I’ve been trying to get “educated” on this topic for a while now.
I feel there is a major flaw in the appoach, and it is causing, in my opinion, the inability to zero in on a good solution.
I have been in the computer/electronic design business about 30 years. This is as a system architect, designer, and “working” manager of engineers.
One thing I learned early on, as a junior engineer, was that if you are going to design anything, you need to work from a set of specifications.
The specifications are a result of requirements. We further broke requirements down into 2 phases, one at a high level, and one at a more detailed, architectural level, that explained the basic philosophies of how we were going to “get there”, and any unique or special technologies involved.
Stan, where is the list of requirements for this proposal? Obviously one very high level, general requirement is to make a standard interface to service many facets of the hobby.
But where is the “list” of things that need to be supported/provided?
Dave brings up one VERY fundamental requirement: The current handling of the interface.
The connectors used have a finite current capability per pin. This hardware was NEVER designed for a high current application.
While “doubling” of pins has been used successfully for years on things like PCMCIA, PCI, and other standard busses, it is not something just as simple as two 3 amp pins EXACTLY splitting a 6 amp load. Even a junior engineer will tell you that slight differences in resistance will make the porportional difference in current flow. (Thus having to engineer an extra margin in).
To me, I keep coming back to the discussion that is swirling around, and cannot “land” because there should be a list of requirements that everyone should agree on BEFORE someone designs the result.
Have you such a document? If 90% of the community can work with 6 amps total, and each pin can handle 3 amps, then there would not seem to be a problem with the current handling part of the design (from the rails) IF the requirement is 6 amps.
The fact that ALL (to my knowledge) LS high current decoders have only screw terminals or soldered wires for track and motor connections tells us something. The “bulletproof” NCD D808, with a 30 amp stall has screw terminals. Whenever someone with a USA trains diesel wants a decoder, this is a consistent recommendation, not the 4 amp D408. Maybe overkill, but overkill gives RELIABILITY.
I do not buy the statement that “it’s because there is no standard”… I maintain it’s because there is no high current connector in common practice. These two are NOT the same statement. (I am not accusing anyone of making this statement).
Similarly, stating that chuff inputs and other inputs must be referenced to ground should be stated as a requirement if it makes sense. BUT, I maintain this would be a result of a requirement that certain “Industry Standard” ways of implementing these functions are supported. (Data from a survey of how the industry leaders do it would be a starting point if it was me making these requirements).
Stan, to garner support, I would strongly advise we all take a step back, and come up with a requirements list.
I think it would be much more straightforward to get the requirements first, and then find the design(s) that support them. Once everyone agrees on the requirements, then the possible implementations become a very objective, unemotional engineering task.
Regards,
Greg
