Adam Wang wrote:
> Although this thread is quite earlier one, do these estimation of current
> above influence to lastest M1R4 we have now ?

With the way you've populated the matrix, the maximum current
would be 3*If for rows 0 and 1, and 2*If for row 2.

With 270 Ohm resistors, we should get an If of about 5.3 +/- 0.5 mA,
assuming Rds(on) = 0 Ohm in the FPGA. (I'm not sure how exactly
drive strength translates into Rds(on), etc.)

Since it's effectively a 3x3 matrix, we could increase If up to a
nominal 8 mA. With 220 Ohm resistors, If would become about
6.5 +/- 0.5 mA.

But what's the objective ?

If you want to avoid adding 270 Ohm resistors to the BOM (we
already have 220 Ohm, but no 270 Ohm), then that's probably fine.
Real-life current will be a bit below the estimates anyway and I
doubt the FPGA's I/O drivers would blow up even if someone added
LEDs in the fourth row, which would bring the maximum nominal peak
current per row to 24-28 mA. The maximum average current (over
>= 1 ms) would still be only a third of that, due to the
multiplexing.

If you want to maximize brightness, it would probably be more
effective to arrange the LEDs in a 2x4 matrix and scan only two
rows. This would yield a 50% increase of maximum duty cycle for
each LED, compared to ~22% when increasing the current.

If you want to minimize strain on the FPGA, the 270 Ohm will be
perfect, keeping current levels well below specified per pad
limits. (I haven't been able to find per bank limits other than
the SSO limits starting at pages 36 of DS162.)

- Werner
_______________________________________________
http://lists.milkymist.org/listinfo.cgi/devel-milkymist.org
IRC: #milkymist@Freenode

Reply via email to