal talkative approach.

I don't mumble when I speak, so there's no evidence of that in my posts.. Quite a ridiculous suggestion, I know you're kidding, but you're still reaching pretty far.. go chase sum kidz dat like 2 type like diz.

anyway.. back on topic.

 

[TorchBoy]

Re: 3x CREE driving options. (Build my own?)

... is a more personal talkative approach.

I don't mumble when I speak, so there's no evidence of that in my posts.. ... go chase sum kidz dat like 2 type like diz.

I didn't say you mumble in your posts, but even if you did there shouldn't be any evidence in your posts. And the example you give, while not quite phonetic, is as you say "a more personal talkative approach." But that doesn't mean it's better. That more personal talkative approach is not easier to read. YMMV.

Anyway, your choice on whether you go for clarity or style. :thumbsup:

 

[VanIsleDSM]

Re: 3x CREE driving options. (build my own?)

Not easier to read in your opinion:laughing:

Taking on the role of speaking for everyone is somewhat ignorant.

I've actually received a comment on another forum about how clear and concise my posts have been.

You wanna know what's much more unclear? pesky little acronyms like "YMMV".. sorry, I don't speak acroenglish... maybe you should be more clear in your posting. Extra dots don't make the content of my post questionable.

Again.. thanks for your criticism, back to building LED drivers now.

 

[Mike Painter]

Re: 3x CREE driving options. (build my own?)

Perhaps, but that drives the grammar and punctuation Nazis here (like me) nuts. :laughing: (Yes, I noticed them immediately.)

oO

Your not going to like what I say about you're post... :laughing:

On a mre serious note, punctuation and grammer are *very* important.

Webster probably saved America from going the way a lot of European countries went. ven though they spoke the same language, the dialects made it very hard for somebody at one end of the country to understand people from the other end. In Germany this was still true in the 60's. Such problems are rare in the USA but I've heard 30's folk music from Appalachia that I couldn't understand at all, uyet it was in English.

 

[VanIsleDSM]

Re: 3x CREE driving options. (build my own?)

The content in this post is degrading fast, I suppose the mods here aren't really sticklers for staying on topic in posts.. again I'll plead.. back to LED drivers please.

I'm going to paste the last thing I spoke about that had to do with building an LED driver.. I really hope the next reply to this post is a comment about this LED driver I am building.. if you're really good at marking a grade student's short story for grammatical mistakes, that's great.. but I'd prefer not to hear from you... If you don't have anything to add about LED drivers.. please stay away from the post button





After a bit of tweaking with the on time I got the frequency a little lower without becoming cost prohibative.. anything lower than I have it and there's no way other than using larger capacitors with a low ESR to keep the efficiency up.. they get really expensive really fast. This on time tweak also gives a slightly higher max current and happens to work out almost perfectly with a 20ohm 25 turn pot for Rsns adjustment. 1 turn on the pot will bring the output current from 1100mA @ ~500kHz to 181mA @ 900kHz. The first half of the pot turn will bring it from 1100mA to 350mA, then the second half the adjustment slows as it reaches 181mA.

 

[ifor powell]

Re: 3x CREE driving options. (build my own?)

I don't know about the technicalities of making a driver but looking at your simulations I think you need to increase the range you are looking over.. With the taskled drivers I use you can set them up with 5 levels, each time you go down a level the curent drops by about half although there is a biger jump generaly to the bottom level if you have the high set up to the max. the lowest level generaly comes in at about 0.04A so well below your 0.18A level.. This is perfectly good for a lot of usage situations. With your pot you should any adjustment you want I would just advicing to make sure you have a good low low..

Ifor

 

[VanIsleDSM]

Re: 3x CREE driving options. (build my own?)

Thanks ifor,

If I set up the driver to use a different switched settings I could achieve a lower level by also modifying the rON resistor.. but I don't want to go that route, I'd like to just have a potentiometer to control the brightness.. and I think 181mA is fine as a low setting for me.. since I'll be using 6x 18650 cells my runtime on the lowest setting would be somewhere around 24h..

 

[VanIsleDSM]

Re: 3x CREE driving options. (build my own?)

Well, I'm done.. Pretty much.

Red and Black are Vin/GND : Blue is +LEDout : Brown is -LEDout : grey is POT back to ground.

It works very well. I ended up sparing no expense and went with an uber low resistance, and therefore quite large inductor. Everything in this driver is over rated, 2.95amp inductor, 2amp schottky diode with 0.42v drop max.. All the capacitors are Japanese, except one, made in USA. Unfortunately the CCCV power supply I had on long term borrow has finally been reclaimed.. I'll be ordering one right away so I can test the efficiencies at different voltages, I'm excited to see how far above 90% it is.

Soldering went pretty well with my $15 iron, I sure could use a needle tip though.. and the one I have is seen it better days as it is.. kind of starting to disintegrate into the solder.. time to invest in a proper soldering station. I also bought some solder paste, but it hasn't arrived yet.. that should make for easier soldering too.

Because of the slight internal resistance of the pot I wasn't getting the current output I wanted, so I had to parallel a few of the current sense resistors together. Now I'm exactly where I was to be on high, 1.14A. I wish I wouldn't have gone with the 20ohm pot, 10ohm would have been better for the extra resolution, and the fact that pushing the switching frequency above 1mHz doesn't seem to matter.. I can bring it down to 40mA no problem without falling out of regulation.. so I think I'll have to order some 10ohm pots, and some lower value current sensing resistors... you can also just bridge the current sense resistor and get a max of about 1.35A before the chip hits the over current shut down, The inductor will start to ring and it'll fall back to about 700mA output.

I plan to add a small heatsink to the chip, I'm thinking a penny will work quite well. It doesn't get that hot, I haven't put a thermocouple to it yet, but my fairly accurate touch-o-meter says it's about 60-70C at 1.14A.. well within spec, but I always like to keep things cool.

I'm going to be using 2 blocks of 3 18650 cells in little trifecto configurations.. so I didn't have to make the driver very small since my battery tube will have a 1.5" ID. You'd have to jam pack the parts in on both sides, but if I redesigned it I might be able to get it down to the standard 17mm size.

So i have the parts to build another 4 of these, although I think I'd like to keep another 1 or 2, I can let go of at least a couple if anyone is interested.. since I did end up splurging for some better parts it would have to be $18 for the driver with a 20ohm pot, if you wanted a 10ohm for better resolution it'll be another $1. Although I plan to do some testing of my own, I'd love to see someone else put one of these through the paces.

 

[TigerhawkT3]

Re: 3x CREE driving options. (build my own?)

Just wanted to chime in and say that this driver is very nice. Adjustable, buck, efficient, and not too expensive. Just make sure it's not too hard to wire up, and you've got a winner.