Sarum Bridge - Electronics and Trains
Trains
This currently consists of a full 7+2 HST set and an assortment of 3D printed locos and wagons. Currently, I have completed assembly and painting of a rake of 26 four-wheel wagons, a short newspaper train made up of ex-GWR siphons and a BR Mk1 full brake, hauled by a class 55 and class 42 respectively. Another 30-odd wagons are painted and simply awaiting wheels and couplers, as are body shells for class 08, 24 and 25 diesels All are done in BR blue to match the HST. Most of the 3D items are by OzRail, with some by ivanf and CCE614.
The 4-wheel wagons were hand-painted with acrylics, then fitted with wheels and couplers. Most of these bits were from two packs of 4 variable-length coach chassis, with a few couplers bought separately plus the spare couplers from the HST sets. I tried the cheaper pin-and-socket type, but these are bulkier and come uncoupled too easily. I have a few of the fixed permanently-coupled couplers, and these do work well. I will be replacing most of the wagon couplers with these in the future (setting up short mini-rakes of 4-6 wagons). The wheels have to be the original Eishindo type, not the new ones available from tgauge.com. The Eishindo wheels have cups in the wheel faces that fit pins on the inside of the journal boxes rather than the opposite technique used in larger scales. The material used by the 3D print process is quite delicate, and I accidentally broke off an axle box from two wagons when I got tired and careless. I ended up having to omit the coupler springs, since these are so strong they lift the entire wagon off the track. Leaving the couplers loosely fitted works well, though coupling up is a delicate operation requiring a pair of tweezers to gently clamp the couplers together.
The diesels were also hand-painted, and look a lot better in reality than they do in the pictures. Some are awaiting chassis. The class 42 (Warship) and 55 (Deltic) bodies have a small strip of plastic glued to the inside of the underframe on one side which acts as a clip to hold the chassis in place.
This currently consists of a full 7+2 HST set and an assortment of 3D printed locos and wagons. Currently, I have completed assembly and painting of a rake of 26 four-wheel wagons, a short newspaper train made up of ex-GWR siphons and a BR Mk1 full brake, hauled by a class 55 and class 42 respectively. Another 30-odd wagons are painted and simply awaiting wheels and couplers, as are body shells for class 08, 24 and 25 diesels All are done in BR blue to match the HST. Most of the 3D items are by OzRail, with some by ivanf and CCE614.
The 4-wheel wagons were hand-painted with acrylics, then fitted with wheels and couplers. Most of these bits were from two packs of 4 variable-length coach chassis, with a few couplers bought separately plus the spare couplers from the HST sets. I tried the cheaper pin-and-socket type, but these are bulkier and come uncoupled too easily. I have a few of the fixed permanently-coupled couplers, and these do work well. I will be replacing most of the wagon couplers with these in the future (setting up short mini-rakes of 4-6 wagons). The wheels have to be the original Eishindo type, not the new ones available from tgauge.com. The Eishindo wheels have cups in the wheel faces that fit pins on the inside of the journal boxes rather than the opposite technique used in larger scales. The material used by the 3D print process is quite delicate, and I accidentally broke off an axle box from two wagons when I got tired and careless. I ended up having to omit the coupler springs, since these are so strong they lift the entire wagon off the track. Leaving the couplers loosely fitted works well, though coupling up is a delicate operation requiring a pair of tweezers to gently clamp the couplers together.
The diesels were also hand-painted, and look a lot better in reality than they do in the pictures. Some are awaiting chassis. The class 42 (Warship) and 55 (Deltic) bodies have a small strip of plastic glued to the inside of the underframe on one side which acts as a clip to hold the chassis in place.
Electrics & Control Systems
The layout uses a bank of four microcontroller-based PWM controllers, one for each track section, replacing an earlier set of home made T Gauge PWM controllers. Both types have performance equivalent to the Eishindo unit, but with full overload protection.
Since T gauge trains usually have multiple powered vehicles scattered through the train, short stop sections will not work well. The track is divided into 4 sections electrically, and the entire section is turned off when a train reaches a stop sensor (LDR) a couple of inches from the end of the section. Each section has a second LDR about two feet back from the stop sensor to provide a slowdown warning. The LDRs are mounted completely below the track, relying on light leaking through between the sleepers. Obviously, no ballasting at those places. LDR trigger thresholds are stored in EEPROM in the microcontroller, and can be recalibrated for new light levels by pressing a button on the control panel. This system worked satisfactorily for home use, but did not stand up well to exhibition conditions - uncertain lighting and people leaning over the sensors caused problems. A future upgrade will replace them with below-track infrared detectors with IR LEDs mounted under a pair of footbridges.
All stop points are in the station platforms, so that trains stop and start at the front of the layout where they can be observed. I originally intended to provide a two-stage slowdown feature using relays to switch in diodes. This looked practical due to the 4V track voltage of the PWM controller, but a few quick experiments showed that while 1 and 2 diodes give decent medium and slow speeds respectively, the slow speed running really suffers with the standard PWM controller (unreliable and very, very jerky). The directional LEDs are also greatly affected - dim with 1 diode, effectively off with 2. In the end I had to use three separate simple PWM controllers built from 555 timers, one for each speed, with relays selecting which one to use. This still turned out to be too jerky and unreliable, so I finally replaced the 3 separate controllers with a bank of 4 microcontroller-based versions with inertia, allowing smooth starts and stops.
It is common practice for block systems to try and make each section roughly the same length to ensure a steady flow of trains. For Sarum Bridge, the lengths are deliberately asymmetric: 3-3-3-1. The short section means that some trains will always be able to run through the station non-stop, even if they are running a bit slower than the others, and that there will usually be something either stopped or moving at the station.
A single PIC16F690 microcontroller handles the four blocks, with additional random station stop and two-speed (passenger train versus goods train) features. I have also set up for simple colour light searchlight signals using below-baseboard LEDs and fibre optics. These are in place and working - they are simply bent pieces of optical fibre with small washers as backshades.
The layout uses a bank of four microcontroller-based PWM controllers, one for each track section, replacing an earlier set of home made T Gauge PWM controllers. Both types have performance equivalent to the Eishindo unit, but with full overload protection.
Since T gauge trains usually have multiple powered vehicles scattered through the train, short stop sections will not work well. The track is divided into 4 sections electrically, and the entire section is turned off when a train reaches a stop sensor (LDR) a couple of inches from the end of the section. Each section has a second LDR about two feet back from the stop sensor to provide a slowdown warning. The LDRs are mounted completely below the track, relying on light leaking through between the sleepers. Obviously, no ballasting at those places. LDR trigger thresholds are stored in EEPROM in the microcontroller, and can be recalibrated for new light levels by pressing a button on the control panel. This system worked satisfactorily for home use, but did not stand up well to exhibition conditions - uncertain lighting and people leaning over the sensors caused problems. A future upgrade will replace them with below-track infrared detectors with IR LEDs mounted under a pair of footbridges.
All stop points are in the station platforms, so that trains stop and start at the front of the layout where they can be observed. I originally intended to provide a two-stage slowdown feature using relays to switch in diodes. This looked practical due to the 4V track voltage of the PWM controller, but a few quick experiments showed that while 1 and 2 diodes give decent medium and slow speeds respectively, the slow speed running really suffers with the standard PWM controller (unreliable and very, very jerky). The directional LEDs are also greatly affected - dim with 1 diode, effectively off with 2. In the end I had to use three separate simple PWM controllers built from 555 timers, one for each speed, with relays selecting which one to use. This still turned out to be too jerky and unreliable, so I finally replaced the 3 separate controllers with a bank of 4 microcontroller-based versions with inertia, allowing smooth starts and stops.
It is common practice for block systems to try and make each section roughly the same length to ensure a steady flow of trains. For Sarum Bridge, the lengths are deliberately asymmetric: 3-3-3-1. The short section means that some trains will always be able to run through the station non-stop, even if they are running a bit slower than the others, and that there will usually be something either stopped or moving at the station.
A single PIC16F690 microcontroller handles the four blocks, with additional random station stop and two-speed (passenger train versus goods train) features. I have also set up for simple colour light searchlight signals using below-baseboard LEDs and fibre optics. These are in place and working - they are simply bent pieces of optical fibre with small washers as backshades.