17mm/22mm MTN-MAXlp - Low-Profile 1A-4A Programmable Buck Driver + 22mm MTN-MAX

Keep at it and you'll get it running. Remember: no pressure on the traces if you go over them with an iron. Pressure and heat is how traces get lifted.

hey richard, I downloaded the file to have stencils made. The top stencil is just fine, but I can’t get the bottom stencil to work out. There is no bottom cream file.

Probably not. You have to go out of your way to have proper cream files made for everything, and I certainly haven't done it and probably won't. I don't use stencils. I've tried them and really don't like them.

n10sivern, did you ever get yours running?

RMM I know you’re busy (especially lately) but have you had a chance to test the HP version?

Yes, and I think that its viability is questionable without some further thought. I built one running at 5.5A and ran it quite a bit in my UF-1405. While it didn't burn up, somewhere along the line it lost the moonlight mode level which told me that something was starting to get fried. I tried cranking up the PWM level and it still wouldn't come back. This tells me that the driver would eventually fail and that a redesign is needed if a reliable 5A+ is to be obtained. I'm guessing that the buck IC or FET is slightly fried, but I haven't had a chance to play with it since then.

I decided that these high-power buck drivers could really benefit from a 4-layer board, both thermally and to electrically. A standard Eagle license is pretty expensive, but for me my time to learn a new program (KiCad) that can do 4-layer boards for free would be more expensive than the money paid. I don't need multiple schematic sheets or autorouter, so it wasn't too ridiculous, but it still wasn't cheap. I designed a new 4-layer board with a lot more thermal transfer capability which I'm hoping will do the trick. They just shipped so hopefully I'll get around to trying one next week.

I haven’t had time to fool with it. I’m going to try on Thursday.

Thanks for the update RMM. My 1405 is patiently waiting, but my wallet needs time to recharge anyways so no hurry.

oh man…how did I miss this!!!

Awesome!

First off, good luck with the move. I hope everything goes well. Secondly, I’m very excited to hear that you think you can get it to 5A reliably with the new board and some time investigating. If you end up getting it to the point where you’re happy I promise to buy them if you put them up for sale. They’d be perfect for insane XP-G2 builds.

Just got my first boards assembled and tested from your batch with the masking issue. I think I’m going to like using guppydrive from now on, it’s going to save having so many divers around with different programming that I have to keep track of.

Keep up the excellent work you’ve always displayed here, this place truly wouldn’t be the same without you.

I agree. BLF wouldn’t be BLF without Richard. He’s a good resource, good guy, good products, and good BLF’er.

Hope the 22's are available soon - great for a JAX Z1.

I'll get 'em posted up soon. They are pretty much ready to go. I've got some sitting here.

Hey Richard, I’ve done 4 layer in eagle. I found it wasn’t really necessary but it’s fun just to build a 4 layer board for oshpark. I learned a lot in that process. The 2 middle layers, I was dead set that I needed them. But in the end you can route around with some ingenuity and get away with a 2 layer board. The thing with osh is they don’t allow you to make connects to the two middle layers exclusively. You have to drill all the way through the board. So it’s kind of limited. Maybe another fab could do it, but found if play around enough, you can route it all on a 2 layer board.

They have the 4 layer rule file in oshpark help but you’ll need a license for eagle 6+ I believe. I would love to tell you what I do but I’m trying to keep it under wraps until finished with my project. I been working on making it smaller (no diode) and running into all the issues. Plus moving from tiny13a to better chips.

I already have a 4 layer board in hand. What you're referring to are blind vias. The 4 layer board has nothing to do with routing in this case but instead is for thermal performance.

I understand replacing diodes with p channel FETs, but there's a reason why switching converters never use one instead of the flyback diode. It's not because of cost. Do you know why?

Yeah osh can’t do it… But thermally, yeah can see your point it’s not for everyone… There’s a lot to the thermal side of things. Your dealing with this kind of heat I don’t think any amount of copper on the board can do the trick. You have to be able to conduct it to the pill. Look to your parts, they have to be capable also. Like the tiny has 125C rated chips and check the rest. The max is fine- I mean pushing 11 amps through the fet here. The trick to thermals is getting key parts to the ring and dumping it all into the pill… Guess doping would work too but I found it’s not needed.

The diode on mine at least is a size thing. I have to mill out some of the pill currently. The pfet is supposed to be more efficient too. I just don’t like the pfet is twice the resistance of the n. I might fit two in and see how that goes first. I’m a ways out myself.

I'll get the 22mm posted up and shared tomorrow. I've had it running in a few lights for a while. Here's one at ~6A, using an external inductor (deep pill) into an XM-L2 (the external inductor lowers overall resistance and definitely heat in the driver). It can also use a Coilcraft SMD inductor. The 22mm is single-sided, which makes building the driver a little less time consuming.

A 6 amp buck driver! I NEED one of those! Or maybe a few! :bigsmile:

Cool! What lights did you have it running in? Or what lights will it fit?

That one is in an X6 clone flashlight, but I've also put it in the Supfire L3. FYI, you can't run the newer XM-L2s at 6A. 5A or so is about it for most of them.