[mythtv-users] DCT2524 cable box suddenly stopped responding to RS-232

[Marc Randolph]

On Fri, Oct 10, 2008 at 1:47 AM, Marc Randolph <mrand at pobox.com> wrote:
> Howdy all,
>
> My MythTV setup has been humming along pretty well for some time now,
> effortlessly controlling the DCT2524 cable box via RS-232 when all of
> a sudden, sometime within the past week, recordings started to fail.
> LiveTV doesn't work either.  Looking in the backend log, I find:
>
>  Communication failed after 5 tries
> [...]
> After this, it suddenly dawns on me that maybe Time Warner has
> remotely disabled the RS-232 port on the cable boxes here in Dallas.
> I'd call them and ask, but I'm sure I'd get the same response as last
> time when I called just asking if they could turn it on: Someone that
> sounded very helpful and knowledgable had never heard of the RS-232
> port, much less it being disabled and someone having it re-enabled.
> Turns out I didn't need to call - it was already enabled.  So...
>
> 1. Anyone else in the Dallas area controlling a Time Warner cable box via RS-232?  If so, is it working or are you having trouble?

Turns out I was right... Time Warner had disabled the serial port.
Found confirmation on this on the TiVo forum - sometimes it's easy to
forget there are other people (and a considerably larger group, at
that) battling the same types of STB issues that we are.  Not only
that, but they discovered some days later that you could dial the
cable company and go through an automated menu system saying that you
are having trouble and can elect to have the system send a reset
signal down to the cable box.  After doing that, the serial port woke
back up!  Don't know if the cable company made another change or what
- but my serial works again.

Of course, by the time I'd read about resetting the cable box via the
phone, I'd hooked up the IR blaster and got it working acceptably - or
at least, I thought.  So I left it that way for the time being...

A couple of days after hooking up the IR blaster, I did some very
minor OS upgrades and then discovered that LiveTV wouldn't work any
more.  Thinking that it must be the upgrades, I reverted them but
still could not start LiveTV.  The channel changer would fire and
change the STB, but mythfrontend would abort out with the following
cryptic messages:

mythbackend.log:
2008-10-22 22:05:59.749 TVRec(1): Changing from None to WatchingLiveTV
2008-10-22 22:05:59.752 TVRec(1): HW Tuner: 1->1
2008-10-22 22:06:00.854 ret_pid(0) child(21866) status(0x0)
2008-10-22 22:06:01.859 ret_pid(0) child(21866) status(0x0)
2008-10-22 22:06:02.863 ret_pid(0) child(21866) status(0x0)
2008-10-22 22:06:03.867 ret_pid(0) child(21866) status(0x0)
2008-10-22 22:06:04.871 ret_pid(0) child(21866) status(0x0)
2008-10-22 22:06:05.875 ret_pid(21866) child(21866) status(0x0)
2008-10-22 22:06:05.913 External Tuning program exited with no error
2008-10-22 22:06:07.670
Not ivtv driver??
2008-10-22 22:06:07.672 AutoExpire: CalcParams(): Max required Free
Space: 17.0 GB w/freq: 15 min
2008-10-22 22:06:07.742 TVRec(1): Changing from WatchingLiveTV to None
2008-10-22 22:06:08.462 Finished recording Cowboys "Baxter Blind": channel 2186
2008-10-22 22:06:08.466 MythSocket(836e578:-1): writeStringList:
Error, socket went unconnected.

and mythfrontend.log:
2008-10-22 22:05:59.683 TV: Attempting to change from None to WatchingLiveTV
2008-10-22 22:05:59.683 Using protocol version 40
2008-10-22 22:06:06.728 MythSocket(b3196260:27): readStringList:
Error, timeout (quick).
2008-10-22 22:06:06.728 RemoteEncoder::SendReceiveStringList(): No response.
2008-10-22 22:06:06.731 GetEntryAt(-1) failed.
2008-10-22 22:06:06.732 EntryToProgram(0 at Wed Dec 31 18:00:00 1969)
failed to get pginfo
2008-10-22 22:06:06.732 TV Error: LiveTV not successfully started
2008-10-22 22:06:06.732 TV Error: LiveTV not successfully started
2008-10-22 22:06:06.797 TV: Deleting TV Chain in destructor
2008-10-22 22:06:06.800 DPMS Deactivated
2008-10-22 22:06:06.800 DPMS Reactivated.

Turns out that these cryptic failures are due to the IR blaster
channel changing script taking too long.  See the repeated lines
containing "child(21866)"... I think that is mythbackend indicating
that it took until 22:06:05:913 for the channel to change successfully
- but that must have been too long for mythfrontend, and it recorded
that it had timed out less than a second later.  Obviously mythbackend
had no trouble with this long change time though, because it recorded
all my shows as requested.

Making the digit delays shorter in the channel change script fixed the
problem.  And then I switched back to serial.

   Marc