FREE DIY open source BOOST driver!!! Tested & working!!

Thank you very much!

I don't think there will be any buck, we already have two great and cheap buck drivers made by members here (x-drive and blitz buck)

There may be a few drivers added to this thread from my personal collection later on. Not until this driver is done though.

I'm in the middle of reflowing the newest version right now! Probably won't get results up until morning though it's getting pretty late

It lives! Drawing 1.9A to make 1.14A on a 4 diode test load

Runs over 1 minute with no heatsink! (aside from the built in one)

Full report and pics coming tomorrow I'm going to bed.

Edit: just ran it for 5 minutes straight with no heatsink, didn't go into thermal protection! Now I'm really going to bed

benmwv said:

Thank you very much!

I don't think there will be any buck, we already have two great and cheap buck drivers made by members here (x-drive and blitz buck)

Click to expand...

You are absolutely right, we do have those two great drivers, and we do have an opensource buck/boost driver by laserluke.

Perhaps LPF community could also benefit from an opensource buck driver. It would not only be educational for beginners but also increase the variety and perhaps through the group effort we will create an even better drivers than currently available.

Open Source concept strongly appeals to me (and I believe I'm not alone in this) and given the choice I'd rather buy opensource unit instead of proprietary one (no pun intended to foulmist, wolman, lazeerer and others developing those awesome units).
I do realize and appreciate ALL of the effort that goes into developing drivers, whether opensourced or proprietary!

Just a thought my apologies if I'm being inconsiderate.

I probably won't, but if I ever do an open source buck it would have to be very different from x-drive and blitz buck. I definitely wouldn't want to use the same IC as either of them, and I wouldn't want to steal their business. This driver was made because there were no good alternatives to using flexdrives for single cell builds, and lots of people were using dual flexdrives to get 1.8A. That costs around $60 and the customer service is crappy, so there was a definite need for something cheaper.

I've been playing around with this benboost mini and it works great!





I decided to try a 4.7uh inductor, and it works fine! This would not work before, I suspect it has something to do with the oscillations caused by the ld+/- cap. That cap has been taken out on this version. I am using a 22uf inout and two 22uf between LD+ and ground.

This was a pretty easy board to reflow too. It has a DFN but there is a lot of space between the pins. I think the pitch is 1mm? Its not hard at all, but you have to use the paste sparingly. However much you think it will take use about 1/4 of that!

For the built in heatsink I cut a strip of copper as wide as the heatsink pad from a pre-1982 penny, sanded it shiny, folded it over a few times, and squished it flat. Then I put some solder paste on the heatsink pad and stuck it on just like it was one of the components I also put a few dabs of paste on the sides to get sucked into the little crevices left from folding it for better heat transfer. It ended up being exactly the same height as the inductor. Its very important that it is at least that tall so you can glue it onto the heatsink in your host.

The board works great:



I ran it for five minutes straight just like that (no extra heatsink) and it never overheated. After five minutes it had dropped to 1.131A because the sense resistors heated up, 16ma lower than where it stated. The input current is also under 2A, about 2/3 of the max. Theoretically we could get another 600ma out of it :eg:, Ill try upping the current later.

I put it on the scope and its not too bad:



AC coupling, each division is 20ma high, 500ns long.





AC coupling, each division is 10ma high, 100ns long.
So the ripple has spikes about 27ma high. Should be safe for any diode we use.





And this is what the waveform looks like when you've ran it so long the battery has drained down to 2.7v and it can't regulated current anymore. It overheated shortly after this. It must have ran at least 10 minutes straight while I was taking pictures and messing around with the scope.

That's awesome. What's the cost of this one compared to the original benboost?

Could probably get cost WAY down if it was sourced to China for production and assembly instead of us getting digikey and mouser parts

Awesome work Ben,

I tried once the LM3410 in constant voltage mode. I was able to get 1.8A out of it at 6V output. I guess it is more efficient in constant voltage but I think you can get a tad little more current in constant current mode too

Give it a shot and let us know :beer:

It was 6.56 plus set resistors. I used some of the 100s that you gave me, but it should be about $7 if you had to buy them too. Im sure the price would go down a lot in higher quantity, the lm3410 chip is half of the entire part cost.

I think it will be able to utilize the full switch current now, which it couldn't do before without overheating no matter how much you heatsinked it. Right now I've got this one set with two 0.33 ohm resistors and I'm going to add another 0.33 which should put me around 1.7A. I'll try that tomorrow and see how it goes.

Also I know there was a problem with the 9mn 445 diodes higher voltage. Having two paralleled drivers at 1.1A didn't work because of the extra voltage the 9mm diodes need. I tested this one with my test load maxed, all 6 diodes, and it still ran fine. That's a bit over 6v load, so the 5v diodes should work just fine.

benmwv said:

I probably won't, but if I ever do an open source buck it would have to be very different from x-drive and blitz buck. I definitely wouldn't want to use the same IC as either of them, and I wouldn't want to steal their business. This driver was made because there were no good alternatives to using flexdrives for single cell builds, and lots of people were using dual flexdrives to get 1.8A. That costs around $60 and the customer service is crappy, so there was a definite need for something cheaper.

Click to expand...

Ben,

If you can release something similar or better and want to, I bet there will be a place for your opensourced buck driver. One more driver will not hurt anybody's business. Instead, it will be beneficial to all, in terms of variety and increased competition on closed source drivers. I think If you organize a thread for that driver with donation button and preliminary designs there will be more than enough willing people (myself included) to support you.
Just a thought.

Sent a bit more to support future development.

If at-least a fraction of those who benefited from Benboost drive, one way or another, could donate...

Thanks! Very generous of you to donate again

What IC is this, not the 3410 I assume?

Same IC just in a different package. It does WAY better with the thermal pad and on-board copper heatsink.

Here:
LM3410XSD/NOPB National Semiconductor | LM3410XSDCT-ND | DigiKey

benmwv said:

Same IC just in a different package. It does WAY better with the thermal pad and on-board copper heatsink.

Here:
LM3410XSD/NOPB National Semiconductor | LM3410XSDCT-ND | DigiKey

Click to expand...

Oh, I see! I thought maybe this was a whole new design. Interesting. There are so many considerations in a design that you don't even think of some of them without a lot of experience. For instance, I would never have thought that IC package might affect whether I could use the full current of an IC. But that is pretty neat that by using a package with a heat-sink pad on the bottom that you basically pick up improved performance vs the same chip in a less heat sink able package!

Of course I still have about 40 of the LM3410's in the old package!

Yeah, that thermal pad helps a lot. The built in heatsink really works good too. It's way more effective than heatsinking the top of a plastic package. I can feel the copper heatsink warming up a lot more and faster than the plastic top of the IC so I know its moving heat far better. RHD gets the credit for that idea, I saw him use it in another driver we worked on and thought it was great so I put it in here!

The other package lm3410's will still be great for up to around 1A, but I will be using this design for anything 1A+.

Big thanks to Blord for his donation!

If anyone wants good priced IC's for this driver you can get them ( they are legit ) for great prices since the original link of them seems to be out...

The DFN ones are 1.9$ US or the sot2235 ( original ) is 1.5$ US.
Contact her anvi@leadingchip.com

Btw I'm not making any refferal money or any benefits for this it's just to help out since the ic is the most expensive part at 3$ +
:beer:

Ben, how'd you reflow the dual sided board without the components falling off out of curiosity?

Thanks!