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

Great! Did you figure out what made the others not work?

The first two I have no clue. The second one I made wouldn’t work so I put it away to take a break. I got back to playing with it a couple of days ago and absent minded me fried it. I forgot to change my test bench setup back to 2S. I inserted the batteries like it was 2S in my side by side battery holder. Well, it was 1S and the way I have it set up also means it was reversed polarity to the driver. It smoked. I dunno what fried but something did. It also blew the LED too. I’m just gonna take all the components off my first one and reflow them to see if I can get it to work.

So for future reference to others, 1S reversed polarity doesn’t get along with this driver :wink:

Good to know! :stuck_out_tongue:

I haven't tried to reverse one, but I think that cerealkiller said that he had reversed polarity to this same IC before and nothing happened. He said he had even measured the LDO output and it also didn't reverse.

As far as the non-working drivers go: I can almost guarantee you that it's your reflow of the MCU, Buck IC, or both. The MCU is usually the culprit.

Hi Richard. Have you done any specific efficiency testing of this driver with an MG-G2 in high, medium and low modes? Alternatively, have you sent any to HJK for testing? Thanks :slight_smile:

Nope, haven't sent any, but you can do a rough efficiency calculation based on the information posted up above.

I built a 17mm MTN-MAX at 2A with the standard inductor [not Coilcraft] and compared it to two zener modded 7135 drivers.

LED: 6V MT-G2
MTN-MAX 2A: 8.4V=1.6A; 6V= 1.2A --- Continuous output. Driver produces far less heat with full cells
QLITE 2.2A: 8.4V=2.5A; 6V=1.5A --- Thermally derates within around 30 seconds w/ full cells
QLITE 1.9A: 8.4V=2.0A ;6V=1.5A --- Thermally derates within around 20 seconds w/ full cells

Thanks. I saw the post and thought I misunderstood something. Would the calculation be (6x1.2)/(8.4x1.6)=0.536 or 53.6%?

The way I understand it, those are two different data points. The driver is outputting 2amps at the emitter. When the input is 8.4v, the input current is 1.6amps. When the input is 6v, the inp[ut current is 1.2amps. You can’t really do the calculation correctly without knowing the forward voltage during his testing, but I guesstimate by just saying Vf is 6v.

So the equation would be (6v*2A)/(8.4v*1.6A) = 89%

.
.
.
I think.
.
I would be more confident, but I feel when input and output voltage are the same (6v) the input current should be much higher than 1.2amps, so I’m probably reading it wrong.

Ahhhhh! This is more what I expected! Thank you. :smiley:

That's correct. Convert to watts, then compare watts in to watts out.

I saw a Cray II running once. There is one at the computer museum in Mountain View, but it doesn’t look so impressive turned off. When it was running, there were streams of tiny bubbles rising from the components. It ran in an aquarium full of freon.
A cooling liquid should be low viscosity and preferably have its boiling point below the highest component temperature. The main practical requirement, though, is that it not affect the components. Water probably won’t do, because it holds ions that make it conduct and it hydrolyzes. The ideal fluid would also have a high heat conductivity, high heat capacity, low surface tension and large thermal expansion.

Can the 22mm version of this driver be used with an electronic switch?

Yes, if it is programmed to use one. The quiescent current drain isn't super high (lower than zener mod), but it still isn't ideal for use as a momentary driver without a lockout of some sort.

I finally read this thread since back when the OP was just getting built. Exciting driver for sure.

Have you let the laser folks at LPF know about this driver? I think it would get a lot of attention. They spend big money on drivers that don't seem nearly as nice as this one. Especially considering the ability to customize the UI.

Does 4 layer mean thicker copper in one layer or 4 separate layers that are isolated from each other for more tighter layout of components?

Thanks RMM. I was going to use it in conjunction with a tailcap switch so it would be locked out. Its a little late now though as the driver is in the mail somewhere. I should of asked the question at the time.

^ I imagine you could wire a momentary switch to the Reset Pin (Pin 1) with just about any FW. I don't know if the OTC feature will work that way though.

It would work, except that you can't even hardly see that pin on that MCU unless it is specifically broken out.

I haven't talked to anyone at LPF about the driver. Honestly, I don't know much about the laser diodes and what specific requirements they may have. I know that they are expensive devices in comparison to our LEDs.

The Oshpark 4-layer boards are the same thickness (1 oz.) on the outside layers, and half thickness (0.5 oz) on the inside layers. The advantage is that you can run a complete pour in the middle layers to take better advantage of thermal vias by using the vias to spread heat directly to the copper in the middle layers which can then more easily transfer the heat into the fiberglass PCB and to the exposed copper around the edges of the driver. I had drivers that were dying after extended use with the two layer board that are still kicking despite lots of extended full power runs without a timer in the same light at the same power level with otherwise the same components. With my thermocouple I measure a 10C-20C difference in the component temperature and a corresponding increase in PCB temperature. I wanted to build a driver that was tough enough to handle lots of use and abuse as long as it had a decent thermal connection to the pill--it had better, given the price of the components alone.

^ I haven't been active at LPF in quite some time. I'll post a link to this thread over there and also to your complete build product page and just tell them I don't know how well it would work with most laser diodes, but it sounds promising. It does sound interesting. I have a couple laser diodes I've been wanting to use. I will probably use this driver for them.

Not sure I understand the 4 layer thing, but I sounds like a very good thing. Thanks for developing it. Definitely the way I will go when I get or make one of these neat drivers.

RMM wrote:

ImA4Wheelr wrote:

^ I imagine you could wire a momentary switch to the Reset Pin (Pin 1) with just about any FW. I don't know if the OTC feature will work that way though.

It would work, except that you can't even hardly see that pin on that MCU unless it is specifically broken out.

Well, hopefully MRsDNF will have enough of the board's contact pad exposed to solder a thin wire to. Momentary switches can use very thin wires.

When I get the driver I'll have a look and see if I can nut out what you guys are talking about. Thanks for the help.