[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Universal PCI signaling
- To: Mailing List Recipients <firstname.lastname@example.org>
- Subject: RE: Universal PCI signaling
- From: Andrew Ingraham <Andrew.Ingraham@digital.com>
- Date: Thu, 15 Oct 1998 23:29:33 -0400
- Cc: "'pci-sig'" <email@example.com>
- Delivered-To: firstname.lastname@example.org
- Resent-Date: Fri, 16 Oct 1998 14:12:01 -0700
- Resent-From: email@example.com
- Resent-Message-ID: <"DqhiZ3.0.jX.Qrh9s"@electra.znyx.com>
- Resent-Sender: firstname.lastname@example.org
>Does anyone know if it is possible for a PCI chip to be completely
>compatible with both 5V and 3.3V PCI signaling if it has absolutely no
>connection to a 5V source and no external glue logic? If so, is there
>one commercially available?
Yes it is possible, depending on how strict you are about "no connection
to a 5V source."
You must change the overshoot clamping between 3.3V and 5V environments.
The clamping level could change, or it could be simply disabled for the
On a PCI expansion board, the only way to tell which environment you are
in, is by way of the "+Vio" pins on the PCI connector. These pins have
+5V when the bus uses 5V signaling. So technically, if you connect +Vio
directly to the chip, you would have the system's +5V source connected
to it. Or you could use external glue logic to sense +Vio and have it
signal the chip.
Intel's (formerly DEC's) PCI chips with dual-voltage capability, come
close. They use 3.3V power only and require no 5V power. But there is
that connection to +Vio to control the overshoot clamps. On these
chips, it is a low-current input pin, not a power pin.