From: austin on
nishad,

Why do you doubt? If the data sheet says an IO may sink 8 mA, then
why not ALL IO at 8 mA?

No problem.

The only concern here is ground/Vcc bounce: yes, 128 * 8mA is a lot
of current, and you should make sure your pcb has good power and
ground planes, and the recommended bypass capacitors.

Austin
From: Gabor on
On Nov 12, 10:22 am, austin <aus...(a)xilinx.com> wrote:
> nishad,
>
> Why do you doubt?  If the data sheet says an IO may sink 8 mA, then
> why not ALL IO at 8 mA?
>
> No problem.
>
> The only concern here is ground/Vcc bounce:  yes, 128 * 8mA is a lot
> of current, and you should make sure your pcb has good power and
> ground planes, and the recommended bypass capacitors.
>
> Austin

On the other hand if you really need 120 I/O's you may want to
consider
breaking it up into two packages. It might even save board area over
one very large package. Of course that depends on how may additional
pins you use to split the design in two. Sometimes it comes almost
free, for example a two-wire serial bus. 1 Amp of switching current
in a large PQFP results in lots of ground and Vcc bounce due to the
large package lead inductance. Generally speaking, when using quad
flat packs, the smaller the better because the die size of the chip
inside gets closer to the overall package size as you get smaller,
so the lead inductance is smaller, too. I'm not sure how it works
with mature CPLD devices, but for FPGA's at least the price per I/O
is better in smaller devices as well.

Regards,
Gabor
From: -jg on
On Nov 13, 2:08 am, "nishad" <abnis...(a)gmail.com> wrote:
> My requirement is to replace fifteen 7 segment display drivers using cpld
> logic.
> Total I/O connected to display will be 15*8=120.
> Each pin has to sink around 8mA, so Im planning to go for Xilinx XC95288XL
> CPLD.
> My doubt is that can the cpld sink 120 lines of 8mA simultaneously?

A XC95288XLis an expensive shift register ? :)

or is it doing other things as well ?

LED drive is generally better done with smaller, daisy chained CPLDs,
or even just shift registers.

-jg

From: nishad on
>nishad,
>
>Why do you doubt? If the data sheet says an IO may sink 8 mA, then
>why not ALL IO at 8 mA?
>
>No problem.
>
>The only concern here is ground/Vcc bounce: yes, 128 * 8mA is a lot
>of current, and you should make sure your pcb has good power and
>ground planes, and the recommended bypass capacitors.
>
>Austin
>

Thanks a lot for the reply...

---------------------------------------
This message was sent using the comp.arch.fpga web interface on
http://www.FPGARelated.com