Re: 3.3v env

[Ted Firlit <ted.firlit@aeroflex.com>]

yuval wrote:
> 
> Hi Ted!
> Thanks for reply!
> The systems we are working with are SBC'S and PASSIVE backplane.

Is it a CPCI system?  The CPCI systems which I have worked with also have a passive backplane, but the VIO voltage is still determined by the backplane (I believe this is required by the spec-- I can look into it if you like.)

> The main issue was how to develop a backplane that can support a "real 3.3v"
> card 64bit /66mhz
> but still to use an SBC.
> The SBC (by the way not only from our manufacturer) is designed to 5v
> environment. The VIO is connected to VCC.
> It seems also that all SBC'S producers  implement 66mhz NOT BY THE BOOK!
> We arrived to conclusion that to implement a 64bit/66mhz by the book in this
> configuration we will have to add
> a bridge that will convert 5v to 3.3v.

I'm a little confused-- is the bridge on your peripheral card?  Will you still have a 66MHz bus running at 5V VIO coming from the SBC?

For future reference, which SBC vendors have you determined do not offer a 64MHz card compatible with 3.3V VIO?

thanks for the info!

Ted Firlit

> 
> Thanks
> 
> Yuval Ben Mordechai
> Customer service and support manager
> Eim System's and components
> 972 - 55 611246
> ----- Original Message -----
> From: "Ted Firlit" <ted.firlit@aeroflex.com>
> To: <yuval@eimsys.co.il>
> Sent: Wednesday, December 04, 2002 12:04 AM
> Subject: Re: 3.3v env
> 
> > Yuval:
> >
> > I ran into a similar situation with a compactPCI "system slot" SBC.  (I
> believe it was a Motorola board.)  The SBC itself had the VIO pins shorted
> to the 5V Vcc pins.  Therefore, it could only be installed in a backplane
> with the VIO set to 5V. (In CPCI, the backplane defines the VIO voltage for
> the entire system, usually selectable by the user via jumpers to either 5V
> or 3.3V.)
> >
> > This 5V limitation of the SBC surprised me, but was not a major concern
> since our peripheral card is universal, and could operate with either VIO
> voltage defined by the VIO pins.
> >
> > I agree with Andy's last paragraph about your SBC manufacturer possibly
> calling '3.3V signaling' the signal swing voltage, not the VIO voltage.
> >
> > If you need a SBC that will operate in a 3.3V VIO backplane, I think you'd
> better keep looking.
> >
> > Ted
> >
> > "Ingraham, Andrew" wrote:
> > >
> > > Yuval,
> > >
> > > My responses below apply to conventional PCI.
> > >
> > > You wrote:
> > >
> > > > From the SBC manufacturer we received a data sheet indicating that VIO
> must be connected to VCC.(5v)
> > > > We received from them assurance that this SBC will work in 3.3v
> signaling.
> > > > So now we are confused?
> > >
> > > I am confused too.  The VIO voltage always equals the PCI signaling
> > > environment voltage.
> > >
> > > In PCI this voltage is also known as VCC.  That is, VCC = 5V when using
> > > 5V signaling, and VCC = 3.3V when using 3.3V signaling.
> > >
> > > Did the SBC manufacturer say that VIO must be connected to 5V, or that
> > > it must be connected to VCC?
> > >
> > > > 1) What will be the VIO level connected to the expansion pci64 slot ?
> > > > Is it 3.3v from the 3.3v rail?
> > >
> > > 64-bit PCI can use either 5V or 3.3V signaling; thus, VIO might be
> > > either 5V or 3.3V.
> > >
> > > > 2) Doesn't the VIO on the expansion slots must be the same VIO on the
> PICMG?
> > > > 3)If it is so how it will work on 3.3v signaling if the VIO on the
> PICMG is connected to the 5V?
> > > > 4) Does the VIO to the pci slots to the 3.3v card must be separate
> (connected to 3.3v) from the VIO to the PICMG
> > > > connected to VCC(5V)
> > >
> > > I am not sure exactly which PICMG environment you are talking about.
> > > PICMG includes CompactPCI, as well as passive backplanes that use the
> > > conventional PCI form factor.
> > >
> > > However, *my* assumption (unless proven otherwise) would be that VIO
> > > on each card would match the VIO on the backplane.  The two connect
> > > together, and the VIO voltage on the card comes from the VIO voltage
> > > on the backplane.
> > >
> > > But also note that the "5V Signaling Environment" includes signals
> > > that may switch only to 3.3V; that is, signals that do not switch all
> > > the way between 0V and VIO = 5V.  Some people might call these signals
> > > "3.3V signaling" even though they are not PCI's definition of "3.3V
> > > signaling."
> > >
> > > Regards,
> > > Andy