Um... I DID IT!!!!!!!!!! YESSSSSSSSSSSSS!!!!!!!

I kept the components as minimum as possible and simple. I got a board with 16.6x15.5 size.. But my final pcb should end up a little smaller than my protoboard (which is made by hand).
No pot for current settings. Fixed resistors are used instead. For pot I need more components on the board which are unnecessary!

so ..

specs in boost mode :

constant current output
2.1A max at Vin=>=3.6V & Vo<=5.0V
1.5A at Vin>=3.6V & Vo=6V
input - you should get one of those 2800-3000mAh 4.2V Lithium Cells I am waiting for mine from Larry to test it with battery instead of bench PSU but as long as the battery can deliver 4A it's good!
CHEAP CHINESE BATTERIES WON'T DO! ALREADY TESTED WITH TRUSTFIRE - PROTECTION KICKS IN! ALSO TESTED WITH NON-PROTECTED CHEAP ONES.. 1A at the output max

To regulate the output at the 1.8A current - a 3.6V is desirable for minimum input when using 445nm diode.
as the batteries starts discharging 3V will still regulate the current with around 300mA lower if the battery can handle the current.

specs in buck mode:
constant current output
max output tested 2.5A @ 4.5V (may go to close to 3A not tested)
max input in buck mode should be 5.5V (6V is absolute maximum - not recommended)

common specs:
input voltage - 2.9-5.5V (3.6V for ~1.8A@4.5Vout)
max output voltage - 6V (but current will drop)
when heatsinked it can run nicely for long periods of time.
overvoltage/overtemp protection

After consistent test on my test load with 2,3,4, and 5 rectifying diodes
At 6V(max) output the max output current drops slightly. ~1.5A in boost
At <=5.5V output current is steady 1.75A (as set) in boost!

At 3.3V output the current can go above 2A in boost and around 3A in buck!

I hooked my diode straight on the output with risk to kill it but I had to try it in REAL conditions. I put my DMM in Amper meter mode and hooked it between the + of the driver and diode. The big direct press heatsink is especially made for this test

In some of my tests the diode survived a 1.95A current. O.o good thing I didn't set it higher the driver was in buck mode. In boost mode I have tested it til 2.1A. It might go higher but it's unlikely.

I spent quite some time trying different inductors and this pretty much is the optimum one and size is PERFECT too.

I present a video of it in action with bench supply, A140 laser diode and current and voltage readings of input and outout. I simulated the battery voltage drop down to 2.8V and up to 4.2V.


I will update this thread with a battery powered operation when I get the... battery itself maybe after new year... (post companies.. shhish )

So far no glitches at the output. Driver is steady. Diode still lives. I am happy. Happy holidays to all



p.s. I will make a few commercial boards after new year if anyone is interested. I can tell you the price after I do some math and figure all the costs.

Incredible! You just killed my market :P Good job and good luck!

4.4V output at 1.75A with 3.6V 3.44A input gives you ~58% efficiency.. are my calculations correct?

So this is only able to be put in to buck OR boost mode at any one time? Just wondering....

Also what size SMD's are you using? They look pretty large. Remember when you scale it down your heat dissipation becomes a bigger issue.

isn't that 62%... at 5.5V its around 77%... hmm I may need to make some improvements.


Um no It can work boost/buck at the same time.
The sense resistors are 1206, caps are 0805, I have tried 0805 resistors, work ok, but I had bigger for that current.

I hope you don't mind me asking if this is a TPS63020 based driver. I did one that is 17.7 x 13.7 mm. Mine is single sided and has a solid ground plane on the back side. I also have open circuit protection, low battery protection, and reverse polarity protection.

I have yet to do full tests but I am currently using this driver to push 1.3A into an M140 diode. I did a second board revision and have sent it off for to be made. I intend to post after new years with a build report for this driver and tests.

Hearing the results you got with this driver (if you are using the TPS63020) gives me lots of hope that I can get mine to work well.

yes It's the TPS63020, it has solid ground on the back side as well
open circuit, low battery protection are both BUILT in the IC

what kind of efficiency do you get... I have yet to determine why mine is so low .. I should be getting around 80-85%....

I think the inductor here is the main reason for such low efficiency... I will order more parts to try out. the caps might not be enough too.... I will try bigger values. and possibly lower ESR

If anyone has any suggestions please share

I am SO glad to see that you're going fixed-resistor. I'm not a fan *at all* of pots.

Congrats, this is fantastic!

yeah I don't like the pots too - way to easy to screw up
I just ordered some more components. I have to make this at least 80% efficient

There is an old thread where someone used this IC i cant remember it. If I can find it ill shot you a PM foulmist. It might help out some with the inductor.

Nice work.


Its an Expensive driver though.

thanks bro, yeah I already gave a lot o-money just for trying different parts
I just ordered 5 different inductors and other caps to try out which sent me back 20 bucks

4.2A at 2.6V. I can do dat!

Keep up the good work! thanks for the vid and showing what the current draw is for the Li-Ion range of voltages.

It might take a while til I receive the new parts. Xmas and New Year and my supplier informed me that they are out of some of the items.... oh shoot .... I hope I can get them soon...

I got about 70% efficiency at 2A out if I recall. About the low battery protection: I used a TC54 with a threshold of 2.9V on the enable pin of the TPS63020. The TPS does have low voltage protection, but the cutoff is too low. By the time it turns off, your battery may already be destroyed.

How do people normally handle low battery protection on these? Maybe everyone uses protected cells? I am using some 26650 cells from battery space (the 4AH ones) with no protection and these work great.

Back to the TPS63020 driver. My inductor is on the large size. It is rated for 4.5A continuous and 7A saturation. It is part number DR74-2R2-R.

Mine is rated 5.6A and it is better than the others I tried but I know there is a better choice. I just have to find it

about the 70% - that's low too.. 85% can be achived at 3.6Vin and 4.5V@2A out! That will be my goal!

I guess protected cells should be used instead of non-protected with this IC.
I can't see a way to modify the undervoltage lockout threshold - it's set to 1.6Vmax. It can be made with sensing the input voltage and controlling the EN pin but that will take way more space.

WoaH !! Dude, You did it.. Now thats some driver your making ..Whats the Official Launch date?!

no official date yet.. I have to figure out why efficiency drops. I did some tests with the test load efficiency is 20% higher at the same voltage and current. O.o but when I hook it up to the diode efficiency drops?

The silicon diodes just give a reference idea of the test load , in order to simulate it you should try a 445 diode which is previously damaged into led.. iirc it is the best way to measure!

I use my 445nm laser diode which is alive for the real testings

Thats somewhat not a wise thing , not recommended either, why would you use a perfectly fine diode as a test load, Things are most likely to go wrong when prototyping.. there is a 50 50 chance that the prototype works according to calculations, almost everytime somethings missed out or is calculated wrong and PUFF! the diode flickers and there it goes..becomes "led".. i never do that to an alive diode, Its almost like killing a living thing :P !!

Lol I guess I didn't say it right.

I first always do tests with TEST LOADS(I even use incandescent bulbs 5.2V as test loads). Also, I check the output on Oscilloscope. After I see a steady output I do the real tests with an alive diode. I can risk the diode but if I don't make that test I cannot be 100% sure it works.

I'm sure foulmist is perfectly capable of evaluating whether he wants to use his live diode to test a boost driver once it has obtained the basic test-load's approval

And where did this "50/50 chance of working as per calculations" figure come from? Is that in the datasheet?

foulmist;

I would not worry about the low voltage cutout.

I have tested U.L. name-brand cells repeatedly to that level, without any cell failures.

Overall driver efficiency improvement is desirable.

LarryDFW

yep, thanks rhd

I figured out why my efficiency was so low - the DMM as A meter in series with the diode to check the output current was messing it. Without the DMM I got 75%, which has to be improved further but that's a LOT of relief. It's one thing to go from 60% to 85% and another is 75% to 85% .

and btw people you can refer to me as Tom ..

-Tom

Yes Offcourse you do! I Just thought the way you are saying implied that..
But my bad!!
Well we must congratulate him nevertheless, he's making something "urgently needed" for higher power 445's and i envy him!!
50/50 is not actually the same all the times, sometimes you get lucky..and sometimes it strikes back so just a stereotype for things that dont go right everytime..
P.S: Is my english too bad!!? lol Cheers!