Forums

Reason for Spikes

Started by mannsurya April 17, 2007
Hi guys,

I am looking for a possible reason for a very interesting spike issue.

The setup is as below,

Box A � 16 Channel D/A converter.
Box B � SHARC 21368 DSP platform.

These two boxes are connected in I2S mode using the SHARC�s SPORT. The Box
A is the master providing WCLK and BCLK for the DSP. 8 I2S links are there
from the DSP to the A/D converter. These two units are powered separately
and their grounds are tied together. Each signal running between the boxes
are provided with an adjacent ground line as figure

http://www.flickr.com/photos/85057938@N00/463045733/.

On viewing the signals using the oscilloscope the following were the
observations.

1.	With ground reference from Box A (A/D converter), spikes are observed
on the BOX B (DSP Platform) ground. The grounds connections between the
Boxes are provided as in the above figure. The spikes are observed at the
switching points of the Data in the I2S link. The amplitude of spikes
increased as the number of active data lines increased.

2.	While monitoring the BCLK using the Ground on the BOX B, spikes were
observed. On falling edge of data, the spike on the BCLK undershoot and on
rising edge the spike overshoot. The data and the clock lines are spaced by
one inch. The behavior of the signals are as in the figure below.


http://www.flickr.com/photos/85057938@N00/463045729/

thx in advance
regards
Mann





_____________________________________
Do you know a company who employs DSP engineers?  
Is it already listed at http://dsprelated.com/employers.php ?
mannsurya wrote:
> Hi guys, > > I am looking for a possible reason for a very interesting spike issue. > > The setup is as below, > > Box A � 16 Channel D/A converter. > Box B � SHARC 21368 DSP platform. > > These two boxes are connected in I2S mode using the SHARC�s SPORT. The Box > A is the master providing WCLK and BCLK for the DSP. 8 I2S links are there > from the DSP to the A/D converter. These two units are powered separately > and their grounds are tied together. Each signal running between the boxes > are provided with an adjacent ground line as figure > > http://www.flickr.com/photos/85057938@N00/463045733/. > > On viewing the signals using the oscilloscope the following were the > observations. > > 1. With ground reference from Box A (A/D converter), spikes are observed > on the BOX B (DSP Platform) ground. The grounds connections between the > Boxes are provided as in the above figure. The spikes are observed at the > switching points of the Data in the I2S link. The amplitude of spikes > increased as the number of active data lines increased. > > 2. While monitoring the BCLK using the Ground on the BOX B, spikes were > observed. On falling edge of data, the spike on the BCLK undershoot and on > rising edge the spike overshoot. The data and the clock lines are spaced by > one inch. The behavior of the signals are as in the figure below. > > > http://www.flickr.com/photos/85057938@N00/463045729/
Are you sure the spikes are real? They could be induced artifacts. Look at a ground point near the spiky signal pin. It it spikes too, it's probably the same artifact. Jerry -- Engineering is the art of making what you want from things you can get. ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯