Model Train Journal

All:

I am wondering if a serial protocol is employed by the Command Base to transmit the TMCC signal?

And, if that is true, an RS-232 or RS-422 repeater would be useful in extending the signal.

George

Ummm, Do you mean the radio signal broadcast through the track and house wiring or the signal sent out via the serial port on the command base?

The Command Base does use an RS-232 port. It communicates at 9600 baud, 8 data bits, no parity and 1 stop bit. It was originally intended to allow a user to connect a personal computer and use software to send commands to the base for controlling trains/layouts. As time went by it was "expanded" by other vendors to allow control of devices not originally sanctioned by Lionel as all of their devices used the RF signaling approach. IC controls used the RS-232 port to control their devices. Z-Stuff used a similar approach on some of their switch machine controllers (and required an opto-isolator/driver wire to accomplish this).

RS-232 was originally designed to connect a teletype machine to a modem, period. Two devices, one wire, closely spaced. With the advent of personal computers it was pushed to connect PC's to modems. There is a transmission rate limit of about 20kb and a cable length limit of about 15 feet (depending on the quality of the wire and skill of construction this can easily be extended to 30'-50'). RS-232 uses one wire to transmit, one wire to receive, and one wire as a "signal ground". There is a set voltage/logic limit and it is referenced to that signal ground. RS-422 was developed to allow for higher transmission rates (10mb) and longer cable length (~4000 feet). RS-422 has a signal ground but it uses paired wires for transmit and receive with a relative voltage/logic level between the paired wires. The wires are twisted pair so their is less of an issue with noise/interference. The signal ground is used to support a limited bus architecture in that one device (master) can transmit to multiple serially connected slave devices.

You can see that RS-422 is probably a better solution for use for device control but it isn't what Lionel used on the Command Base. You can install an RS-232 to 422 converter but you will have to but an inverse device to knock it back to RS-232 on the other end as most devices are expecting RS-232 signaling protocols.

chuck wrote:Ummm, Do you mean the radio signal broadcast through the track and house wiring or the signal sent out via the serial port on the command base?

The Command Base does use an RS-232 port. It communicates at 9600 baud, 8 data bits, no parity and 1 stop bit. It was originally intended to allow a user to connect a personal computer and use software to send commands to the base for controlling trains/layouts. As time went by it was "expanded" by other vendors to allow control of devices not originally sanctioned by Lionel as all of their devices used the RF signaling approach. IC controls used the RS-232 port to control their devices. Z-Stuff used a similar approach on some of their switch machine controllers (and required an opto-isolator/driver wire to accomplish this).

Chuck, do you mean Z-Stuff's DZ-2001 Data Wire Driver? I use that device. In fact, Panhandle 2.0 has 3 of them to deal with the large and spread-out nature of the switches. I would dearly love to find a way to extend the range of this device. I've been told the range is limited to 75'.

chuck wrote:RS-232 was originally designed to connect a teletype machine to a modem, period. Two devices, one wire, closely spaced. With the advent of personal computers it was pushed to connect PC's to modems. There is a transmission rate limit of about 20kb and a cable length limit of about 15 feet (depending on the quality of the wire and skill of construction this can easily be extended to 30'-50'). RS-232 uses one wire to transmit, one wire to receive, and one wire as a "signal ground". There is a set voltage/logic limit and it is referenced to that signal ground. RS-422 was developed to allow for higher transmission rates (10mb) and longer cable length (~4000 feet). RS-422 has a signal ground but it uses paired wires for transmit and receive with a relative voltage/logic level between the paired wires. The wires are twisted pair so their is less of an issue with noise/interference. The signal ground is used to support a limited bus architecture in that one device (master) can transmit to multiple serially connected slave devices.

At the beginning of my career in 1980, I worked for Digital Equipment Corporation. Their serial line drivers (DZ-11C) were capable of driving a signal up to 1200'. They were that good. We regularly specified shorter distances for "dumb" terminals connected to computers.

chuck wrote:You can see that RS-422 is probably a better solution for use for device control but it isn't what Lionel used on the Command Base. You can install an RS-232 to 422 converter but you will have to but an inverse device to knock it back to RS-232 on the other end as most devices are expecting RS-232 signaling protocols.

Thank you for the explanation.

George

I have not run across any documentation on a 75 foot limit on the serial cable. That's a pretty long cable. The wiring isn't indicative of how the data communication is being run. The wire is actually being used as an electrical bus. The TMCC commands have the device type/address embedded in them. The system just blasts the data stream out and (in theory) only the addressed device should respond, everything else should ignore it. What I am trying to say is that you can run a pair of short wires for data/common to a terminal strip and then run the feeds off of that. You don't need to run the wire from switch controller to switch controller daisy chain other than a way to save on wire. There is an exchange from CTT's forum on wiring up the DZ-2001. It may have some useful information

http://cs.trains.com/ctt/f/95/t/254298.aspx

If Lionel had implemented RS-422 this would all be moot BUT it would probably have increased the complexity of the wiring and cost of the devices to support the logic.

We used to have to use line drivers before we switched over to 10 base T ethernet. There was a "black box" (literally) on the computer side that connected to the RS-232 interface and then telco wiring to the comm closet where there were interface boards which consisted of eight "modules" on a big perf board that mirrored the individual black boxes and broke out into a 50 pin telco cable on the back side. Those telco cables mated with another 50 bin telco cable with either an octopus cable of 8 DB-9's for interface to the VAX serial ports or they plugged into a 50 pin telco interface on a ethernet terminal server. We had some of the earliest Cisco terminal servers in service. One didn't even have a backplane. Just a bunch of ribbon cables and molex connectors running to the network interface and power supply. First on was called "Cisco", second one "Pancho". Two of us hand made every one of the octopus cables.

Chuck,

Thanks for the link. I have wired several DZ-2001's, but this had some other very useful info I had not seen.

Kudos for making cables. Way beyond my skillset. Do you remember a Pittsburgh company called "Black Box"?

George

Do you remember a Pittsburgh company called "Black Box"

Remember the company, bought a bunch of stuff from them for work and personal use. Never made the connection with Pittsburgh.