Saturday, July 30, 2011


The debt-limit "discussion" in Washington makes me sad.

This non-stop posturing for political gain on both sides (but particularly the Republicans) is so lame ... even Kindergarteners have a better sense for when to stop being silly.

Thursday, July 28, 2011


I couldn't resist. ... Talheim now has all semaphores selected, though not yet functional. Tonight I installed the two on the north side of the station.

Talheim north end
The twin-coil machines are installed into the roadbed to hide them a bit from view. Where space is tight between tracks, I dug the hole even a little bit under the track. I will fit a piece of styrene over the opening to protect the machine from landscaping, dust and dirt. Yet I'll try to keep the cover removable for maintenance.

Twin-coil detail
Exit signals at the north end of Talheim
These signals are about 30 years old, and even though they are definitely not up to modern detailing standards, they look reasonably good.

Yes, that Roco 211 still needs detailing...

Wednesday, July 27, 2011

Talheim Local Panel - part 5

The remaining buttons are now connected, and I can finally reset a badly chose route before throwing it with the left black button in the back. The black button right next to it will be used to toggle whether to set signals, too, when setting a route.

One minor snag is that the cross-overs at the bottom don't throw consistently in unison as I press the button to toggle them. I'm not sure yet if the computer is sending the turnout commands too fast, or if the commands get lost. However, since I'm experiencing zero problems with other turnouts, and it's always the same turnout that doesn't toggle, I think communication is fine, and JMRI is sending the turnout commands a little bit too fast for the Tam Valley servo controller to handle. JMRI routes allow to add an extra delay between commands when executing the route, so I'll try that later tonight.

I'm very tempted to pull out the semaphores and connect them, even if only temporarily ... It would substantially change the look and feel of the Talheim area, and would be really cool. Hmmmmm...

Update 23:00:
My stupid perfectionism found a few issues:

  • The old signals need a color refresh. Particularly the red color on some of the semaphore arms is in sad condition.
  • To my utter surprise I'm short 2-3 signals. Had I realized this earlier, I could have picked up a few real cheap last weekend at Eurowest... However to my delight I noticed that there are two signals that can show Hp0/Hp1/Hp2. Those will be used as entry signals to Emsingen, and/or maybe Talheim. Though for Talheim I can pretend the entry signals are inside the tunnel.
  • The semaphore masts feel small, but it turns out that a mast height of 10.5cm above the bottom of the rail works out to just over 9 meters in HO which is right in the middle of the prototypical heights of 8 or 10 m. I'm mounting the semaphors lower than intended to partially hide the huge Maerklin twin-coil machines in the base.
  • In some locations the semaphore arms are partially above the neighboring track. My literature says that is prototypical as well. Tough one location in Talheim is very tight, and the prototype would have used a different kind of mast with tracks that close together.
  • At least one track in Emsingen needs the semaphor location adjusted by a few centimeters.
Just toggling the cross-over turnouts in Talheim doesn't give reliable results, apparently because the QuadP can't handle the turnout commands right after each other. Adding a delay helps, but it looks like there's a race condition around releasing the sensor button, so the logix still didn't work correctly.

I broke up the one logix into three: One logix to toggle an internal sensor based on the pushbutton sensor going active, and one logix each to explicitly set the turnouts to thrown or closed based on the state of the internal sensor, including a small delay between the turnout commands. That works very reliably, and I couldn't get the turnouts to go out of sync any more.

Sunday, July 24, 2011

Talheim Local Panel - part 4

Yup, still spending time on this. Tonight I finished up the basic JMRI logic, and even came up with a way to conditionally throw signals as the routes are set based on the state of a sensor. As usual apparently I'm over-engineering this ... again. Oh well, we'll see how long this one will last.

Saturday, July 23, 2011

Seen at Eurowest 2011

To bad this was the only signal I could find on the whole layout, and it was always green :-)

Thursday, July 21, 2011

Talheim Local Panel - part 3

It's starting to come together. The bad push-buttons are swapped out. The buttons for locally controlled turnouts are added, as well as the LEDs put in place for show (though not connected yet).

Wednesday, July 20, 2011


I cleaned up the remaining mess from the 63400 testing. All systems are back to go, ... well, with the exception of the Talheim panel. The push buttons are in, plus some LEDs, electronics hooked up, too.

I'm now writing the routing logic in JMRI to get two-button routes. I.e. the way it works is that the operator presses one route button on the panel to start the route. Then another route button for the end of the route, and JMRI will align the switches (and later semaphore signals) accordingly. Since Talheim has tracks going into 5 directions there's no way to use the more regular way to control turnout ladders with a single button per route.

Some of the cheap push-button switches I use are already breaking, partly because I need to be really careful when soldering wires to the connectors, since the casing is plastic and easily melts, displacing the contacts inside. So annoying.
Does anyone have a source for reliable push-button switches that don't cost the world? The ones I use right now are similar to these.

Thursday, July 14, 2011

Google Apps transition test

I completed transition of my Google Apps hosted domain to Google's new authentication system. This is a test post to make sure, Blogger still works with the new credentials.

Tuesday, July 12, 2011

Saturday, July 09, 2011

LPAC @ Belwood

Today was the last regular swim meet before Champs. Both Tatjana and Pascal are swimming great.

And I finally managed to convince kdenlive to not letterbox the funky SD mpeg2 wide-screen videos that my Panasonic SDR-H80 camera produces. I'm setting the project settings to edit in 720p, import the 480p SD material, force pixel-aspect ratio to 1.19 on the clips, edit the video, then finally render into HDV 720 30p 2-pass. This produces ginormous videos (sorry, storage and processing folks), but the final quality doesn't suffer nearly as much as when I edit and render in straight SD.

Thursday, July 07, 2011


What do you do when you're stuck at an airport overnight? ... Have some fun!

Tuesday, July 05, 2011

LocoNet issues with Uhlenbrock 63400

Wow. This is really confusing. 

The Uhlenbrock 63400 is connected to its own 16V AC transformer. The Intellibox is on its own transformer as well. The 63400 is connected to the Intellibox via short LocoNet cable. The track connection from the Intellibox is directly connected to a portion of the layout that doesn't have trains on it (one of the three power districts on the layout). And LocoNet communication is still affected.

If I unplug the transformer connected to the 63400, LocoNet communication is still affected.

If I remove the 16V AC feeder to the 63400 (i.e. make it a passive device), no issues.

Similarly, there are no issues if I remove the tracks from the Intellibox track connection. 

I.e. the mere fact that about 40ft of Maerklin track and the respective bus wiring is connected to the Intellibox moves the needle from "LocoNet works" to "LocoNet doesn't work".  I double-checked the cabling to ensure no crossed wires. There is no connectivity between the power districts, either.

While Uhlenbrock Support responded to my first two emails within a couple days each, the last update apparently got lost in the ether.

So, at the moment I'm putting this issue to rest and may get back to it when I build the panel for staging. Meanwhile, a TeamDigital SRC16 in combination with jmri logix is pressed into service to drive the panel. Not exactly the most economic or simplistic option, but it should do the job nicely, plus allows for some bells and whistles that make the Talheim block a bit more fun, too.

Unfortunately, the 63400 is too light as a paper weight.

Update 07/08/2011:
Uhlenbrock got back to me and suggested another test regime only involving the Uhlenbrock components and a short piece of track with a loco to weed out any influence from other components or wiring. I'll do that over the weekend.

Update 07/10/2011:
And that test had all the components working correctly.

I removed (even) more components from power and busses. By now all twin-coil turnouts are disconnected, the somewhat messy Talheim panel is completely removed and cabling cleaned up. No DS64's have push-buttons connected any more, DS64 power is disconnected (not that it matters for the test).

There are no obvious ways any more I can think of that would inject non-DCC power into the system, yet LocoNet communication is still affected.

I don't get it. There must be something wrong with how the track is wired, but I have no idea what it could be.  Need to find an oscilloscope, and re-read Wiring for DCC for any hints.