[mythtv-users] IR blaster missing channel changes - Scientific Atlanta 3250

[Jack Perveiler]

On Thu, Mar 18, 2010 at 11:13 PM, John H <arizonamythtv at gmail.com> wrote:

> >Here's a couple of refinements I had to make to get things working
> >perfectly:
> >1) My Home-made blaster 'eyes' were too strong - I had to move them a
> couple
> >of inches AWAY from the receiver to avoid over saturating the STB's IR
> >receiver.
> >2) One of the pre-made 'stick-on' emitters I was using wasn't 'strong'
> >enough - I had to tape a small piece of white paper on the tip of it at a
> 45
> >degree angle to reflect more of the IR into the receiver. (the IR was
> >emitted from the end opposite of the wire, instead of going directly into
> >the receiver)
> >
> >For one of the receivers I had, I also had to increase the "min_repeat" as
> >Jarod mentoned. (Basically that just simulates a human holding down the
> >button slightly longer, thus repeating the signal "x" number of times.)
> >
> >Also, if you've got any fluorescent lights nearby (especially the CFL
> type),
> >they can cause lots of IR interference. Sunlight can too.
> >
> >What is your IR blasting hardware by the way?
>
> Thanks for the help. I'm using the blaster found at the link below.
>
> http://www.irblaster.info/
>
> I thought I was having an over saturating issue but it also happens on
> startup. By startup, if the TV hasn't been on and I start watching
> LiveTV naturally it'll blast the last channel I watched. Even on
> startup sometimes I see the issue so this rules out over saturating
> issue.
>
> I wrote a simple script just to repeat/loop the irsend command and I
> moved the blaster around to find the optimal location. I then stuck it
> on with painters tape and then covered this with electrical tape. I
> figured this would avoid any reflections and interference from
> fluorescent lights.
>
> I'm going to try the min_repeat.
>

I'm using that same blaster and having the same problems you are.  It worked
perfectly under ubuntu 8.04 but started failing every so often when I
upgraded to ubuntu 9.10.  No amount of electrical tape/delay adding/nicing
seemed to help either :)  My guess is that something in kernel-land changed
timing wise.

I ultimately gave in to the lists' conventional wisdom and bought a usb mce
transceiver because everybody says it "just works".  Unfortunately I got a
gen1 device which now can blast anything that isn't 56 kHz so I was out of
luck there with my 56 kHz Scientific Atlanta box.

I read a couple of weeks or so ago on this list that someone else with this
problem switched to using Jarod's git tree and things went back to perfect
again so I figured I'd give that a try.  But his current git tree wouldn't
build against the kernel source that the ubuntu 9.10 packages provide.  I
can probably make it work but my first attempts failed with:

/bin/sh: cannot create /home/me/jarods_lirc/include/asm/asm-offsets.h:
Directory nonexistent
make[1]: *** [/home/me/jarods_lirc/include/asm/asm-offsets.h] Error 2
make: *** [_module_/home/me/jarods_lirc] Error 2

I'm sure it's something stupid on my part, though.

At this point I've come to 2 conclusions:
1) LIRC hates me.  Passionately.
2) Meh, it's almost summer time :)

Here's the thread where Jarod's git tree is reported to work better than
lirc cvs for modern kernels:

http://mythtv.org/pipermail/mythtv-users/2010-February/281282.html

Good luck,

--Jack
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mythtv.org/pipermail/mythtv-users/attachments/20100319/bb5e4514/attachment.htm>