[hpsdr] Potential Metis Bug and Protocol Ambiguity

[George Byrkit]

Dear Jeremy,
I don't know about the apparent Metis bug with the endpoint 6 packets continuing to
arrive.  What version of Metis firmware would that be?

Re the sequence number on EP4 and EP6:
KISS Konsole certainly keeps separate sequence numbers on EP4 (full/wide spectrum) and EP6
(normal data).  Without separate sequence numbers, it would be hard to implement 'missing
frame' checking (on skipped sequence numbers), or know which sequence numbers to bundle
together on the '16k wideband samples' feature.

My attempt to check out your observations failed on my Metis setup, as that doesn't seem
to be working after Dayton!  I'll have to check that out a bit more...

However, Hermes works, and even with the wide spectrum 'de-selected', EP4 (wide/full
spectrum) data keeps arriving.

It seems that in the Metis document (using 'Metis - how it works, rev 1.26), it is
reasonably clear that the sequence numbers are per-endpoint.  It says that the sequence
number is 4 bytes, and is a '32 bit unsigned integer starting at zero and incremented each
frame and independently for each end point'.

KISS only ever sends '3' in the start command, so that it doesn't need to stop, then
restart, metis (or hermes) if you select or deselect Wideband.  If I modify KISS to send
'1' in the start command, only EP6 data is returned.

Given the Metis unit problem, I'm sorry that I cannot test that for you and confirm your
observations on Metis.

Good luck at SeaTac!

73,
George K9TRV


 1369949749.0