Can there be a standard electrical interface for "plug and play" electronics in locomotives?

Some time ago I joined an NMRA working group that was hoping to find a solution to a standardized electrical interface to Large Scale locomotives.

The idea was to write the requirements for this standard.

Since then, some very interesting things happened.

It became clear that the manufacturers and just a few individuals were driving this and Bachmann is making some changes. Aristo, of course, has not changed anything, especially since they do not support sound, so no extensions to their socket.

USAT was not doing anything, and has always seemed disinterested.

Bachmann was the only major manufacturer that seems interested, and was getting a lot of direction from Stan Ames.

So, since I had a full time job and cannot play politics all day, and the working group was a sham. What else is new. No wonder G scalers hate the NMRA.

So for now, I think the answer is NO! A very revealing and disheartening experience.

Update, years later, the situation is the same.

The Bachmann K-27, the first "new" implementation of the Aristo Socket:

A version of an idea for the socket was given to Bachmann for their K27. Well, as of today, February 2008, this has been a very good learning experience in what NOT to do!

Bottom line is that in order to support the various types of power and control systems, you need to make careful choices.

Now the following may upset some people, but I'm going to point out the various aspects of the Bachman implementation, and show why they are problematic.

"Logic problems"

For most systems, like lights, there is often a common between them. For example in typical DC circuits, negative is used for common between lights. But in DCC, positive is the common. (There is a reason for this).

The K is wired like a DC, that the lights are effectively positive common, so to turn on a light, you bring the light "wire" to ground. (This is not completely true, more later).

This problem is further exacerbated by the fact that the light is turned on by a transistor. It is designed to turn on when you bring the base of the transistor to ground. To make it work to turn on by going high, you need to invert the signal, so you need another transistor. 

"Analog vs. Digital control of lights"

Well, there is another curve ball in the K. Turns out that instead of using a dropping resistor to drive the LEDs, Bachmann came up with a clever idea: use a transistor to switch the light on and off. I also believe that the transistor acts as a current limiting device.

Very cute and clever. Unfortunately a bad idea.

First, the obvious problem is that if you change the type of bulb, instead of changing a current limiting resistor, you need to "reprogram" your solid state "switch and regulator".

To pick a dropping resistor value, you only need Ohms law: V=IR. To change the circuit in the K you need to understand gain theory for transistors.

Also, you are still stuck with the logic problem, the system is designed to turn on the transistor (which turns on the LED) by going to ground (as mentioned above).

But the fun is not over yet. Most new controllers (DCC and otherwise) are capable of modulating the intensity of lamps and LEDs, by modulating the voltage (or current).

Well, this circuit is either ON or OFF, i.e. DIGITAL. So your headlight can no longer dim.

I believe this is a mistake.

And to top off the fun, certain DCC decoders use the headlight as a "load resistor" to dump off extra voltage and/or it's involved in transponding. I won't go into the details, but since the transistor that turns on the LED only draws a few milliamps or less, it does not "act" electrically like a lamp or LED attached to the decoder, and thus these functions do not work.

Chuff circuitry: 

In most sound systems, the chuff "trigger" is assumed to be a signal that is high impedence (open circuit) most of the time and goes to ground when you want a chuff. Some sound systems will handle you going from open to high. Also some circuits detect the level of voltage, and some detect the transition.

Once this is very clear, a circuit that requires about 5 volts to even produce chuff signals is not good for a track powered loco. You become forced to have an on-board battery to have the circuit powered up. (Take for example a DC powered loco with a  Phoenix sound system).

In order to support track powered people, the circuit needs backup power, i.e. it needs to function from 0 volts on the rails. Obviously when running DCC, you have power all the time. If you were battery powered, you need to use it to power the chuff electronics, and of course a way to shut it off when running, not to mention a regulator.

One other thing is that the K chuff circuit is reversed, in logic. Most sounds systems require you to reverse the logic with the addition of a transistor. Dumb. Shows what happens when someone that is not an electrical engineer trys to make electrical system specifications.

Conclusions of what is needed to standardize electronics:

We need to standardize the electronic interfaces in locos before we can even THINK of standardizing a socket.

For lights, we need analog control of the voltage, and we need to be able to facilitate changes between LEDs and lamps. We also need to have either a positive or negative common.

On the common: costs of wiring and the amount of wires being a concern, running both leads of every light back to a socket would be way too many connections. So some common needs to be picked. Obviously this is of no concern with incandescent lamps. It is a major problem with LEDs.

At this point, I think the best way may be to "pick" a common, and then be sure that individual lights can be "reversed" if needed to change the common. Whether this means all lamps plug into a board somewhere and can be reversed, or there are sockets, something needs to be done. This might not be easy to do due to costs. Technically, it's simple.

The next concern is accomodating different lights that need different voltages. One alternative is to allow the user to "reprogram" the current limiting/voltage characteristics himself. (Change out resistors on a board). I believe this is very dangerous, since most people would wind up damaging lamps with improper "programming" 

The best way is probably to pick a standard voltage to be applied at the "lead" for each lamp. I would suggest that this voltage be 24 volts DC maximum. If you pick a lower voltage, then you are in danger of destroying lights. Many battery power systems run at 21-24 volts. DCC runs in G scale at a minimum of about 20 volts, but better at the max of 24.

The light should operate in an analog manner, i.e. varying the voltage on the input lead should be capable of modulating the light output at some point. Different "curves" between LEDs and incandescent lamps is acceptable.

As of now, I have no hope that this logical approach will ever happen. What is happening is Bachmann continues to produce locomotives with various implementations, and we will see if the various "decoder" manufacturers will make a "version" to fit the upcoming "versions" from Bachmann.

I've deleted all my previous stuff about getting information from real users and manufacturers.

All that will follow is what is available and the changes to the "Aristo Standard".

Aristo socket:

This is the Aristo interface as was published on the Aristo site before they folded. It consists of 2 sockets with 2 sip connectors, one 12 pins and the other 10 pins.

Not all sockets are consistent in all locomotives.

The below diagram is looking down "into" the socket.

Aristo says 5 & 8 must be bridged for DCC, why? (Current idea is to allow the unit to function from the existing smoke on/off switch)

Upcoming socket in new Bachmann Forney:

Things that need to be addressed/ my comments:

One big problem I see is that there are only 2 pins for the motor power. This means a single socket pin must pass ALL of the current for ALL motors in the loco. This is clearly INADEQUATE. CLEARLY. Five amps continuous rating is probably marginal. With DCC decoders having stall capability up to 30 amps, this just does not cut it. The pins for the power input are doubled, so is it thought that the decoder and peripherals will take 50% of the total current? 

Stan says "An example part number for the male plug that goes on the decoder is Samtec TSW-112-07-T-S and an example part number for the female socket that goes in the locomotive is Samtec SSW-112-01-T-S"

The assumption that the chuff switch can always be switched to ground needs to be researched. It's my understanding that this is NOT correct. I will research and list this situation on existing decoders.

The assumption that the reed switch is active low is likewise not proven to be correct. It's not clear what this is proposed for, but some decoders use a reed switch to allow programming on a DC track by activating a reed switch to enter programming.

The control for a smoke unit is a ground. To be consistent this means that grounding this pin causes the smoke unit to be turned on. Again, this is an inference that is not necessarily supported by documentation. I believe this is a minor point, and can be accomodated in most cases.