[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: JTAG pins



hello,

Just to follow up on the posting I sent a few weeks ago about some
motherboards having non-compliant connections to the TMS signal:

The motherboards we have discovered with non-PCI-compliant 1K pulldowns
on TMS are:

Tyan Thunder LE S2510
Supermicro 370DLR 

>> TYAN:
Here's the response I obtained from our motherboard distributor
regarding Tyan:

>Tyan said that at the present time, all their boards that have the
>ServerWorks chipset (just as a coincidence) have that problem.  You are
>right, it does not have anything to do with the chipset.

>There new board (Tyan S2518) will have that solved and should be available
>within a few weeks.

>> SUPERMICRO:
The only response we have received within the last couple weeks is that
they have indeed received our question and are looking into it.

As far as other vendors, most user's manuals I have looked at do not
describe the connection to TMS, but Intel manuals usually do contain
some info.  Here's what I've found with respect to TMS:

Advanced/EV motherboard: no-connect
AN430TX: +5V
Desktop D815EEA: +5V

(the +5V may be pullups or direct connects-- I haven't checked)

I have heard that an example Intel schematic contains a pull-down, but
the Intel checklist for motherboard design says to pull-up TMS with 5K
(which should take precedence over the schematic-- of course the PCI
spec takes overall precedence.)



regards,

Ted Firlit


Lame Brooks-G14738 wrote:
> 
> I strongly suggest that we try to enforce the compatibility intent of
> specifications by making it widely known (first to the source) when there is
> a real problem due to a manufacturer not adhering to the spec(s).  What is
> the mfg and model of the non-compliant motherboard.  Distributors should
> also be informed.
> 
> I just want the products I buy to work!    -- BrooksL      :)
> 
> > -----Original Message-----
> > From: Olaf Birkeland [mailto:Olaf.Birkeland@fast.no]
> > Sent: Monday, 03 September, 2001 00:03
> > To: pci-sig@znyx.com
> > Cc: ted.firlit@utmc.aeroflex.com
> > Subject: RE: JTAG pins
> >
> >
> > > -----Original Message-----
> > > From: firlit@utmc.aeroflex.com [mailto:firlit@utmc.aeroflex.com]
> > > Sent: Friday, August 31, 2001 20:18
> > > To: pci-sig@znyx.com
> > > Subject: JTAG pins
> > >
> > >
> > > Hello,
> > >
> > > I have just recently come across a fairly new motherboard
> > which has the
> > > PCI connector TMS signal pulled down to ground with a 1K resistance.
> > > This is crippling to my PCI card, since I use the JTAG signals for
> > > production test of the card (and this pulldown puts my PCI interface
> > > device into JTAG mode).  This brings me to 2 questions:
> > >
> > > 1.  Has anyone come across this problem and researched into the
> > > magnitude of it?  Are there few or many motherboards out
> > there which are
> > > not compliant to the rev 2.2 spec, paragraph 4.3.3 --
> > requiring a pullup
> > > on TMS and TDI?
> >
> > I've seen a lot of variation on this, ranging from incorrect
> > pull-up/down to
> > no resistors at all.
> > >
> > > 2.  Are most PCI expansion cards designed without using the
> > JTAG pins of
> > > the PCI connector, hence most cards operate fine in these
> > non-compliant
> > > motherboards?
> >
> > At least to my experience. I've opted for bringing JTAG out
> > to a separate
> > connector in my designs, leaving the PCI JTAG as is (i.e.
> > connecting TDI to
> > TDO). This makes my JTAG chain independent of whatever the
> > motherboard and
> > other expansion cards are doing. The drawback is the need for
> > a separate
> > test connector and plugging into this during production test.
> >
> > Regards,
> > - Olaf
> >