[sdiy] Matched transistors - critical parameters?

[Mattias Rickardsson]

Speaking of parameter matching:

On 7 February 2012 23:35, Matthew Smith <matt at smiffytech.com> wrote:
> The DMMT5551 (and DMMT5401 for PNP) from Diodes Inc., are "intrinsically
> matched," which is explained as being built on adjacent dies from a single
> wafer.

Is "intrinsically matched" mostly a buzzword from the sales department
:-) or do the transistor parameters really end up so similar that it
can be considered as transistor matching? If so, what parameters are
matched by design? hFE? UBE? Something else?

(Diodes Inc. writes in the datasheet:
"Key Features
1. Intrinsically matched PNP – DMMT3906W or NPN – DMMT3904W die
selected from adjacent locations on the wafer to ensure the closest
match possible.
2. Matched tolerance for hFE (DC current gain), Vbe(SAT), and
Vce(SAT), guaranteed within 2% maximum, and less than 1% typical. 1%
matching available upon request")

On 8 February 2012 06:37, Brock Russell <brockr0 at shaw.ca> wrote:
> I'm using some unmatched high gain dual transistors from On Semi,
> PNP BC857CDW1T1G and NPN BC847CDW1T1G. These are smaller parts,
> SOT-363. They are cheap, seem very well matched and fine for filters.

Nothing seems to be mentioned about matching or selection in these
datasheets, what kind of matching could be expected here? Silicon
fabrication gurus out there, welcome to give some selected wisdoms.
:-)

/mr