Rotating LEDs revisited

This is the place for any magazine-related discussions that don't fit in any of the column discussion boards below.
Post Reply
Bigglez
Posts: 1282
Joined: Mon Oct 15, 2007 7:39 pm
Contact:

Rotating LEDs revisited

Post by Bigglez »

Last month there was a thread about LEDs arranged in a
circle, more info here.

I happened to be doing a "thought exercise" recently, to drive
a cluster of bi-colour LEDs. The tricky part (for me at least) was
coding sixteen PWM channels with 127 level resolution, and
keeping the refresh rate above about 40Hz. To reach this level
I had to increase the uC clock to 20.0MHz (its rated max).

The second hurdle was finding a way to 'linearize' the steps.
In reality it means the opposite! The human vision requires a
higher power transfer function, much like audio controls require
'log taper'. I implemented this for the 16 LED channels with an LUT
(look up table) and experimented with x^1.1 through x^2.5 values,
while reducing the steps to 64 levels (instead of 128) to keep the
overall refresh rate above visible flicker.

One thing lead to another and I put together a very quick SBB
circuit with sixteen LEDs (single colour to start), two SRs and a
tiny261 AVR uC. Youtube link

Here's a better quality still PIX
Image

From this prototype, which sat on my tech bench for a few
weeks, I decided to place the just 12 LEDs in a circle based
on the previous thread noted above. The placement of these
LEDs appears to be quite critical, as you can see from the first
video clip of a linear row of LEDs. (NOte the wiggle as the LEDs
are not aligned on the SBB). Youtube link

Here's a better quality still PIX
Image

The video is not optimum, meaning this project looks spectacular
in real life. The rotating single LED 'dot' has a comet tail of four
LEDs at lower brightness.
Post Reply

Who is online

Users browsing this forum: Majestic-12 [Bot] and 73 guests