85SES3 Display Module Tester V3

Following on from the Train Describer Display Module Clock project I thought I'd resurrect another old piece of kit from around 1985, the TD display module tester. I thought it would be a good idea if I get could at least one example of display module from each version of our Train Describers all lit up at once! From Aviemore through to Cathcart, that's around six different modules.

Around 1985 we were building and commissioning Dundee TD, we had taken delivery of about fifty display module PCB's and each one required testing prior to soak testing and installation. A simple test setup was required, so we designed a simple circuit that would allow us to enter any character in any order without the need of having the TD card frame, keyboard etc. available. The unit was stand alone, it had it's own PSU's and only needed a single cable to plug between it and the module. Most faults found were either in the ribbon cable to the display front PCB or with the LED Matrix. When later display modules required a different a connection only a cable was required with the appropriate connector i.e. Dundee used 'Ringlock' and Carhcart used 'D' type connectors.

Looking at the circuit there will only be a few small changes to build it today. The Hall Effect switch used to initiate the transmit of a character is no longer available and the 'computer' interface will be updated to work with a RPi Pico instead of the giant MDS of yesteryear - The MDS was disposed of many years ago.

I replaced the Hall Effect switch with a basic push switch and a spare 'D' type flip-flop to debounce the contacts. The 'computer' interface was also a simple enough change, the Raspberry Pi Pico has 3.3V outputs so all I had to do was buffer these with with push switches using 74HCT08 AND gates. Usefully 74HCT inputs 3.3V logic tolerant. I also added a 4x20 line display for the RPi Pico.

The build!

I redrew the circuit and updated it with my changes, then designed a PCB. It turned out a bit unussual for me as the components are mounted on both sides of the board.

The PCB with it's cutout for the display and a mockup of how it will look once built.

Before adding the IC sockets I thought I'd select the componets for the oscillator, to get close to the 80KHz I used a 33nF capacitor, a 270R resistor and a 120R selected on test.

At this point I soldered the headers onto the Pico and the PCB.

Unfortunately once the oscillator timing components were soldered on the board the frequency changed. I swapped out the 270R resistor for a 330R and replaced the 120R with a 100R multiturn preset potentiomenter. This made it very easy to calibrate.

Next the switches went on and the IC's went in.

Then the bit switches went on and the 4x20 display is fitted.

Testing commences.

The hardware logic worked as expected but there were a few silkscreen errors that I rectified by sticking new labels over the screen printing. It's not worth ordering new boards for just these errors. I can set up descriptions one character at a time on the switches using the character binary codes so it's on to writing the program for the Raspberry Pi Pico. With he Pico I'll be able to send the full four characters fast enough for the description to look as if they appear simultaneously.

Comparing the character set for the Aviemore single character display and the 8748 four character display shows that there are differences that will need to be catered for in the Pico.

The software is ready which can run several different tests, scrolling text or just simply displaying train descriptions. The Aviemore TD required a different character set so to simplify the final display with all modules running simultaneously I built a second test unit.

Restoration of the display modules

First module up is the Aviemore single character display c.1980, this module was returned as faulty sometime in the late 80's. A replacement would have been sent out with this one, berth C335/C334, being put in the queue for repair. Well today (June 2021), some 30 odd years later, it has reached the front!.

The first picture shows the fault on the display, three columns not illuminating. Dismantling the unit and checking all the tracking on the pcb's and display front did not show up any cracked tracks or dry joints. Next I checked the display front connectors for discolouration, no problems there so it was time to bring out my logic probe. Checking the column drives did indeed show there was no drive signal so the obvious course of action was to reseat the 74LS138 in its socket which actually fixed the fault!.

A quick clean of the faceplate and filter then put back in its case and its ready for the big display..

Next up is the Cathcart display c.1987.

A similar story for this module, berth M142, came back faulty after years in service..

The fault, as can be seen, was a simple one. A faulty LED matrix and simply replaced.

Third up is from the Dundee TD display module build in 1985. This one worked worked straight out of storage. The subtle difference between the Dundee and Hight St modules is on the matrix PCB, the ribbon cable connector on the Dundee module is on the rear of the board.

Fourth up is from Hight St. TD c.1983.

I've been using this one for testing the Display Tester Unit so it's working ok.

Fifth up is the Hot Axle Box Detector in-line display used for Auchengray HABD at Motherwell SC. c.1981. Not many of these went into service, only two plus two spares!. The other twenty odd were kept in storage but were eventually superseded by the microcontroller based display.

The module I had was the prototype which shows us trialing the red and green LED matrices but I seem to have miss placed it but luckily I have found a bare main board PCB.

So I've started to update the design drawing with all the chip pin numbers and create a rough PCB overlay. There is no LED PCB which isn't a problem as I cannot find the original LED's on-line so I'm in the process of designing a new front PCB to be used the red versions of the High St./Dundee matrix. At the time we had settled on using red LED's to differentiate the HABD alarms from TD descriptions.

I'm also in the process of re-creating the character table to be programmed into 2716 EPROMs. This data has been lost so when tested the .bin file will be added the system software zip file.

Six, seven and possibly eight - Two Westinghouse Brake & Signal Modules which I think use the same data input configuration, one we used on the Hyndland TD expansion and the other from an Edinburgh TD fringe Box. c.1980ish and an in-house built module from either Eastern, Western, Midland, or Southern region - I don't recall which one!