Model T Serial Cable

From Tandy Tech
Jump to navigation Jump to search

Summary

The ideal cable to go from a PC to a Model 100, 102, 200, or 600, would be described as "9F/25M serial null-modem full-handshake".

That is an uncommon configuration to find in a single cable without needing adapters or needing to build it yourself custom.

The serial ports on

TRS-80/TANDY Model 100, 102, 200, 600
NEC PC-8201, PC-8201A, PC-8300, PC-8401 / Starlet, PC-8500
Olivetti M10
Kyotronic KC-85

have a DTE pinout, the same as a com port on a PC, but with a female connector, unlike a pc.

This is a backwards configuration from everything else today. This may have been more common before the first IBM pc, and IBM came along later and created a new standard by using male connectors for the COM ports on their PCs, probably to distinguish them from the printer port which used a female DB25. But the universal standard today is that DTE ports are male (9-pin or 25-pin), and a DB25F (25-pin female) is either a parallel printer port or a DCE port on a peripheral (like a modem), not a DTE port on a host (a computer).

This means that most serial cables you will find today will need some sort of adapter to connect a "Model T" to anything else.
If you have a null-modem cable, it will probably have the wrong connectors, in which case you will need either a gender-changer or a 9-to-25 adapter or both.
If you have a cable with the right connectors, it is probably a straight-though "modem" cable, in which case you will need at minimum a male-to-female null-modem adapter.

So a single cable with both the right wiring on the inside, and the right connectors on the outside, without needing a stack of adapters, is a little bit special. They do make them, but you just have to search for them specially. Sometimes you can find a cable with that combination of properties by searching for a "Serial Printer" cable. Several known-good ideal cables are linked below.

The ideal cable is this:

DE9F-DB25M Serial/RS-232 Null-Modem cable
looking at cable end
female
male
DTE DTE
Signal DE9F DB25M Signal
RX 2 2 TX
TX 3 3 RX
DTR 4 6
8
DSR
DCD
SG 5 7 SG
DSR
DCD
6
1
20 DTR
RTS 7 5 CTS
CTS 8 4 RTS



Ideal Cables

These cables are wired null-modem, have all the connections for hardware flow-control, and the right physical connectors on both ends, all in one factory-molded piece. You don't need any null-modem adapters or gender-changer adapters.

No RI
S+PG
No RI
S+PG
No RI
Includes nuts (removable) on the 9-pin plug screws, so if you have the wrong kind of usb-serial adapter you can still screw it to the cable.
No DCD
No RI

Software Flow Control Only Cables

These cables actually work fine for everything(*). They are "less-ideal" only because the RTS/CTS hardware flow control lines are either missing or faked, or wired in some non-standard way like connecting DSR/DTR to CTS. But they have the correct physical plugs, and correct TX/RX wiring, and at least faked DSR/DTR, and so they work for software flow-control, which is just about everything.

Software flow-control is the only thing any software ever used on these machines, because the system ROM doesn't even contain anything that accesses the RTS/CTS pins. Neither TELCOM nor even BASIC has any hardware flow control features.

(*) There is one exception to "everything". Even though the system ROM contains no code to use the RTS/CTS pins, the hardware does actually connect the RTS/CTS pins to the UART, and it is possible to access the UART registers and employ those pins from a machine-language program. The only software I know of that does this to date is HTERM. So these cables are NOT suitable for HTERM. They ARE suitable for TELCOM and TPDD_Emulators.

Extra note about TPDD, TPDD clients all require that the DTR pin be connected to something. Either shorted to the machine's own DSR, connected to the other ends DSR, or even wierder setups like connecting to RTS+CTS, but it has to be connected to something. A 3-wire cable can do TPDD, but it still needs at least DTR shorted to DSR inside the DB25 connector shell in addition to the 3 wires going to the other end of the cable. These cables all work for TPDD.

DCD<-->RTS+CTS
RI<-->RI
DCD<-->RTS+CTS
RI<-->RI
DCD<-->RTS+CTS
No RI
DCD<-->RTS+CTS
No RI
DTR<-->DCD+RTS+CTS
No RI
DTR<-->DCD+RTS+CTS
No RI
DCD<-->RTS
DSR+CTS<-->DTR
No RI
This is a very unusual type of cable. If you don't have an HP or IBM plotter, I would not recommend getting this. But, strictly for the TX/RX and the physical connectors, as long as the software is ignoring RTS/CTS and DSR/DTR, it does work, so if you do have an HP or IBM plotter, then you could use this for both your plotter and your M100.
DCD<-->RTS+CTS
RI<-->RI

USB-Serial Adapters

Any one will work well enough. But some considerations are:

Models based on FTDI chips are generally better than the ones based on Prolific chips.

Try to get a usb-serial adapter that has nuts, not screws.

Here are a few ideal examples, with FTDI chips and jack nuts.

Legend

For all the cable wiring notes above, the notes indicate how the cable differs from a canonical RS-232 DTE-to-DTE null-modem reference.

RS-232 full-handshake null-modem


Any signal that is not mentioned, is wired according to the reference. IE, if TX or RX is not mentioned, then TX on one end of the cable is connected to RX on the other end of the cable.
Unless otherwise noted, the indicated wiring is symmetrical, the indicated connections are the same on both ends of the cable.

S+PG

The cable Shield (& connector shell) is connected to Protective Ground or "frame ground" (pin 1 on the DB25 connector, no pin on the DB9 just the connector shell)

RI<-->RI

Ring Iindicator is connected to Ring Indicator
DB25 pin 22 (RI) is connected to DE9 pin 9 (RI)

DCD<-->RTS+CTS

Request To Ssend is connected to Clear T Send on one end, and both are connected to Data Carrier Detect on the other end.
DB25 pins 4 (RTS) & 5 (CTS) are connected to each other, and to DE9 pin 1 (DCD).
DE9 pins 7 (RTS) & 8 (CTS) are connected to each other, and to DB25 pin 8 (DCD).