- February 2
System 3 Progress Report
So, today, after everyone else had left, I ran four trains around in a loop, all
at the same time. How did I manage? Jeff was listing off some features he'd like
to see, so I just went ahead and added one.
Sys3 now will stop the train automatically and flash the cab signals if it gets
a red light, until either: the user presses the (i guess it can be called
call-on) button, or the signal changes, at which point it resumes its previous
speed. This makes it actually possible to run many many trains, and even insert
more into the loop while the rest are still running, without having to worry
about them running into each other. Very cool. :) Now we just need more cabs so
you can run multiple trains without having to use the text console.
In other news, the next three block cards are on their way to being completed
(which will make 8 block cards, 64 blocks), which means soon S-yard will be
operational as well.
- February 6
John Shriver paved the carbarn ladder tracks and a small portion of the
trolley loop that is near the turntable. Andy Miller painted some
locomotives. Ken Terrell tested a three-car articulated set, one of three
he bought at the Springfield show.
Roger Neumann continued soldering components onto the two block cards he
has been building. James Knight and Tom O'Reilly debugged some track
continuity problems that were affecting System3 performance. For part of
the evening there were TWELVE trains running simultaneously, an all-time
record for TMRC. John McNamara worked on a high-level description of
System3, and here it is:
The principal component of System3 is the Server, a PC running the main
control program, which is written in Java. The Server is connected to an
RS422 serial bus that connects to:
1) The IMP
2) The Block Cards
3) The Cab Controllers
The IMP is a PIC microprocessor that sends data to the Switch Cards via a
special shift-register-like bus. Each switch card can operate up to 8
switches (turnouts) on the layout. The IMP also controls the aforementioned
RS422 serial bus via a token passing scheme.
The Block Cards each use a PIC processor to control the generation of power
in 8 blocks. Unlike System2 where each control box ("cab") had its own
power source ("throttle"), in System3 each block has its own pulse power
throttle that receives control information from the cab controlling that
block. Thus one cab can control several throttles. The association of cabs
to the pulse power throttles in the blocks is provided by the Server.
The Cab Controllers use PIC processors to monitor the speed and direction
settings in the hand-held devices used by layout operators. The updated
speed and direction information is forwarded many times per second to the
throttles in the blocks occupied by the train.
As a train moves into a new block, the program in the Server causes the cab
controlling that train to become logically associated with an additional
throttle/block on a block card. The speed and direction information
provided by the cab, via the cab controller, is forwarded to the newly
associated block. Blocks that are no longer occupied by the train have
their throttles set to zero.
Switches are currently operated by seperate PC running a switch control and
layout display program called RSIG. Since these PCs control the switches,
and did so before the advent of the Server and the IMP, they are presently
connected to the shift-register-like bus that interconnects the switch
cards. Eventually, they will have an Ethernet connection to the Server so
that they are coordinated with the database contained therein. Switches
will be controllable via these computers, via keypads on the cabs, or via
layout pushbuttons that are monitored by the switch cards, which can
forward button-push information to the Server and IMP.
Status: Switch cards for the current layout have been constructed and
installed. More will be needed as the layout grows. With the completion of
a few more block cards, the same will be true for block cards. Cabs, cab
controllers, and the software to operate therein are the next parts of the
System3 project to be completed. Assistance in any of these construction
and programming efforts is most welcome!
- February 13
John Shriver's trolley system paving work has been completed in the car
barn area and is now extending across the Whatahack River to the loop area
immediately adjacent to the roundhouse area. Tonight, John took the initial
paving step, which is the construction of sidewalk supports that form the
boundaries of the paved area. Andy Miller continued his painting work on
some North Shore locomotives and joined with Ed Drozd to refill the Coke
machine. Ed also did some work on the KATO SD40's that he purchased for the
Club. John McNamara worked on his TNP superliners, installing new
trucksets. John Purbrick installed a set of jacks for System3 cabs. James
Knight and Alvar Saenz-Otero worked on Club finances. Brandon Gunn and
Roger Neumann continued their construction of a couple of System3 block
cards. Malcolm Laughlin tried some LED mounting alternatives for his
proposed Berkmannville control panel and consulted with various people on
panel construction alternatives. Tom O'Reilly demonstrated a nifty set of
cab control windows on the Server. Each is labelled with a train
identifier, such as "Amtrak Loco", has a speed control slider, and buttons
for Reverse, Stop, Dropout, and Foo. The demonstration included multiple
trains operating with the System3 "auto-stop on red" feature.
- February 20
The trolley line that serves downtown Gifford City runs between two loops.
One loop is in the downtown area and was brought to the new layout from the
old layout in Building 20. The other loop is near the roundhouse in a (not
very classy) residential area. Street paving activity, in preparation for
sidewalks and trolley poles, has been gradually moving from the downtown
area toward the residential area. To secure the new paving to the plywood,
in the same fashion as elsewhere, John Shriver installed round-top staples
like those used for telephone wiring. Andy Miller found a multi-story
residential and storefront building that was suitable for installation on
the new street. It was slightly too deep, so he shortened it. Andy also
repaired three freight cars and made some sketches of the proposed new
More buildings are needed in this area, as are volunteers to build them.
Work on the new System3 control system is continuing on several fronts.
James Knight did some work on it this evening, and Roger Neumann wired some
network connectors for the block control cards. Malcolm Laughlin continued
his work on a Berkmannville control panel. It features rectangular yellow
LEDs in the track diagram that indicate switch positions. Yellow LEDs and
red LEDs in the track diagram indicate which blocks are occupied and the
direction in which each block is powered. It is going to look really
spiffy. John Purbrick reconditioned several TNP cabooses such that they
hold together, their ladders are intact, and their LED marker lights don't
flicker. John McNamara did some work on his TNP superliners, under the
supervision of Tom O'Reilly.
System3 received another test - a 37-car freight pulled by two Athern
Genesis SD70M's brought in by Dave Mitton.
- February 27
John Shriver completed the paving around the trolley loop nearest the
roundhouse. He's now got this down to a science; it looks great. John
Purbrick installed an additional switch control card in the Berkmannville
area. He and James Knight also did some work on the block card feedback
feature. John McNamara loaded the Coke machine, which was once again
gratifyingly close to empty despite being loaded on Saturday. He also
brought in some plastic solvent for Genya Zaytman to use when he resumes
work on his mill building. Malcolm Laughlin continued using 3DPlanIt to
draw up proposed versions of the new freight yard. Mike Patton dropped by