Forums

ADSP-21065L SPORT termination problem

Started by Steve Holle July 1, 2004
We have a system with a backplane with five slots for ADSP-21065L
DSP's based custom cards.  Each DSP card has SPORT 0 DR0A and DT0A
tied together. TCLK0, RFS0, and the common DT signals are then each
run through a 33ohm series to the backplane bus.

The problem is that we are having significant noise problems.  One
slot in our back plane seems to have problems on the TCLK line as the
audio crackles until we connect a scope probe to the TCLK line.

We have tried various resistor and capacitor combinations which only
seem to make the problem worse.  We have noted item 21 in the Anomaly
list but are not sure how to proceed to fix the problem.  Has anyone
out there dealt with sport signal termination on the 065L
successfully?  If so, what was your solution, if you wouldn't mind
sharing it.

Thanks for your help.
Steve,

Take a look at this application note:

"Analog Devices Serial Port Development and Troubleshooting Guide"
http://www.analog.com/UploadedFiles/Application_Notes/712188301ee_74.pdf

On our system we were not able to eliminate the problems completely by
proper termination. We had to include high-current drivers for the long
serial lines.

HTH,
-- Georgi



"Steve Holle" <sholle@link-comm.com> wrote in message
news:ba83847d.0407010813.bbee8f6@posting.google.com...
> We have a system with a backplane with five slots for ADSP-21065L > DSP's based custom cards. Each DSP card has SPORT 0 DR0A and DT0A > tied together. TCLK0, RFS0, and the common DT signals are then each > run through a 33ohm series to the backplane bus. > > The problem is that we are having significant noise problems. One > slot in our back plane seems to have problems on the TCLK line as the > audio crackles until we connect a scope probe to the TCLK line. > > We have tried various resistor and capacitor combinations which only > seem to make the problem worse. We have noted item 21 in the Anomaly > list but are not sure how to proceed to fix the problem. Has anyone > out there dealt with sport signal termination on the 065L > successfully? If so, what was your solution, if you wouldn't mind > sharing it. > > Thanks for your help.
Our solution required buffers on the TCLK0 and RFS0 signals from the
card designated to drive those signals.  We use an 74xxx08 because we
could get them very small and squeeze them into the design.  We are
concerned about the DT0A and DR0A signals since they are tied together
for multichannel mode and there is no way to buffer them because they
need to be bidirectional and there is not direction line.

I'm having trouble determining what the load and capacitance limits
are for these signals from the data sheet.  Anyone know where I can
find them?

"Georgi Beloev" <gbH8SPAM@beloev.net> wrote in message news:<HaGdnQSKKYikHHndRVn-vA@megapath.net>...
> Steve, > > Take a look at this application note: > > "Analog Devices Serial Port Development and Troubleshooting Guide" > http://www.analog.com/UploadedFiles/Application_Notes/712188301ee_74.pdf > > On our system we were not able to eliminate the problems completely by > proper termination. We had to include high-current drivers for the long > serial lines. > > HTH, > -- Georgi > > > > "Steve Holle" <sholle@link-comm.com> wrote in message > news:ba83847d.0407010813.bbee8f6@posting.google.com... > > We have a system with a backplane with five slots for ADSP-21065L > > DSP's based custom cards. Each DSP card has SPORT 0 DR0A and DT0A > > tied together. TCLK0, RFS0, and the common DT signals are then each > > run through a 33ohm series to the backplane bus. > > > > The problem is that we are having significant noise problems. One > > slot in our back plane seems to have problems on the TCLK line as the > > audio crackles until we connect a scope probe to the TCLK line. > > > > We have tried various resistor and capacitor combinations which only > > seem to make the problem worse. We have noted item 21 in the Anomaly > > list but are not sure how to proceed to fix the problem. Has anyone > > out there dealt with sport signal termination on the 065L > > successfully? If so, what was your solution, if you wouldn't mind > > sharing it. > > > > Thanks for your help.
"Steve Holle" <sholle@link-comm.com> wrote in message
news:ba83847d.0407020653.616424a8@posting.google.com...
> Our solution required buffers on the TCLK0 and RFS0 signals from the > card designated to drive those signals. We use an 74xxx08 because we > could get them very small and squeeze them into the design. We are > concerned about the DT0A and DR0A signals since they are tied together > for multichannel mode and there is no way to buffer them because they > need to be bidirectional and there is not direction line.
I think TFS is used as "Transmit Data Valid" for external logic in multichannel mode. -- Georgi
> > I'm having trouble determining what the load and capacitance limits > are for these signals from the data sheet. Anyone know where I can > find them? > > "Georgi Beloev" <gbH8SPAM@beloev.net> wrote in message
news:<HaGdnQSKKYikHHndRVn-vA@megapath.net>...
> > Steve, > > > > Take a look at this application note: > > > > "Analog Devices Serial Port Development and Troubleshooting Guide" > > http://www.analog.com/UploadedFiles/Application_Notes/712188301ee_74.pdf > > > > On our system we were not able to eliminate the problems completely by > > proper termination. We had to include high-current drivers for the long > > serial lines. > > > > HTH, > > -- Georgi > > > > > > > > "Steve Holle" <sholle@link-comm.com> wrote in message > > news:ba83847d.0407010813.bbee8f6@posting.google.com... > > > We have a system with a backplane with five slots for ADSP-21065L > > > DSP's based custom cards. Each DSP card has SPORT 0 DR0A and DT0A > > > tied together. TCLK0, RFS0, and the common DT signals are then each > > > run through a 33ohm series to the backplane bus. > > > > > > The problem is that we are having significant noise problems. One > > > slot in our back plane seems to have problems on the TCLK line as the > > > audio crackles until we connect a scope probe to the TCLK line. > > > > > > We have tried various resistor and capacitor combinations which only > > > seem to make the problem worse. We have noted item 21 in the Anomaly > > > list but are not sure how to proceed to fix the problem. Has anyone > > > out there dealt with sport signal termination on the 065L > > > successfully? If so, what was your solution, if you wouldn't mind > > > sharing it. > > > > > > Thanks for your help.