my customer found some timing mismatch of t16 (CS Falling Edge to isoSPI Data Transmit) and t14 (SCK Non-Latch Edge to isoSPI Data Transmit) from the datasheet of LTC6820
below test results are tested at slave side , the setting of the SPI mode is 1 (POL=0, PHA=1), Slow=0 (SPI CLK = 750kHZ)
Also ,it used Capacitive Isolation Barrier for the ISO SPI but not transformer (the cable length is under 1m)
please see the waveforms below:
1. Purple waveform is the ISO SPI (IP-IM) yellow waveform is the CS
the timing is 8.43us , which fail the t16 spec (<345ns)
2.Purple waveform is the ISO SPI (IP-IM) yellow waveform is the CLK
the timing is 930ns, which Fail the t14 spec (<190ns)
May I have some of your suggestion if the test or the setting was wrong?
May I know if you have any idea on this issue?
I will move your question since this does not belong to Power by Linear. Somebody should respond through Interface and Isolation portfolio.
Got it and thanks for the help
These photos are taken at the slave side, so the timing will be from isoSPI to SPI and those times appear appropriate (close cause & effect). The isoSPI waveforms are showing a slow tail and this may be extending the decode time somewhat. Normally the isoSPI signals should look fairly rectangular, not slow RC as shown. Capacitive coupling is ok but termination should be ~100R differential (we use two ~50R resistors with the center node bypassed to V- with 10nF) and coupling caps of 22nF which makes relatively low-droop pulses as intended.
I will try your suggestion for the signal integrity of the isospi data.
Quoted from your words" timing will be from isoSPI to SPI and those times appear appropriate (close cause & effect" )
so your opinion is that you think the timing (t14, t16) that doensn't match the spec in the datasheet is normal?
or the failure was caused by bad signal integrity of ISOspi?
please correct me if my understanding is wrong.
on the other hand , a 10nF cap is suggested to connect to the middle point of two 50R (termination resistors) and connected the other side of the cap to " V-" , may I know what exactly you meant V-?