[ell-i-developers] Re: "LeBulli" Led shield for PoElli

  • From: Pekka Nikander <pekka.nikander@xxxxxx>
  • To: ell-i-developers@xxxxxxxxxxxxx
  • Date: Sat, 21 Jun 2014 09:38:02 +0300

Quick thoughts on some of the points below:

> I'm doing a schematic for a high power led driver shield for Ellduino, it 
> resembles a bit the schematic we have for hackathon. .... Since selected LED 
> has low forward voltage, buck topology is used.
> Design has following assumptions, please correct me if I'm wrong:
> 1) 1 A of current can be drawn from 5V pin, which comes from PoE SMPS.

For Ellduino, I would not recommend.  It was not meant to be used in that way, 
for any larger amounts of current.  I don't remember the details of the 5V 
supply, but I wouldn't be surprised if it tops in less than 1A.

For the Nucleo and RasPi shields, the situation is different.  Timo has tested 
that the design does provide 2.5A 5V, so within those limits it can be used.  
You have to count for the 5V used by the MCU, though.

And, as you know, in the long run we should go for a forward or flyback 
topology where we regulate directly from the PoE nominal 48V.  

> 2) Device can receive and transmit information through ethernet port

Yes, of course.

> 3) There are 3 open-drain PWM  pins available.

I have to admit what the open drain here means, but otherwise yes.

> 4) Core can run at 48 MHz, so 8-bit PWM can be run at 48 MHz / 2^8 = 187,5
> kHz

Yes.  And that should be enough for the kinds of bucks we are considering at 
this phase.  At the very low end, say below 10% or 5% duty, due to nonlinearity 
of LEDs and human light perception, we might also want to lower the switching 

> 5) Pinout of Ellduino is correct (please check this if nothing else)

I don't think this matters, since I don't think we would be using this design 
with Ellduino but in the very beginning.  I think it would be better to check 
the Nucleo pinouts.  If possible, of course a pinout that works both with 
Ellduino and Nucleo would be nice.  For ADC channels it should be easy.  For 
PWM, the situation is somewhat trickier.

> I'd also appreciate recommendations for pins to use for PWM. Open-drain
> ability is a plus but not necessary. Even better would be to have phase
> difference between PWM-signals to avoid unnecessarily high ripple current
> on 5V pin.

Based on my experience such phase shifting is more or less a must at those 
frequencies, especially if the PoE PSE is an 802.3af compliant one.

> Do you think I should add power arbiteration from Vin pin of Arduino Mega
> for Arduino-friendliness? OTOH, Arduino Mega cannot reach required PWM
> frequencies, so using a lot larger chokes and capacitors would be
> required.

As I wrote above, I would design for a Nucleo PoE shield + Nucleo combination.


Attachment: smime.p7s
Description: S/MIME cryptographic signature

Other related posts: