[linux-dvb] dvbtune-0.5 inverts diseqc pol and band bits
zoilo at xs4all.nl
Mon Apr 30 18:26:49 CEST 2007
dvbtune-0.5 source apparently contains a bug where it inverts both the
polarization and high/low band bits in the diseqc command ... as I
cannot find any info on this anywhere, I am posting my results here.
I discovered this while using a KNC1 DVB-S card to compare diseqc
handling with VP1034 (because diseqc is not (yet) working on my VP1034).
I am using a Spaun 17089-NF 16-input switch (4 sats) + 8 outputs, which
supports up to Diseqc 2.0.
First I got confused, because the Mantis driver simultaneously uses the
old 18/13V+22kHz method to control polarization and band, as well as the
Diseqc method. So for a while I believed that Diseqc was (partially)
working, until I found out that in fact 18/13V+22kHz seems to be
responsible for the (partial) job!
As a result of no Diseqc-functionality on VP1034, it seems that with
VP1034 on my switch, sat-selection is always fixed to sat=1 (makes
sense, = the first 4 inputs, which happens to be Hotbird 13.0E in my
setup), however input selection from the 4 feeds V/H and H/L works fine
and correctly, although based on 18/13V+22kHz.
When I compared this with my KNC-1 DVB-S card (to find out why
Sat-selection is not working with VP1034), I found that with KNC1,
Sat-selection does switch to Sat=2 (so it must be using diseqc), but
polarization and band-selection were inverted ... So when I issued the
"dvbtune -D 2 -f 10832000 -p H -s 22000"
I found that it was actually selecting sat=2 pol=V band=H ("e0 10 38
f5"= switch input 7), instead of sat=2 pol=H band=L (= switch input 6 /
"e0 10 38 f6").
The Diseqc-specs at
(table on page 13) confirm that dvbtune is in fact sending the wrong
With a patched dvbtune (pol-bit and band-bit both inverted) it works OK,
and is also according to the Eutelsat specs.
Hope someone benefits from this.
More information about the linux-dvb