Re: JTAG in expansion slots

[David Cary <d.cary@ieee.org>]

Dear designers,

I've seen several embedded CPUs that the manufacturer recommends pull-ups
on all
the JTAG inputs.

ARM Application Note 72
http://www.arm.com/Documentation/AppNotes/

Is this an error in the ARM documentation ? Does the JTAG spec really require
TCK to be pulled down ?

>From: "Olaf Birkeland" <Olaf.Birkeland@fast.no>
>To: <pci-sig@znyx.com>
>Subject: JTAG in expansion slots
>Date: Fri, 31 Mar 2000 08:59:04 +0200
...
>I'm developing a PCI card where we will use JTAG for production tests. I'm
>considering whether to have the expansion card JTAG chain as a separate
>chain, or connecting it to the JTAG port in the PCI connector. I would
>appreciate some thoughts on these two:
>
>1) If the PCI JTAG port is connected, I'm depending on the motherboard to
>drive these signals properly (pullup on TDI/TMS, pulldown on TCK/TRST#) I've
>so far seen two motherboards that doesn't do this (either pullups on all, or
>floating!!) I can't find any requirements for motherboard implementation in
>PCI 2.2 that requires pullup/-down. There is also no checkpoint for this in
>the compliance checklist. The Intel PC chipset reference designs have the
>pullup/-down in place though......
...
>Regards,
...
>Olaf René Birkeland, M.Sc.    Mail:  Olaf.Birkeland@fast.no
>Senior R&D Manager            Web:   http://www.fast.no/
...

>From: "Ingraham, Andrew" <Andrew.Ingraham@COMPAQ.com>
>To: "'Olaf Birkeland'" <Olaf.Birkeland@fast.no>
>Subject: RE: JTAG in expansion slots
>Date: Fri, 31 Mar 2000 18:21:20 -0500
...
>See the section on Pull-ups (4.3.3) in the PCI 2.2 spec.  TDI and TMS are
>supposed to be bussed and pulled up, TCK and TRST# should be bussed and
>pulled down.
...
>Regards,
>Andy



--
David Cary "mailto:d.cary@ieee.org" ><> <*> O-
  http://www.rdrop.com/~cary/
"icbmto:N36 09.055' W95 58.730'" (was: "icbmto:N36 08.830' W97 03.443'")