FPGA Central

Hi everybody,

I have designed a Cardbus Card with Spartan 3 and Xilinx PCI Core,
which works very well in all kind of notebooks, but when I try to use
the card within a PC (with a PCI CardBus adapter card - TI PCI-1520
chipset) nothing happens not even the new hardware window comes up. But
the PCI CardBus adapter seems to be OK, because if I use a bought 3COM
Ethernet Card the new hardware window appears and the driver will be
installed. The PCI Cardbus adapter card has exactly the same chipset as
one of my test notebook, where the card is working very well. I have
also tried to set the boot clock (CCLK) to maximum speed - but the
result did not change. I do not really see the difference between a
chipset mounted directly on the motherboard and the adapter card.

I hope somebody has an idea to solve this problem.

With kind regards
Nico Presser

Are you sure you're getting your required voltage(s) from the
cardbus adapter? If your requested voltages are not available
you normally don't get powered up at all. I've found that 3.3v
is the only voltage you can reliably expect on a cardbus adapter.

HTH,
Gabor

Weltraumbaer wrote:
> Hi everybody,
>
> I have designed a Cardbus Card with Spartan 3 and Xilinx PCI Core,
> which works very well in all kind of notebooks, but when I try to use
> the card within a PC (with a PCI CardBus adapter card - TI PCI-1520
> chipset) nothing happens not even the new hardware window comes up. But
> the PCI CardBus adapter seems to be OK, because if I use a bought 3COM
> Ethernet Card the new hardware window appears and the driver will be
> installed. The PCI Cardbus adapter card has exactly the same chipset as
> one of my test notebook, where the card is working very well. I have
> also tried to set the boot clock (CCLK) to maximum speed - but the
> result did not change. I do not really see the difference between a
> chipset mounted directly on the motherboard and the adapter card.
>
> I hope somebody has an idea to solve this problem.
>
> With kind regards
> Nico Presser

Hi Gabor,

As far as I know CardBus does only support 3,3V. In older PCMCIA
systems you could choose between 5V and 3,3V. Nevertheless my card is
based on 3,3V.

Nico

Weltraumbaer wrote:
> Hi Gabor,
>
> As far as I know CardBus does only support 3,3V. In older PCMCIA
> systems you could choose between 5V and 3,3V. Nevertheless my card is
> based on 3,3V.
>
> Nico

Actually table 3-1 "Card Detect and Voltage Sense Connections"
in the PC card standard shows that Cardbus cards can request
other voltages on the Vcore/Vpp pin. The main power is always
3,3V but you can for example jumper CVS2 to CCD2# and ground
CCD1# and leave CVS1 open to request 1.8V Vpp without the
possibility of change due to CIS configuration settings. I tried this
with a PCI to Cardbus adapter and it did not power up my card
because it could not supply 1.8V

After grounding CVS1 the card powered on normally. I did not
actually use the 1.8V, but my device was PCI and not stricly
CardBus compliant, so I was hoping the fixed voltage strapping
would prevent troubles from attempting to read non-existent
cardbus configuration registers.

You are right. But in practice I have never seen a Card Bus controller
chipset that supports other then 3.3V/5V.
By the way - during burst transfers the ac portion of the 3.3 supply
voltage does really rock of about +-200 mA pp composed of all possible
pc frequencies. Depending on the notebook the voltage swing is very
different, as well as the frequency spectrum, so I do not believe that
crosstalk on my card is the reason. But the card shell be used in
different notebooks and the FPGA directly needs 3.3V, so I am also
looking for a well working dc-filter (-circuit) or something else to
reduce the voltage swing to a minimum. The old known things like filter
capacitor, pi-filter or series inductor I have already tried, but the
results were not satisfying me. Maybe you or someone else has had
similar problems and some hints how to get manage.

With kind regards
Nico Presser