For their first half a century or more on the railroad scene, our electric locomotives and internal combusion engine powered locomotives relied heavily on relay logic electrical control technology. This provided a significant advantage for preservationists because relay logic is reasonably understandable and also usually maintainable through substitution. Generally no matter what kind of switchgear device could not be found any more, there was something else in the electrical suppliers catalogs that could do the same job, although sometimes bringing a few extra sets of contacts to the application.

That began to change in the 1960s and 1970s when transistor equipped "black boxes" arrived on the scene, to be followed by modules that combined many of the excitation functions on the same card. When controlled rate of slip wheelslip systems gained prominence, they brought with them opto-isolators, solid state amplifiers, and other new devices that are now getting to be obsolete and are being replaced by newer systems in surviving locomotives. Finally the locomotive control microprocessor systems arrived, not just one or two but a dozen of them, many built by companies that are now merged or out of business, while a few others decided the locomotive business was not worth the trouble and left on their own.

So this brings us to the future of preservation, when the prize to be preserved is a locomotive with a half dozen computer systems all two to three decades old, filled with devices that are no longer manufactured and software that has been extinct for a generation or more. It is going to be a fascinating new world when you go to test your newly acquired prize and find out that something is wrong and there is no source of service replacement not only for the troublesome device but also for virtually everything else in the carbody. Sticking in a substitute relay with too many contacts will not be an option, making that locomotive into a viable operating display may require an entire spare microprocessor installation. Just something to think about. Relay logic systems, anyone?

PC

I have seen Allen-Bradley PLC's (programmable logic controller) used as retrofits on at least one GP40 and a couple of passenger cars. They have some inexpensive units and they are well known for supporting their stuff for years - you can still get parts for their circa-1980 PLC-2 lines - they still have software to program them.

Most electronics (that would end up in a locomotive or most industrial equipment) made up to about 1990 or so-or even later-are no problem to repair, substitute, or modernize if needed, for someone who knows basic electronics. The biggest thing that I would suggest is archiveing any ROM's on microprocessor controlled equipment. Most old microprocessers are still made, or could easily be re-created with newer units if need be, but the software can be lost forever pretty easy-it doesn't take much to zap an old ROM. If a replacement program can't be obtained, then you have to imagine what the programmer was trying to do and reproduce it...not easy.

There are some impressive stories in the memory banks of EMD and GE employees who worked on the early development of microprocessor control systems for locomotives. Available space and time make it impractical to tell them here, but suffice to say that a diesel electric locomotive is a heck of an impressive welding machine when something goes wrong with the sequencing of the switchgear for the transmission circuitry. There were some very spectacular events along the route to properly working locomotive microprocessor systems.

The high voltage side is not forgiving of mistakes.

PC

Preston, the problem is very much already here. One of our clients has a 1950's roadswitcher in the collection that was 'rebuilt' from original relay technology into a computer-controlled, repowered configuration. At the risk of a flame war I won't identify the locomotive and railroad, but some will figure it out.

They meant well. But the control system was nearly obsolete when installed, and now it's become glitchy with things like refusal to load and false wheel slip. The alternatives rapidly became whether to try to repair it or completely replace it. In the end, we determined that replacing it ran the risk of the same exact problem we already had, what would this be like in another 10 years? Would we we just as far ahead to simply replace the faulty cards (now effectively a custom build) or start over with somebody elses system that would also age out?

We rapidly hit all the issues you're describing with the repair parts - including a commentary by one of the original designers that reverse-engineering the original cards, making new circuit boards, etc., could easily run $250K. EMD turned us down cold with changing it over to -2 electricals due to the engineering cost.

So we're in the unenviable position of realizing that no matter what happens, we're going to face the same decision probably in another 10 years. For right now, the supplier can still supply original cards from stock. So we're buying a stockpile that will carry us for a while. But then what? So much of the original equipment was removed (including the prime mover) that going backwards isn't a rational decision either. The shop crew has observed that it's a whole lot easier to blacksmith and machine-shop the steam locomotives than to try to fix the computer systems on this rebuilt diesel.

Suddenly relays and original locomotive technology are looking very user-friendly. As the resident 'tech guy' I know I consider anything older than 3 years on critical microprocessor hardware on the planned replacement list, and at the same time I'm PM for several steam locomotive projects, some days the contrasts boggle me.

When we did the Savannah Streetcar hybrid it was entirely with Allen-Bradley 'off the shelf' stuff that we hope will survive well over time and we didn't try to make it any more complicated than we absolutely had to. EMS in Altoona did the design, and they kept it as much to standard industrial design as possible.

I'm no longer feeling so bad at the prospect of rebuilding this control equipment on a certain 1930 oil-electric!

Weren't we recently hearing about stranded trains that couldn't cope with the recent European blizzard because the digital control systems were not happy in the cold and wet? I think there's an article about an old school electric locomotive pulled out of a museum in Sweden in the current flimsies to plow out so the new ones with digital control could then get moving again.

I'd almost refuse to upgrade from old reliable relays - that you could jam a flag stick into and limp home of necessary - with anything you couldn't find a low tech solution to on the shop floor level. Marginal savings in efficiency are rapidly canceled out by total losses in availability.

A few months ago I lost an air handler at work that kept a substation cool......fixed by jamming an old worn out belt into the starter relay to hold it closed until the long job of troubleshooting the digital control system could be done. The joke was on the digital folks - the suck-in coil in the starter was gone. Hours to find that out, 15 minutes to change it. Was there a lesson learned? Jury still out......

dave

Hardly a new concept or unique to electronics: the PRR GG1's were felled en masse by a freak fine-powder snowstorm in February 1958. This led to the partial retrofit of new filters on many of them (the huge boxy filter grids at the top of the noses, befouling the as-built Loewy lines--note that only one preserved GG1 had those refits, 4890 now at Green Bay.....).

As Randy noted, several of the early and prominent players in locomotive microprocessor controls who looked like they were going to be in it for the long run left the market after they had developed systems that are now museum pieces in and of themselves, with modules that can no longer be obtained, are impossibly costly to manufacture, and difficult to repair. In general they found that there is not such thing as a "universal" locomotive control microprocessor, that they all require extensive customization, particularly for retrofit applications, and that brings with it a high overhead expense of having experienced electronic technical people on staff and available for customer service. Curiously a couple of those outfits subsequently got "cold feet" about their departure and came sniffing around again to try a new generation of controls on locomotives. But the control market is pretty much owned these days by the OEM builders of locomotives, each with their proprietary in-house systems.

The economics of retrofitting a modern control system in an older locomotive are very challenging. Most management teams tend to look dimly on the idea of spending over $250K on a locomotive that has a book value of $50K.

PC