Serial interface and logic analyzer

[john]

Hello,

I have following signals coming out of the CPLD

1. Tag
2. Data
3. SCLK ( 3MHz)
4. DCLK ( 1.5MHz)


Please go to the following link to look at the timing relationship of
the above mentioned signals

http://img174.imageshack.us/my.php?image=timingdiagramsy1.png

I am trying to look at these signals using logic analyzer ( HP
1661ES). The logic analyzer offers three different analyzing modes to
look at the signals

1. timing mode
2. state mode
3. State and timing ( both together, timing analyzer gets trigger by
state analyzer)

I want to trigger the logic analyzer in the right way to see the right
results. Can somebody advice me which mode to choose?

Thanks
John

 

[Joerg]

Hello,

I have following signals coming out of the CPLD

1. Tag
2. Data
3. SCLK ( 3MHz)
4. DCLK ( 1.5MHz)


Please go to the following link to look at the timing relationship of
the above mentioned signals

http://img174.imageshack.us/my.php?image=timingdiagramsy1.png

I am trying to look at these signals using logic analyzer ( HP
1661ES). The logic analyzer offers three different analyzing modes to
look at the signals

1. timing mode
2. state mode
3. State and timing ( both together, timing analyzer gets trigger by
state analyzer)

I want to trigger the logic analyzer in the right way to see the right
results. Can somebody advice me which mode to choose?

If HP uses the same terminology as my Dolch, then state mode. TAG low as
a qualifier plus rising edge of DCLK -> Sample the data line. If you
have to organize the data in a more fancy fashion you'd have to set it
up more nifty so it first looks for a TAG falling edge, then keeps
looking for DCLK rising edges and samples the data, then stops sampling
once TAG rises and writes the next set into another line so it's easier
to read.

One word of caution here. I've had situations with SPI buses where a
device for some reason wanted to read data on the rising edge even
though the datasheet said falling edge or vice versa. One was an ADC. So
I called and after some hmms, ohs and ahems the app engineer said "Well,
then just use the slope that works".