The 40 IO, and LED PWM bitbanging

Shoot; I hadent considered some of these points. Will have to figure it out.

(Prob should have moved this discussion to another thread; will know better next time)

Based on the LED’s that I got ([url]http://www.ebay.com/itm/190541345624?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649[/url]) I’ve calculated the current limiting resistors to be 15 ohms.

If the Fez Supply voltage is ~3.3, the voltage drop is ~3.0 and the forward current is 20mA then the limiting resistor should be 15ohms.

With such a small resistor value, should I even use limiting resistors? I’m currently using 330 ohm resistors.

Few questions…

[ol]Will you be driving the led directly from the fez pin?
Can the fez pin supply that much current?
What happens if the supply is 3.2V or 3.4V?[/ol]

@ Errol,

  1. Yes, directly from the pins. Since only 1 LED will be lit at anyone one time, the current draw through the Fez will never exceed that of 20mA.

  2. Yes a Fez’s pin can supply that much current. I believe it’s 150mA actually.

  3. The supply voltage is regulated so the input voltage should never fluctuate.

I think Gus usually suggests no more than 20-30 mA.

I’d suggest you sacrifice some LEDs and see how low you can go on the resistor at max speed to get max brightness :slight_smile:

I’ve done some more calculations: In order to be able to create the illusion of motion, I would need to do 24 “frames” per second. That means I have 40ms to address all of the LED’s at least once. Therefore each LED would only be on for 0.04 ms :frowning:

Eek.

Even with no resistor I doubt you’ll get any light in that time period :frowning: You might want to consider Charlieplexing each layer and then have a pin to activate each layer. Then you could light 10 LEDs (1 on each layer) at the same time.

@ Ian, It takes 11 pins to light 100 leds. That’ 110 pins in all; the Fez doesn’t have that many. I would have to use 10 Fezes

Actually, two Panda-IIs could handle it. But coordinating them would be a bit of a challenge :wink: Maybe use Decoders to reduce that further?

@ ian, the PandaII has only 40 IO’s on it’s extra. I would need 3 of them to make it work. If GHI still sold Rhino’s then I could just buy 2 of those…

But even if I drive 10 sets, an LED would only be able to stay on for 0.4ms (up from 0.04)… humm.

It has 54 total IO’s. My bad. I was thinking 56…

Well, so far so good I suppose:

Don’t you mean 5x5 instead of 25x25 :wink:

Looking good so far! Those bright blue LEDs may be helping out with the brightness problems I was concerned about.

@ ian, Nice catch and corrected; thanks.

Well it turns out that my Fez Rhino only has 30 Interupt pins; which means a 10x10x10 led matrix is impossible with 1 Fez.
Further, since my rhino has the wiznet module on it, I have even less interupt pins to work with.

It seems that I will have to wait on multiple Fez Ant’s or design my own custom multi Fez Solution to the problem.

Which would you guys prefer: Custom Solution or Wait on Fez ANT?

Is that still an option? Anyone got an update on that board?

Why do you need interrupt pins for output pins?

@ Errol, only interrupt pins can be used as tristate ports.

Well it looks like a Fez Cobra will drive it; toping out at 33 Tristate pins.

Does anyone know if it will come back in stock?

I just counted the number of tristate capable ports that are on the Usbzi 144 chip. There are 42! (What an interesting number)