Metastability in 74LVC1G175?

I am not able to find any new data on metastabilty.  I only found this one old appnote.

I found the following graph at:

ftp://ftp.dii.unisi.it/pub/users/vignoli/old/elettronicaii_cl/materiale_didattico/digital/per%20approfondire/metastab.pdf

Which shows the metastability characteristics of LVC.  But my concern is that the single gate parts are so different that this graph does not apply.

It looks like that the rule is the faster and higher drive the better. If that is the case then The little logic should be much better since the will run faster than the standard logic parts.

That is a guideline.  But as you can see from the graph, 74LV is one of the worst for metastability even though one of the fastest families.

Thinking about the cause of metastability, speed would make it less likely to occur but would not necessarily affect the resolution time.  It seems to me that the analog gain of the circuit would have more effect on resolution time.

Also the graph has two lines labeled 74F.  This is either a mistake or indicates that even within a family there can be significant variation. 

It would take a lot of effort to set up a test and get a statistically significant sampling; I was really hoping there would be some data available.

The SN74LV family is low speed low drive with slow edges.  The SN74LVC family is High speed High drive with the faster edges. They are on opposite ends of the spectrum in CMOS  logic.  The LVC is in the same range as F and ABT which are also high speed high drive BiCMOS.

The LV is with HC and LS which are low speed devices.  

I dont know what factor is linked to metastability but it seesm to be either drive or speed.

I hate to prolong this discussion, but perhaps we are using different definitions for speed.  The datasheet only specifies propagation delay, not edge rates, so that is what I am using.

The following data clearly show that speed does not always correlate with good metastability performance and so the doubts which spawned my original question remain.

Comparing typical prop delays at 5V:

SN74LV74 - tpd = 5.6ns (poor metastability)

SN74AC74 - tpd = 6ns (very good metastability)

SN74HC74 - tpd = 20ns (poor metastability - but better than LV)

The problem with the chart is we dont know what Vcc they were tested at. For instance LVC at the time this appnote was written was a 3.3V part so I am sure that is where it was tested. The new single gate parts will go to 5V and are much faster.

 Some of the other parts shown could be tested at either 3.3V or 5V. We have no Idea whic Vcc they used. LV HC and AC could have been tested at either Vcc.

SInce LV is a considered a low voltage part I suspect it was tested at 3.3V.

If you look at 74AHC and 74LV there is a large difference. Thes are basically the same design and sometimes the same die. The AHC is much better and I think it is because it was tested at 5V while LV was tested at 3.3V.

I know I am not much help here but I dont think the chart is very useful either since the propagation delay gets much faster at 5V.

.