Author Topic: High power LED driver  (Read 2508 times)

0 Members and 1 Guest are viewing this topic.

Offline LabRat

  • Contributor
  • Posts: 15
  • Country: gb
High power LED driver
« on: January 19, 2017, 10:25:27 pm »
Hi folks,

I'm pretty new to the world of electronics so thought I would post in the beginner section, see if I can get some advice on my current project.  I am busy exploring the world of switch mode power supplies and in particular I am working on a boost converter for high power LEDS (10W Cree U2s), hoping to make my own bike light to replace the cheap Chinese jobs currently adorning the front of my rig (really looking forward to milling the case, going to be a lot of attention to heat dissipation!).

My brief is to run off 2S lipo pack, be able to run 8 LEDs in series at 2A which comes in around 60W.  Starting to realise this is not as easily said as done, plenty of magic smoke!

I have found the TI LM3429 and have used the seemingly helpful TI webbench to populate a schematic as per the attachment.

On the breadboard I have used a beefier TO220 FET and have managed to get my LED string to light for 5-10 seconds before the FET burns a hole in the breadboard.  On my protoboard using the recommended FET at M1(Infineon BSZ100N06LS3) in a TSDSON-8 package.  There is generally a subtle "pop" and then nothing.

Peak current through the inductor is about 10A.  Am I expecting too much of the tiny package or am I missing something more fundamentally important?

Thanks for thoughts and insight.
 

Offline Cupcakus

  • Contributor
  • Posts: 43
  • Country: us
Re: High power LED driver
« Reply #1 on: January 20, 2017, 07:30:05 am »
You're gonna need a heat sink on the FET, and on the LEDs have you done that yet?  FET resistance increases with temperature.
 

Offline Ice-Tea

  • Super Contributor
  • ***
  • Posts: 1499
  • Country: be
    • Freelance Hardware Engineer
Re: High power LED driver
« Reply #2 on: January 20, 2017, 08:55:38 am »
I'm confused? You want to use a relay instead of a FET for a boost converter?

If I got that right: no. Just... no.
eBay shop with all the gear you need!
FS: Agilent 54825A, R&S CMU200, CRTU, Marconi 6201B, Lecroy WP 950, 9354TM, 9354M, 9374L LC584A, Tek THS720P, TDS7154B, Anritsu MG3671A 2.75G I/Q RF gen, Keithley 238 SMU 707 matrix, HP 8642B, 8903A, 8110A, 8156
 

Offline tpowell1830

  • Frequent Contributor
  • **
  • Posts: 750
  • Country: us
Re: High power LED driver
« Reply #3 on: January 20, 2017, 09:32:18 am »
I'm confused? You want to use a relay instead of a FET for a boost converter?

If I got that right: no. Just... no.

That will teach me to read more carefully. :palm: You're right it wouldn't work for a boost converter. I was under the impression he just needed to drive the 60 watt DC load. I have removed my post. Thanks for catching my silly mistake ice-Tea.

PEACE===>T
PEACE===>T
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 5736
  • Country: nl
Re: High power LED driver
« Reply #4 on: January 20, 2017, 01:11:25 pm »
Hi folks,
My brief is to run off 2S lipo pack, be able to run 8 LEDs in series at 2A which comes in around 60W. 
A car headlight full on power will be aprox. a 35W LED equivalent and you want to put 60W on your bikehelmet?
I really hope I will never encounter you in traffic with that lamp and don't be surprised if the cops stop you and impound that device.
 
The following users thanked this post: LabRat

Offline LabRat

  • Contributor
  • Posts: 15
  • Country: gb
Re: High power LED driver
« Reply #5 on: January 20, 2017, 01:20:06 pm »
heat sinking the FET and the LEDs on test are on star boards bonded to a large lump of aluminium.  I will keep playing around and see if I can get a more stable solution.  Is 500KHz high enough to cause difficulties with trace length and board layout (esp on the breadboard)?

Re light power, strictly off road use (think forest and 6 foot drops in a Scottish winter), hoping some decent light output (~100 lumen/watt) will allow me to get some decent video.  Thanks for the comparison though, food for thought as to how much power I actually need, less power would reduce the amount of engineering.
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 5736
  • Country: nl
Re: High power LED driver
« Reply #6 on: January 20, 2017, 01:28:48 pm »
Don't forget to take the lightloss of the lens and enclosure into account.
If you start with >100Lm/W leds that would be a good start and for very good lighting for video you probaly would like a lens with a large angle, so not a spot but better broad light, also for yourself as a biker nice, esp when mountainbiking you would like to see the treeroots and other dangerous obstacles on the side.
 

Offline LabRat

  • Contributor
  • Posts: 15
  • Country: gb
Re: High power LED driver
« Reply #7 on: January 21, 2017, 08:50:40 pm »
I seem to be making progress.  Built a new board up (thanks OSHPark for providing 3 boards a time!), added some extra capacitance on the output (470uF - I misread the original schematic which recommended 3x100uF caps and I was only using one) and dropped my test rig to 6 LEDs and getting a much more steady response.  The FET is barely breaking a sweat now - touch warm after 5 mins operation.

I was getting worried this project was beyond my skills but the stark, blinding glow of high power LEDs has rekindled my enthusiasm  :phew:

Will probably start a new thread under projects to chart my progress.

Thanks again for everyone's replies, has been enough to keep my spirits up  :-+
 

Offline Rick Law

  • Super Contributor
  • ***
  • Posts: 2712
  • Country: us
Re: High power LED driver
« Reply #8 on: January 22, 2017, 01:42:18 am »
Hi folks,

I'm pretty new to the world of electronics so thought I would post in the beginner section, see if I can get some advice on my current project.
...
My brief is to run off 2S lipo pack, be able to run 8 LEDs in series at 2A which comes in around 60W. 
...
Thanks for thoughts and insight.

...
Re light power, strictly off road use (think forest and 6 foot drops in a Scottish winter), hoping some decent light output (~100 lumen/watt) will allow me to get some decent video.
...

If I may first make a non-electronic suggestion...

You may be better off with two separate lights.  Lights and electronics do fail, and home made electronics made by someone (as you put it) "pretty new to the world of electronics" likely has a higher probability of failure.  (Please don't feel I am putting you down, not my intend...  I am a novice myself.)

If you are off-road, and as you put it, forest with 6 foot drops in Scottish winter, a failure will put you in danger.  With two lights, should one fail, you still have some light.  You wont suddenly be in total darkness with one of those 6 foot drops just ahead of you.

Back to electronics, and from another novice.

On that same note, parallel LED's is probably better.  If you are running 2 lights each has 4 in parallel.  Instead of high current driver, now you can use simple single LED driver.  You can eliminate the booster and that may increase efficiency as well.

So, each light has 4x (LED+driver) running in parallel; and two separate such lights.  For simple LED failure, you likely will have 7/8 of your lights to get out of danger.  If it is a power-supply failure, you have 1/2 your lights to get out of danger. 

« Last Edit: January 22, 2017, 01:48:18 am by Rick Law »
 

Offline LabRat

  • Contributor
  • Posts: 15
  • Country: gb
Re: High power LED driver
« Reply #9 on: January 22, 2017, 10:28:03 am »
Hi Rick,

thanks for the electronic and non-electronic advice.  I never ride with just one light - at the moment I have 2 bar mounted and one on my helmet, learned from experience it takes one loose connection somewhere for a light to fail (never trust a cheap battery pack!).  I will certainly look into running multiple smaller units or maybe even 2 drivers in one unit, might allow me to implement some interesting features.
 

Offline tkuhmone

  • Regular Contributor
  • *
  • Posts: 76
  • Country: fi
Re: High power LED driver
« Reply #10 on: January 22, 2017, 12:01:48 pm »
Interesting thread, like too see end results when ready :-) Is this controller using one power level in your case? I understood so based on the input connection of nDIM -pin at LM3429.

I have been also using two lights, one in the helmet mount and main light at handlebar.  I use ready-made solution, and the battery has 7.2V system. Did not have any faults yet, even the lampheads has suffered rainwater, cold winter temperatures, shaking and so on...
Timo, OH7HMS
 

Offline LabRat

  • Contributor
  • Posts: 15
  • Country: gb
Re: High power LED driver
« Reply #11 on: January 22, 2017, 03:55:04 pm »
The plan is to have the nDIM pin PWMd by a micro (KL05Z) with presets for "climb", "traverse" and "descend" (button selectable), the micro will keep an eye on battery charge (did think of a proper fuel gauge but thought this was overdoing it.  To try and reduce battery consumption and aid with thermal management the light will automatically dim to a predetermined level when the rider has stopped, was going an accelerometer to determine this.  This project has already suffered a bit of "scope creep", maybe ready in time for *next* winter!
 

Offline tkuhmone

  • Regular Contributor
  • *
  • Posts: 76
  • Country: fi
Re: High power LED driver
« Reply #12 on: January 22, 2017, 09:25:13 pm »
ok with comments,,,,

What I know for some brands of cycling led lights, at least Lupine uses battery capacity check circuit (built in, to a smartcore battery line). And acceleration is detected in the red tail light (Rotlicht), for brake light functionality...
Timo, OH7HMS
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf