portable adjustable power

17 views
Skip to first unread message

Anton Kast

unread,
Jun 9, 2013, 1:27:13 PM6/9/13
to led-...@googlegroups.com
My wife likes this wearable light I made her: http://www.youtube.com/watch?v=L5E-lXaf9ik  Her only request is that she would like to adjust the brightness.

There must be standard answers, but I haven't found them.  The thing gets up to about 30W at 3.7V (7A).  I've found lots of motor speed controllers for model cars, but they work at higher voltages.  Low-voltage drivers I've found are for much lower powers.  Reducing the brightness programmatically works, but it degrades the dynamic range of the effects.

Surely someone has dealt with power control at such levels in lithium-ion powered wearable lights.  Any advice?

Anton Kast

EkriirkE

unread,
Jun 9, 2013, 3:45:00 PM6/9/13
to led-...@googlegroups.com
Are you regulating this at all right now, or driving straight off a 1S LiPo (w/ resistors)?  I would look into an adjustable buck converter if looking to reduce voltage.  Or are you looking for a PWM option that runs a faster freq than you currently run the fades programatically?

Anton Kast

unread,
Jun 9, 2013, 4:39:57 PM6/9/13
to led-...@googlegroups.com
This particular project is one long bus of WS2801 pixels powered directly by a LiPo pack.  There are no resistors in the power path, nor should there be at such high power.  The signal to the pixels comes off a separate, regulated 5V supply.

The effects in this project involve lots of subtle brightness variations.  These variations are not produced by PWM but by commanding the WS2801's.  Since the resolution of the WS2801 is only 256 values per channel, scaling the brightness in software loses a lot of the subtlety.  This must be a common problem.

A buck converter would be great, if there were one that worked with the low-voltage supply and could sustain the high power.  I have failed to find one.  Does anyone know of such a component?


--
You received this message because you are subscribed to the Google Groups "LED Nerds" group.
To unsubscribe from this group and stop receiving emails from it, send an email to led-nerds+...@googlegroups.com.
To post to this group, send email to led-...@googlegroups.com.
Visit this group at http://groups.google.com/group/led-nerds?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

EkriirkE

unread,
Jun 9, 2013, 5:25:07 PM6/9/13
to led-...@googlegroups.com
Ah, I though you wired single green LEDs and were driving them yourself, those do do PWM internally...  I understand, what I meant was on top of any PWM you(or the WS2801's internal) do, you could pulse the power line at a much higher rate so as it doesn't interfere/harmonize with whatever rate(s) the inherent dimming uses.  I remember reading something years ago where someone wanted to do just that with one of these self-contained ICs and found a frequency which didn't cause the ICs to reset while "off" and it worked.  So you got the scale(s) of the IC with the additional "master" PWM.  All the bucks I see are rated ~4V input, I'll do an experiment with one here and see what I get off a 1S lipo

googlegr...@slkgroups.us

unread,
Jun 11, 2013, 11:25:58 AM6/11/13
to led-...@googlegroups.com


On Sunday, June 9, 2013 1:27:13 PM UTC-4, Anton Kast wrote:
My wife likes this wearable light I made her: http://www.youtube.com/watch?v=L5E-lXaf9ik  Her only request is that she would like to adjust the brightness.
<SKIP>

Anton Kast

Anton- great idea!!! love the design. do you have any pictures of the backside or the internals? as far as dimming- since they are basically a microprocessor-based unit, they cannot be dimmed by lowering the voltage. you need to adjust your program to dim them. maybe you could post your code so we could look at the how behind the lighting? there might be some features you are not using that we could point out?
or- maybe you have done anything i could think of- but some one else might have other ideas...
again- great project- keep us posted@!
Keith Digi-dome
 

Anton Kast

unread,
Jun 11, 2013, 12:55:38 PM6/11/13
to led-...@googlegroups.com
Here are some details about this project: http://sonic.net/~rat/cuirass/

Since you asked, I put the source code up at https://github.com/AntonKast/cuirass  It's crap, but it runs.  Please be gentle.

Keith, I'm not sure you're correct that addressable pixels can not be analog-dimmed.  Each has separate pins for power and for data.  Maybe I'm doing it wrong, but I just connected the data and clock to 5V (microcontroller-driven) and the power to 3.7V (direct to Li-ion batteries) with a common ground.  I had to shunt the power around so the more distant lights wouldn't turn red, but otherwise the arrangement works great even as the batteries fade.

Tekket, thanks for the pointer to FastSPI_LED2.  I have not tried it, but I have tried scaling brightness in software.  It certainly dims everything, but as you can tell from some of those videos, I'm aesthetically interested in probing brightness variations.  Digital scaling throws away dynamic range at the low end of brightness.  At fixed supply voltage, the difference between "1" and "0" remains the same perceptually, and so software scaling just throws away dynamic range.  It makes things less smooth and more blinky as pixels get closer to black.

It sounds like maybe there isn't a good pre-packaged device to step down current efficiently at such power and voltage levels.  If I come up with a design, I'll be sure to post it back here.

Anton


--

Daniel Garcia

unread,
Jun 11, 2013, 2:14:32 PM6/11/13
to led-...@googlegroups.com
You may want to check out the TM1829's then - which I have a strand of, but won't get supported in FastSPI_LED2 for another few weeks - but in addition to PWM based brightness, it also does 32 levels of current based (e.g. no PWM, no blinky) brightness control.  I can't wait to play around with it!

--Dg

Digi-Dome

unread,
Jun 11, 2013, 2:28:31 PM6/11/13
to led-...@googlegroups.com


On Tuesday, June 11, 2013 12:55:38 PM UTC-4, Anton Kast wrote:
Here are some details about this project: http://sonic.net/~rat/cuirass/

Since you asked, I put the source code up at https://github.com/AntonKast/cuirass  It's crap, but it runs.  Please be gentle.

Keith, I'm not sure you're correct that addressable pixels can not be analog-dimmed.  Each has separate pins for power and for data.  Maybe I'm doing it wrong, but I just connected the data and clock to 5V (microcontroller-driven) and the power to 3.7V (direct to Li-ion batteries) with a common ground.  I had to shunt the power around so the more distant lights wouldn't turn red, but otherwise the arrangement works great even as the batteries fade.

Tekket, thanks for the pointer to FastSPI_LED2.  I have not tried it, but I have tried scaling brightness in software.  It certainly dims everything, but as you can tell from some of those videos, I'm aesthetically interested in probing brightness variations.  Digital scaling throws away dynamic range at the low end of brightness.  At fixed supply voltage, the difference between "1" and "0" remains the same perceptually, and so software scaling just throws away dynamic range.  It makes things less smooth and more blinky as pixels get closer to black.

It sounds like maybe there isn't a good pre-packaged device to step down current efficiently at such power and voltage levels.  If I come up with a design, I'll be sure to post it back here.

Anton
Anton- great idea!!! love the design. do you have any pictures of the backside or the internals? as far as dimming- since they are basically a microprocessor-based unit, they cannot be dimmed by lowering the voltage. you need to adjust your program to dim them. maybe you could post your code so we could look at the how behind the lighting? there might be some features you are not using that we could point out?
or- maybe you have done anything i could think of- but some one else might have other ideas...
again- great project- keep us posted@!
Keith Digi-dome


Anton- be gentle??? that was SO Cool!!!!!! and- you went to the Big Burn? very cool. A very appropriate execution of playa-wear. I am jealous.
now-= you mentioned that you tried software dimming- did you (as was my first thought) simply scale down all values? or- (my second thought) did you set a ceiling to all values output? (as in- maxvalue= outvalue, 127, maximum) (just a guess on commands)) i think the programming was great.
my third thought is to replace 255 (that i saw in a few places) with MaxValue and make this a global variable. then, maybe a button routine that would check for a press and change MaxValue - maybe dim up and down 'live' until button released?

again- forgive me if i am being too forward. i LOVE the project and plan on stealing a few bits of code for a Dome I am working on.
Where are you located? there is a regional not too far from us and we are planning on going in the Fall.
best regards,
keith

 
Reply all
Reply to author
Forward
0 new messages