Diseqc switches

[Andy Burns]

My previous satellite setup had a Smart Oxid quad LNB on 28.2E

two of the LNB outputs fed a dual tuner DVB-S2 tuner in PC
other two LNB outputs headed off to down to lounge via Blake starBox

Recently I've added a Smart Oxid single LNB on 19.2E

Initially I connected the new LNB's output direct to one of the tuner
card inputs, I could scan for and receive all the expected channels from
19.2E

So, happy that I'd found the correct satellite, but before final
positioning and tweaking, I inserted

Smart SUS4 2x1 diseqc switch

With the output of the single LNB on port1 of the switch
and the first output of the quad LNB on port2
the output of the switch goes to one of the tuner card inputs

Reconfigured the PC's software (mythtv) so it is aware the switch is
there, and which LNBs connect to which ports.

When I tell it to scan 28.2E, it selects port 2, finds the 28.2E
transponders/channels during the scan, and is able to receive them.

When I tell it to scan 19.2E, it selects port 1, it does find various
19.2E transponders/channels during the scan, but jumbled in with them,
it sees 28.2E transponders/channels too, afterwards it isn't able to
receive the 19.2E channels.

I get the same "jumbled" results when using w_scan and scandvb instead
of mythtv.

I also tried NOT telling mythtv about the switch, so that (AIUI) without
diseqc commands the switch will default to port1 and the PC should
simply see the 19.2E signal, this gives the same jumbled result.

Since today has been wet'n'orrible I've not been out to fiddle with the
dish and disconnect the switch, but does this sound like a faulty switch?

What are the different committed/uncommitted switch types, would getting
that wrong give the results I'm seeing?

[Stephen H]

i have seen this before. its basically poor inter-satellite isolation
within the diseqc switch

change the switch for a good quality one. Avoid technomate TM4-s's

The best I've personally worked with in the 4 port global communications
switch. I do not know if these are still in production.



Regards,

Stephen

[Andy Burns]

Stephen H wrote:

> Andy Burns wrote:
>
>> When I tell it to scan 19.2E, it selects port 1, it does find various
>> 19.2E transponders/channels during the scan, but jumbled in with them,
>> it sees 28.2E transponders/channels too
>

> i have seen this before. its basically poor inter-satellite isolation
> within the diseqc switch

Well, it certainly was a cheapy, but Smart is at least a name I
recognised, claims 25dB inter-LNB isolation.

> change the switch for a good quality one. Avoid technomate TM4-s's
> The best I've personally worked with in the 4 port global communications
> switch. I do not know if these are still in production.

I can see Global Invacomm 2x1, but it only claims same isolation as the
Smart one I have.

[John Legon]

Andy Burns wrote:
> I get the same "jumbled" results when using w_scan and scandvb instead
> of mythtv.
>
> I also tried NOT telling mythtv about the switch, so that (AIUI) without
> diseqc commands the switch will default to port1 and the PC should
> simply see the 19.2E signal, this gives the same jumbled result.
>
> Since today has been wet'n'orrible I've not been out to fiddle with the
> dish and disconnect the switch, but does this sound like a faulty switch?

It certainly sounds to me like a faulty switch.

>
> What are the different committed/uncommitted switch types, would getting
> that wrong give the results I'm seeing?

An uncommitted switch wouldn't respond to the diseqc 1.0 switching codes
that committed switches respond to, but they also wouldn't jumble up the
signals from two LNBs. I've used both 2-way and 4-way Konig switches
without any problems - about £6 through Amazon for a 2-way.

[Stephen H]

I used to have 3 single LNB's via a single Global 4 port switch. I
changed the single LNB's for quad LNBs and added 3 technomate switches
to feed additional receivers.

Any receiver connected to the technomate switches exhibited the problem
you described. The one attached to the global switch was fine. Swapping
the receivers around proved that it was the 3 technomate switches at
fault. Anything attached to the global switch was fine and could switch
satellites no problem.

I ended up going to 3 quattro LNB's and a 17x16 multiswitch. All my
probs went completely. The technomate switches were cheap, I got them
for a fiver each. The global switch cost me 20 quid in the first place.

HTH,

Stephen

[John Legon]

I suspect that many of the different brands come from the same
manufacturer. Price isn't necessarily an indication of quality.

[Andy Burns]

John Legon wrote:

> I suspect that many of the different brands come from the same
> manufacturer.

Several of the ebay/amazon ones do bear a family resemblance, but some
are obviously different.

> Price isn't necessarily an indication of quality.

Bearing in mind your other reply, a while back when I ordered a Konig
satfinder, what arrived looked the same but was badged different anyway,
I didn't really object.

I've just had a look, and the Konig switch is identical to the Smart
switch I have anyway.

I've just had a thought ... I do tend to leave a long tail on the inner
conductor when making up F connectors, wonder if one has poked somewhere
it shouldn't?

[Andy Burns]

Stephen H wrote:

> Avoid technomate TM4-s's

Hmm, searching for those, they do like like yet another rebadge of the
4-port big brother of my 2-port one ...

[John Legon]

Andy Burns wrote:
> John Legon wrote:
>
>> I suspect that many of the different brands come from the same
>> manufacturer.
>
> Several of the ebay/amazon ones do bear a family resemblance, but some
> are obviously different.
>
>> Price isn't necessarily an indication of quality.
>
> Bearing in mind your other reply, a while back when I ordered a Konig
> satfinder, what arrived looked the same but was badged different anyway,
> I didn't really object.

Might be Konig, Labgear, SLX and a few others

>
> I've just had a look, and the Konig switch is identical to the Smart
> switch I have anyway.

No surprises there then! My Konig switches seem solidly made and the
4-way has been in regular use for a couple of years.

>
> I've just had a thought ... I do tend to leave a long tail on the inner
> conductor when making up F connectors, wonder if one has poked somewhere
> it shouldn't?

The inner wire is usually allowed to protrude by about 2-3 mm beyond the
outer rim of the plug.

As a basic test that the switch is switching and not shorted between the
two ports, I would suggest measuring the voltage on the inner conductor.
This should be 13 or 18 volts when the port is selected, and 0.00 when
it's off.

[Andy Burns]

John Legon wrote:

> The inner wire is usually allowed to protrude by about 2-3 mm beyond the
> outer rim of the plug.

I seem to remember I left one of them about 1cm.

> As a basic test that the switch is switching and not shorted between the
> two ports

The odd thing is that port2 appears to "leak" into port1, but not
vice-versa.

> I would suggest measuring the voltage on the inner conductor.
> This should be 13 or 18 volts when the port is selected, and 0.00 when
> it's off.

Yeah, when it's less soggy I'll get on the roof (only the garage
fortunately) and have a look, easiest way to check the switching will be
to re-attach the satfinder which lights up when powered, while SSH'ed
over WiFi to the mythbox from a netbook.

[Michael Chare]

On 18/03/2012 19:29, Andy Burns wrote:

I use three of these for a few years without problems:

http://www.octagoneurope.co.uk/cart/index.php?l=product_detail&p=5

- and I think they cost be more than that too!

Diseqc 1.0 switches (2 and 4 port) use Committed commands.
You might see the ports denoted as A, B, C and D or AA, AB, BA, BB


Diseqc 1.1 switches (up to 8 port) use Uncommitted commands
You might see the ports as 1, 2, 3, ...


It just might be the Mythtv software that is causing your problem.
Is there a Myth TV forum?

Windows Home Premium or DVBviewer will drive a 1.0 switch OK. (IME)

I have had problems with a receiver (and PC card) not supplying enough
power (current) to drive a switch and LNB properly.


--
Michael Chare

[Andy Burns]

Michael Chare wrote:

> I use three of these for a few years without problems:
> http://www.octagoneurope.co.uk/cart/index.php?l=product_detail&p=5

Thanks, in the search for another type, that one was on the list as
unlikely to be a re-branded version of the the one I have.

> Diseqc 1.0 switches (2 and 4 port) use Committed commands.
> You might see the ports denoted as A, B, C and D or AA, AB, BA, BB
>
> Diseqc 1.1 switches (up to 8 port) use Uncommitted commands
> You might see the ports as 1, 2, 3, ...

This is a 2.0 switch.

> It just might be the Mythtv software that is causing your problem.
> Is there a Myth TV forum?

More or less ruled that out but using other Linux commands (w_scan and
dvbscan) as well as mythtv, but it could be a card driver issue, I can
see software issuing the expected diseqc commands

> Windows Home Premium or DVBviewer will drive a 1.0 switch OK. (IME)

Windows would have problems, this is a virtual machine and the BIOS
doesn't support VT-d.

> I have had problems with a receiver (and PC card) not supplying enough
> power (current) to drive a switch and LNB properly.

That's possible, I suppose.

[Glenn Millar]

On 18/03/2012 19:29, Andy Burns wrote:

Had a similar problem at a customers home but with a Monoblock LNB
randomly switching. In the end it turned out that the cable RG6, that
was used for some years, became the problem. Replacing with a clean run
of WF100 cured the problem.

Glenn.

[Andy Burns]

Glenn Millar wrote:

> Had a similar problem at a customers home but with a Monoblock LNB
> randomly switching. In the end it turned out that the cable RG6, that
> was used for some years, became the problem. Replacing with a clean run
> of WF100 cured the problem.

Thanks, this is all fairly new cabling, nicely supported, sealed by SA
tape, under a year old.

The existing four runs were foam cored WF100, the extra run where I
inserted the switch is air spaced 'cos that's all I had handy in black.
I've not checked them particularly closely for stray strands of outer,
but I think I'd have noticed the sparks!

Weather looking promising ...

[John Legon]

Andy Burns wrote:
> John Legon wrote:
>
>> The inner wire is usually allowed to protrude by about 2-3 mm beyond the
>> outer rim of the plug.
>
> I seem to remember I left one of them about 1cm.

Oh dear! I can push a wire into the sockets for only 9mm before meeting
some resistance. The F-plug screws on to a depth of about 5 mm, leaving
just 4 mm of space for the protruding conductor. That leaves 6 mm of
your 1 cm unaccounted for.

So depending on how hard you've been screwing the plugs and the length
and stiffness of the protruding conductor, I think it's possible that
you've shorted something behind the F-socket. Maybe later today I'll
post a photo of the inside of mine which will show what you may have
come up against.

>
>> As a basic test that the switch is switching and not shorted between the
>> two ports
>
> The odd thing is that port2 appears to "leak" into port1, but not
> vice-versa.

Well, I'm not sure what's going on but assuming that the ports are
effectively shorted together (for whatever reason), and that the signal
from 28E on port 2 is somewhat stronger than 19E on port 1 - because
your 19E LNB is offset and has not been optimized - then I think it's
conceivable that the tuner will lock on to 28E even when you're looking
for 19E. Most of the transponders on the two clusters share the same
frequencies and polarities, and AFAIK the capture effect could/should
mean that the stronger signal will win out.

>
>> I would suggest measuring the voltage on the inner conductor.
>> This should be 13 or 18 volts when the port is selected, and 0.00 when
>> it's off.
>
> Yeah, when it's less soggy I'll get on the roof (only the garage
> fortunately) and have a look, easiest way to check the switching will be
> to re-attach the satfinder which lights up when powered, while SSH'ed
> over WiFi to the mythbox from a netbook.

Whatever is easiest. But I'd leave the diseqc switch connections
undisturbed to begin with and put the satfinder at the LNB end.

[John Legon]

Andy Burns wrote:

> Michael Chare wrote:
>
>> Diseqc 1.0 switches (2 and 4 port) use Committed commands.
>> You might see the ports denoted as A, B, C and D or AA, AB, BA, BB
>>
>> Diseqc 1.1 switches (up to 8 port) use Uncommitted commands
>> You might see the ports as 1, 2, 3, ...
>
> This is a 2.0 switch.

That just means it's 2.0 compliant. The switch control codes that it
responds to are diseqc 1.0.

[Andy Burns]

John Legon wrote:

> That just means it's 2.0 compliant.

2.x meaning the PCI card could query status info from the switch, rather
than just send it commands?

I'm still a bit hazy about the uncommitted/committed, is that the same
as diseqc x.0 vs x.1?

> The switch control codes that it responds to are diseqc 1.0.

I see mythtv sending

e0 00 00 to reset the switch
e0 10 38 fX to change to portX

[Andy Burns]

Stephen H wrote:

> The best I've personally worked with in the 4 port global communications
> switch. I do not know if these are still in production.

Do you mean the type shown here?

http://www.cardman.com/diseqc.html

[John Legon]

Andy Burns wrote:
> John Legon wrote:
>
>> That just means it's 2.0 compliant.
>
> 2.x meaning the PCI card could query status info from the switch, rather
> than just send it commands?

Meaning the switch won't throw a wobbly when 2.0 commands are sent down
the line. The switch won't return status info.

>
> I'm still a bit hazy about the uncommitted/committed, is that the same
> as diseqc x.0 vs x.1?

Yes, x.0 allows 4-way switching, x.1 allows 16-way, and cascaded
switching gives 64-way.

>
>> The switch control codes that it responds to are diseqc 1.0.
>
> I see mythtv sending
>
> e0 00 00 to reset the switch
> e0 10 38 fX to change to portX

Those are diseqc 1.0 control codes for committed switches.
The sequence: e0 10 39 ... would apply to uncommitted switches under
diseqc 1.1.

[Andy Burns]

Michael Chare wrote:

> http://www.octagoneurope.co.uk/cart/index.php?l=product_detail&p=5

I went up and prodded things, I don't think my inners were long enough
to have caused a problem, they were 6-7mm but didn't extend beyond the
"turrets" on the switch sockets.

Put the satfinder in-line with LNB1 and it was on/off intermittently,
all over the place, not reacting properly to diseqc commands (tried both
committed and uncommitted commands, powering down between).

Disconnected everything and swapped LNBs, made a difference, but the
best result I could get was that it would usually power LNB1 when it was
supposed to, and remove power when supposed to, but it never powered
LNB2 when was supposed to, occasionally it would take six seconds (or
perhaps automatically repeated diseqc commands) until it reacted.

All seems very unreliable, could still be insufficient power, or linux
driver issues, but I've ordered one of the Octagon switches and
temporarily wired each input of the PCI card straight to each LNB for now.

[John Legon]

Andy Burns wrote:
> I went up and prodded things, I don't think my inners were long enough
> to have caused a problem, they were 6-7mm but didn't extend beyond the
> "turrets" on the switch sockets.

Having now looked inside my Konig 2-way switch, it does seem unlikely
that the inner conductor of an F-plug could penetrate the plastic insert
at the back of the socket. If it did get through, though, then in the
case of the central socket, the wire could come into contact with one of
the RF PIN diode switches, and maybe cause the problem as described:

http://www.john-legon.co.uk/temp/switch.jpg

Anyway, I hope you will post the results obtained with the new switch in
due course.

[Andy Burns]

John Legon wrote:

> Having now looked inside my Konig 2-way switch, it does seem unlikely
> that the inner conductor of an F-plug could penetrate the plastic insert
> at the back of the socket. If it did get through, though, then in the
> case of the central socket, the wire could come into contact with one of
> the RF PIN diode switches, and maybe cause the problem as described:

No, I'm pretty sure that even when screwed tight, the inner wouldn't
intrude too far. My reason for leaving them a bit longer is they serve
as a useful 'guide' to aligning the 'nut'.

> http://www.john-legon.co.uk/temp/switch.jpg

Looks familiar externally, was it sealed with white silicone and an
aluminium plate?

> Anyway, I hope you will post the results obtained with the new switch in
> due course.

I will ... with the current bypassed wiring I can actually connect the
switch indoors, which is a bit more conducive to fiddling to find what
it might actually be doing that sitting on a roof ...

[John Legon]

Andy Burns wrote:
>
>> http://www.john-legon.co.uk/temp/switch.jpg
>
> Looks familiar externally, was it sealed with white silicone and an
> aluminium plate?

The backplate is soft aluminium but there was no sign of any sealant.

>
>> Anyway, I hope you will post the results obtained with the new switch in
>> due course.
>
> I will ... with the current bypassed wiring I can actually connect the
> switch indoors, which is a bit more conducive to fiddling to find what
> it might actually be doing that sitting on a roof ...
>

Incidentally, in a previous post you say:
~~~~~

I see mythtv sending

e0 00 00 to reset the switch
e0 10 38 fX to change to portX

~~~~~
Are you saying that mythtv resets the switch each time it changes the
port? I'm not sure it should be doing that, or if it is, it should be
waiting 150 ms before sending the next command in order to comply with
the spec. It's just possible that there's a timing problem that is
causing the switch to behave erratically. Personally, I would want to
test it on an ordinary satellite box before throwing it away.

[Andy Burns]

John Legon wrote:

>>> http://www.john-legon.co.uk/temp/switch.jpg

> there was no sign of any sealant.

Looks like mine's the deluxe waterproof version then :-P

> Incidentally, in a previous post you say:
> ~~~~~
> I see mythtv sending
> e0 00 00 to reset the switch
> e0 10 38 fX to change to portX
> ~~~~~
> Are you saying that mythtv resets the switch each time it changes the
> port?

No, the reset only happens once when it opens the tuner, then I see it
sending the port change commands as and when it needs to switch
satellite, just like it changes tone/voltage

> I would want to test it on an ordinary satellite box before throwing
> it away.

Unfortunately the only other satellite box I have is a sky one, which
obviously declines to talk diseqc.

I won't throw it away, though it only cost £3.89 inc P&P

[Pete Shew]

On 19/03/2012 07:30, Andy Burns wrote:
> Michael Chare wrote:
>
>> I use three of these for a few years without problems:
>> http://www.octagoneurope.co.uk/cart/index.php?l=product_detail&p=5
>
> Thanks, in the search for another type, that one was on the list as
> unlikely to be a re-branded version of the the one I have.
>
>> Diseqc 1.0 switches (2 and 4 port) use Committed commands.
>> You might see the ports denoted as A, B, C and D or AA, AB, BA, BB
>>
>> Diseqc 1.1 switches (up to 8 port) use Uncommitted commands
>> You might see the ports as 1, 2, 3, ...
>
> This is a 2.0 switch.
>
>> It just might be the Mythtv software that is causing your problem.
>> Is there a Myth TV forum?
>
> More or less ruled that out but using other Linux commands (w_scan and
> dvbscan) as well as mythtv, but it could be a card driver issue, I can
> see software issuing the expected diseqc commands
>

I am using MythTV and recently added a 19.2 LNB to my Zone2 Sky dish.
After deleting and reconfiguring the DVB-S2 card to know about the
diseqc switch (cheapie off eBay) I scanned both 28.2 and 19.2 without
problem. I set the 28.2 to get EPG from the radio times xmltv feed and
for the 19.2 channels I wanted, to get the EPG off air. MythTV worked
without a glitch.

>> Windows Home Premium or DVBviewer will drive a 1.0 switch OK. (IME)
>
> Windows would have problems, this is a virtual machine and the BIOS
> doesn't support VT-d.
>
>> I have had problems with a receiver (and PC card) not supplying enough
>> power (current) to drive a switch and LNB properly.
>
> That's possible, I suppose

I am using a Hauppauge DVB-S2 card with a diseqc and Sky type LNBs with
no problem.

Pete

[Andy Burns]

Pete Shew wrote:

> I am using MythTV and recently added a 19.2 LNB to my Zone2 Sky dish.

Seems popular at the moment, anyone would think people were avoiding
throwing cash at Murdoch ...

> After deleting and reconfiguring the DVB-S2 card to know about the
> diseqc switch (cheapie off eBay) I scanned both 28.2 and 19.2 without
> problem. I set the 28.2 to get EPG from the radio times xmltv feed and
> for the 19.2 channels I wanted, to get the EPG off air. MythTV worked
> without a glitch.

Thanks, any reason for not using EIT for the EPG (other than 14 days
rather than 8 days)?

> I am using a Hauppauge DVB-S2 card with a diseqc and Sky type LNBs with
> no problem.

Mine's a TeVii dual DVB-S2 card, been flawless in other regards so far.

[Pete Shew]

On 19/03/2012 20:26, Andy Burns wrote:
> Pete Shew wrote:
>
>> I am using MythTV and recently added a 19.2 LNB to my Zone2 Sky dish.
>
> Seems popular at the moment, anyone would think people were avoiding
> throwing cash at Murdoch ...
>
>> After deleting and reconfiguring the DVB-S2 card to know about the
>> diseqc switch (cheapie off eBay) I scanned both 28.2 and 19.2 without
>> problem. I set the 28.2 to get EPG from the radio times xmltv feed and
>> for the 19.2 channels I wanted, to get the EPG off air. MythTV worked
>> without a glitch.
>
> Thanks, any reason for not using EIT for the EPG (other than 14 days
> rather than 8 days)?

14 days was the main reason. The other is that I have a DVB-T card and
after scanning I fix up the channel numbers to match and to make sense
to me. I use two variants on xmltv configuration tailored for FreeSat
and Freeview as the 'sources' which means I get the same EPG data for
both. That way I don't usually know which card is actually receiving the
channel as long as it isn't HD.

[Andy Burns]

Pete Shew wrote:

> On 19/03/2012 20:26, Andy Burns wrote:
>
>> any reason for not using EIT for the EPG (other than 14 days
>> rather than 8 days)?
>
> 14 days was the main reason. The other is that I have a DVB-T card and
> after scanning I fix up the channel numbers to match and to make sense
> to me.

With just 28.2E and DVB-T channels, I didn't get (too many) clashes, but
I notice 29.2E channels intrude into the 3-digit range and mingle with
the DVB-T channels

Are there various 'standard' channel rename/renumber scripts out there?

[Andy Burns]

Andy Burns wrote:

> With just 28.2E and DVB-T channels, I didn't get (too many) clashes, but
> I notice 29.2E channels intrude into the 3-digit range and mingle with
> the DVB-T channels

that should say 19.2E obviously.

[Pete Shew]

I based mine on an SQL script by Stephen Robertson originally posted on
the mythtv users mailing list. This sanitises the different channels by
making the name and callsign identical for Freeview and FreeSat channels
that are the same, including setting your own choice of regional TV as
the 'real' one for FreeSat. It starts by setting all channel numbers
+100000 and hiding them and then bringing back the ones that are wanted.
e.g.

UPDATE channel SET name='BBC One', callsign='BBC One' WHERE name='BBC 1
Yrks&Lin';
UPDATE channel SET name='BBC One', callsign='BBC One' WHERE name='BBC ONE';
UPDATE channel SET channum=101, visible=1,
xmltvid="yorkslincs.bbc1.bbc.co.uk" WHERE name='BBC One';

I set BBC1 HD to Channel 1 and my regional SD version to 101, etc. The
result is that MythTV sees the Freeview and FreeSat channels as identical.

I also used the script to use the range 60 to 70 for the 19.2E channels
I wanted, however I printed a list of all the channels, including the
high numbered hidden ones and any of these can be selected to watch
although you can't record without EPG data. It impresses visitors when
you select their own regional news at 18:30 :-)

I can send you a copy if you can't find it.

Pete

[John Legon]

Andy Burns wrote:
> Put the satfinder in-line with LNB1 and it was on/off intermittently,
> all over the place, not reacting properly to diseqc commands (tried both
> committed and uncommitted commands, powering down between).
>
> Disconnected everything and swapped LNBs, made a difference, but the
> best result I could get was that it would usually power LNB1 when it was
> supposed to, and remove power when supposed to, but it never powered
> LNB2 when was supposed to, occasionally it would take six seconds (or
> perhaps automatically repeated diseqc commands) until it reacted.

One more question: Have you tried connecting up the switch with the
LNBs and associated cables disconnected, and putting just the satfinder
on the LNB1 and LNB2 inputs alternately? If the power LED on the
satfinder lights up appropriately when you select 19E or 28E then that
might rule out the switch as the cause of the problem.

[Andy Burns]

John Legon wrote:

> One more question: Have you tried connecting up the switch with the LNBs
> and associated cables disconnected, and putting just the satfinder on
> the LNB1 and LNB2 inputs alternately? If the power LED on the satfinder
> lights up appropriately when you select 19E or 28E then that might rule
> out the switch as the cause of the problem.

Not tried that, just need to make up an extra cable to use the switch
indoors ...

[Stephen H]

Yes indeed.

In case its out of production, I have used diseqc switches from the
following manufacturers without any problems so far:

Triax
EMp Centauri

There are some other makes worth looking into but I have no direct hands
on experience of them, they are Spaun and Chess.

All of the four above also make mutiswitches as whereas Technomate do
not. So As long as you buy a diseqc switch from a supplier that also
does Multiswitches should hopefully know how to design them properly.

Stephen.

[John Legon]

Andy Burns wrote:
> John Legon wrote:

>> The switch control codes that it responds to are diseqc 1.0.
>
> I see mythtv sending
>
> e0 00 00 to reset the switch
> e0 10 38 fX to change to portX

I've just been running my little diseqc analyser program (what I wrote
in assembler) to look at the output from one of my satellite boxes, to
check on the codes being sent to the diseqc switch. It's been a while
since I last did this and I forgot that the format is not as simple as
the above would suggest.

If mythtv is indeed sending e0 10 38 fX to change to port X then this
won't work. I've now found the definition tucked away in annexe B.2 of
the update for the implementation of bus diseqc from Eutelsat.com, in
binary:

1110-0000 Framing byte
0001-0000 Address byte - all LNBs and switches
0011-1000 Command byte - write to port group 0 (committed switches)
1111-xxxx Data byte - select port and also LNB polarity and band

The format of the data byte is

1111-xxxx
^^ select port A,B,C,D using 00, 01, 10 or 11
^ select polarity 0 = vertical, 1 = horizontal
^ select band 0 = low, 1 = high.

It's strange that LNB modes are mixed in with the port selection in this
way, but as shown by my analyser tests this actually happens.

[Andy Burns]

John Legon wrote:

> If mythtv is indeed sending e0 10 38 fX to change to port X then this
> won't work.

[snip details]

The postie tried to deliver the new switch on Wednesday, but I was out,
and I'm never near the sorting office when it's open, so had to collect
it this morning.

Results are better, but not much, seems the switch never moves from
port1, but at least I'm not seeing port2 signals "intruding" now.

I suspect by PCIe card isn't sending diseqc properly, I see quite a few
complaints about the TeVii S480 card and diseqc

I do think mayth/w_scan are aware of the bitwise nature of the 4th byte,
depending on transponder and whether I treat the switch as committed or
uncommitted I see it sending e.g.



# w_scan -f s -a 3 -s S19E2 -vvvv -C UTF-8 -D 1u

w_scan version 20120112 (compiled for DVB API 5.3)
using settings for 19.2 east Astra 1F/1G/1H/1KR/1L
scan type SATELLITE, channellist 6
output format vdr-1.6
output charset 'UTF-8'
-_-_-_-_ Getting frontend capabilities-_-_-_-_
Using DVB API 5.5
frontend 'Montage Technology DS3000/TS2020' supports
INVERSION_AUTO
DVB-S
DVB-S2
FREQ (0.95GHz ... 2.15GHz)
SRATE (1.000MBd ... 45.000MBd)
using LNB "UNIVERSAL"
using DiSEqC uncommitted switch 1
-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_
diseqc_send_msg:429: DiSEqC: e0 10 39 f1 00 00
DiSEqC: switch pos 0, 18V, loband (index 1)
diseqc_send_msg:429: DiSEqC: e0 10 38 f2 00 00

diseqc_send_msg:429: DiSEqC: e0 10 39 f1 00 00
DiSEqC: switch pos 0, 13V, loband (index 0)
diseqc_send_msg:429: DiSEqC: e0 10 38 f0 00 00

diseqc_send_msg:429: DiSEqC: e0 10 39 f1 00 00
DiSEqC: switch pos 0, 18V, loband (index 1)
diseqc_send_msg:429: DiSEqC: e0 10 38 f2 00 00

diseqc_send_msg:429: DiSEqC: e0 10 39 f1 00 00
DiSEqC: switch pos 0, 18V, loband (index 1)
diseqc_send_msg:429: DiSEqC: e0 10 38 f2 00 00
signal ok:
S f = 10744 kHz H SR = 22000 5/6 0,35 QPSK



# w_scan -f s -a 3 -s S19E2 -vvvv -C UTF-8 -D 1c

w_scan version 20120112 (compiled for DVB API 5.3)
using settings for 19.2 east Astra 1F/1G/1H/1KR/1L
scan type SATELLITE, channellist 6
output format vdr-1.6
output charset 'UTF-8'
-_-_-_-_ Getting frontend capabilities-_-_-_-_
Using DVB API 5.5
frontend 'Montage Technology DS3000/TS2020' supports
INVERSION_AUTO
DVB-S
DVB-S2
FREQ (0.95GHz ... 2.15GHz)
SRATE (1.000MBd ... 45.000MBd)
using LNB "UNIVERSAL"
using DiSEqC committed switch 1
-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_
diseqc_send_msg:429: DiSEqC: e0 10 39 f0 00 00
DiSEqC: switch pos 1, 18V, loband (index 5)
diseqc_send_msg:429: DiSEqC: e0 10 38 f6 00 00

diseqc_send_msg:429: DiSEqC: e0 10 39 f0 00 00
DiSEqC: switch pos 1, 13V, loband (index 4)
diseqc_send_msg:429: DiSEqC: e0 10 38 f4 00 00

diseqc_send_msg:429: DiSEqC: e0 10 39 f0 00 00
DiSEqC: switch pos 1, 18V, loband (index 5)
diseqc_send_msg:429: DiSEqC: e0 10 38 f6 00 00

diseqc_send_msg:429: DiSEqC: e0 10 39 f0 00 00
DiSEqC: switch pos 1, 18V, loband (index 5)
diseqc_send_msg:429: DiSEqC: e0 10 38 f6 00 00
signal ok:
S f = 10744 kHz H SR = 22000 5/6 0,35 QPSK



Off to update to latest kernel ...

[John Legon]

Andy Burns wrote:
> John Legon wrote:
>
>> If mythtv is indeed sending e0 10 38 fX to change to port X then this
>> won't work.
>
> [snip details]
>
> The postie tried to deliver the new switch on Wednesday, but I was out,
> and I'm never near the sorting office when it's open, so had to collect
> it this morning.
>
> Results are better, but not much, seems the switch never moves from
> port1, but at least I'm not seeing port2 signals "intruding" now.

Although the isolation between ports on my Konig switches certainly
isn't 100%, it's impossible to get a result through a port that isn't
selected. The receiver indicates no signal.

>
> I suspect by PCIe card isn't sending diseqc properly, I see quite a few
> complaints about the TeVii S480 card and diseqc

More likely than the switch being faulty, I would think.

>
> I do think mayth/w_scan are aware of the bitwise nature of the 4th byte,
> depending on transponder and whether I treat the switch as committed or
> uncommitted I see it sending e.g.

The bits for the committed switch data byte appear to be correctly set.
Not sure about the data for uncommitted switches (codes e0 10 39 fx)
or why these are being sent.

[...]

> using DiSEqC uncommitted switch 1
> -_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_
> diseqc_send_msg:429: DiSEqC: e0 10 39 f1 00 00
> DiSEqC: switch pos 0, 18V, loband (index 1)
> diseqc_send_msg:429: DiSEqC: e0 10 38 f2 00 00

[...]

[Andy Burns]

John Legon wrote:

> The bits for the committed switch data byte appear to be correctly set.

I take it that a "normal cheapo" 2 way diseqc switch should be assumed
to be committed (in the absence of any documentation to the contrary)?

> Not sure about the data for uncommitted switches (codes e0 10 39 fx) or
> why these are being sent.

with w_scan the "-D xy" option chooses diseqc port 'x' (0 based) and
mode 'y' (c or u) I just wanted to make sure it made a difference to the
commands being sent.

[John Legon]

Andy Burns wrote:
> John Legon wrote:
>
>> The bits for the committed switch data byte appear to be correctly set.
>
> I take it that a "normal cheapo" 2 way diseqc switch should be assumed
> to be committed (in the absence of any documentation to the contrary)?

Yes, the common 2-way and 4-way switches are "committed". Uncommitted
switches are generally 8-way and much more expensive.