Diff-Mode S-Parameter files exported from HyperLynx not viewable

[EdgarMerger]

S-Parameter-Files converted from standard mode to differential mode using HyperLynx touchstone viewer are not viewable.

Differential Modes S-Parameters from HyperLynx_v27.zip

[EdgarMerger]

Made some adaptations to the non-working file. With that adaptations it worke. ASCII-compare shows the difference.

DP1_TX0_N__made_working_with_SignalIntegrity.zip

[PetePupalaikis]

These s-parameters are in Touchstone 2.0 format, while my tool supports only Touchstone 1.0.

It would not be to hard to make a Touchstone 2.0 parser and handle this.

Can HyperLynx output Touchstone 1.0?

Some advantages to supporting Touchstone 2.0 would be:

1. quicker reading of files due to knowledge of number of frequencies ahead of time.
2. ability to handle different reference impedance on a per-port basis.

I do see mixed-mode keyword, but do not see that in the Teraspeed document. I'd need to look at the standard.

In the past, I looked at the new specification with the intent of supporting it, but the committee who invented that thing made it so complicated with not much advantage and the lack of availability of an open-source golden parser to test against made me shy away from it.

By the way, our commercial instruments do not support this standard either, and I've never heard a complaint.

I'm going to move this to a feature request, although it's method of failing I will treat as a bug initially.

[EdgarMerger]

I don´t see any way to configure HyperLynx with respect tp the Touchstone version. I´ll have to ask Mentor/Siemens about it. But there needs to be a way for it. Since commercial instruments don´t support 2.0, for proper simulation-to-measurent correlation there must be a way to output Touchstone 1.0 from HyperLynx. Will check.

[PetePupalaikis]

Just a slight correction to what you said. "Our (Teledyne LeCroy) instruments do not support it". I'm sure Keysight or Anritsu support it. Since our instruments are mostly a producer of s-parameters and because a Touchstone 2.0 or 2.1 parser must be backwards compatible (able to read Touchstone 1.0), it doesn't cause a problem. The problem is on import, which does change things. I do know, since I wrote the code, that our Virtual Probe oscilloscope component will not accept Touchstone 2.0 files.

[EdgarMerger]

Response from HyperLynx-Support:
Unfortunately when you convert to differential, the default is to save the file as Touchstone 2. But there is an easy workaround to convert the file back to Touchstone 1. You just need to strip out everything in file that has brackets [ ].

[PetePupalaikis]

That will work in this case, for other cases, where the port reference impedances have been changed, for example, that will lead to quite erroneous results.

At some point, I will consider supporting Touchstone 2.0.