SN74CBT3257C: SPI clock signal cross talk

Hi Henry,

I have a few questions so I fully understand the scope of the issue you are having:

1. are the "spikes" the interference you are referring too (I have reattached the scope shot with the "spikes" circled) ?

2. Is the waveform from the SN74CBT3257C or is it from the TMUX1574? Could you provide scope shots of both waveforms so I can see if there is something that differs between them.

3. What are the signals at SDI and SDO look like? If you can have the scope shot with all the signals labeled going into the device (as if the issue is cross talk other signals would "leak" into the active channel. 

4. What state is "S"  in during your tests? 

5. Could you also provide layout of the switch for this application - as a lot of these issues are layout dependent and it may not be a switch problem.

If you can provide that information it will be very helpful. From the schematic itself, nothing looks wrong - but there could be issue in layout that could be occurring, so providing that would be great. The 3257C is not as good as the TMUX1574 in terms of performance - so this is why I want to see comparisons between both parts as if they are having the same issues - I wouldn't think its the mux alone - but in addition to the other system parameters there are issues being caused. If it is cross talk - other signals on the device would correspond with the "leakage" that is being seen - but I can't really see any signal besides the blue signal which I assume is pin 5 measurement. 

Please if you could answers my questions I will see what I can do to get to the bottom of this problem for you!

Best,

Parker Dodson 

Hi Sir,

1. are the "spikes" the interference you are referring too (I have reattached the scope shot with the "spikes" circled) ? => This is a scope waveform of the pin-5 and these conditions is the "S" is High, the "#OE" is Low and removed "R59" to measure the pin-5 side. They think the waveforms of pin-6 cannot leak into the pin-5.

2. Is the waveform from the SN74CBT3257C or is it from the TMUX1574? Could you provide scope shots of both waveforms so I can see if there is something that differs between them. => They have not save the waveforms of the TMUX1574.

3. What are the signals at SDI and SDO look like? If you can have the scope shot with all the signals labeled going into the device (as if the issue is cross talk other signals would "leak" into the active channel. => They are a common communication of the SPI format.

4. What state is "S" in during your tests? => It's the "High".

5. Could you also provide layout of the switch for this application - as a lot of these issues are layout dependent and it may not be a switch problem. => I'm so sorry, The customer can not provide the layout file due to the confidential reason.

The customer mentioned to use the signal generator be a signals source to test the SN74CBT3257C and TMUX1574, also found the signal leak into the inactive channel. BTW, they also try to test the Diodes/ PI5C3257, have not found the signal leak into the inactive channel.

Hi Henry,

1. Can the scope shot be taken again with all the signals shown - I can only really see one signal and I am not sure which signal is what.

2. Can they retake pictures with the 1574 - as if the scope shots are the same its very unlikely it is the switch and something else is more than likely causing the problem. 

3. I just want to emphasize a clear scope shot with the signal labeled can help me determine if there is actually cross talk- I only really can see one signal; so I am not sure what the problem is.

4. Thank you.

5. That is okay - but I do want to stress that higher frequency leakage between adjacent circuit nodes can be caused by layout - so even if everything else is perfect this can cause issues. If you can't share that is okay - but I would advise the designer to look at layout and if  possible try to simulate the traces to see if there is any interference coming from the layout itself - I do suspect to be a likely culprit of at least part of the interference.

6. Leakage  into inactive channels - yes that is possible as there is capacitive coupling between the I/O of the device - This is a physical reality  of all semiconductor switches - including the comp. I am  a little skeptical that the part isn't showing issues and the 1574 is. 

Please I need clearer scope shots - and I would need the scope shot of the 1574 as well. I don't know where the cross talk is coming from as there are no other signals present on the line that are shown. I want to see clock, data, and CS lines and stacked, not overlapping - for both devices. These devices should not be performing the same way so I am very skeptical that its an actual device issue.

Best,

Parker Dodson