Which RS232 control signals do you really need?

[Dylan Hall]

I've built an initial version of a UART module for a 286 system I'm developing.

So far I've only connected TX, RX, RTS and CTS. This seems to be enough to get the module talking to my laptop.

This fits nicely with the MAX232 which has 2 drivers and 2 receivers.

However, I started reading about serial mice which need DTR and RTS for power and reset.

This got me wondering, what other devices might not work because I don't have the full set of signals?

Thanks,

Dylan

[Sergey Kiselev]

The ones that you've listed generally should be enough for serial communications with hardware flow control. Newer UARTs, e.g. TL16C550C (the 'C" suffix is important) can even be programmed to do automatic RTS/CTS.

Serial mice do "steal' power from serial port (as RS-232 does not include any power signals vs. XT, AT, or PS/2 keyboard/mouse ports). So they might use either RTS or DTR lines to get the power or both... I think usually there is a diode or a two are connected to these signals, plus a Zener diode or a voltage regulator to reduce the voltage to 5V or so.

For example, see the datasheet for the EM83701 mouse controller, it has the schematic.

If I recall correctly, the PC side of the RS-232 should have 3 drivers and 5 receivers. You can use two MAX232's but then you'd not have one of the receivers (and one extra driver). If you're not planning to use a modem, you can get away without implementing the RI (Ring Indicator) signal, and save a receiver that way. At least that what I did while back when building an ISA serial port board (see attached schematic).

Alternatively, if you have +/-12V available, you can use a GD75232 driver/receiver or similar. It has 3 drivers and 5 receivers as required for an IBM PC compatible RS-232 port.

-Sergey

[T Gerbic]

In most cases you just need TX, RX, ground, and a pair of flow control signals if there is a chance of needing buffering management (mismatched speed capabilities). Beyond that most people just force signals up, which can be done with pull-up resistors.  It is nice to have DTR, DSR and DCD(CXR) support but outside of providing power and possibly a reset, they were mostly for MODEM control or indicating something like a terminal/computer is connected and powered on. RI is only really needed if you have a MODEM for some situations but most newer MODEMs also dump out a text message when rung-up. 

It comes down to if you want to be able to just plug something in (have configurable signal paths/pull-ups) or if you want to just put out the RS232 standard pin-outs for a DTE end and let people wire their cables/adapters to their needs. 

So at a minimum you should implement four signals and maybe allow a couple of pull-ups, or go whole hog and implement eight (with two MAX chips) to be able to run anything.  Would be nice to show a dial-up connection operate from an old IMSAI/Altair to a machine acting as a BBS/mainframe just the way it did in the past, well at least at a computer fair. 

On Thursday, December 4, 2025 at 1:31:55 AM UTC-8 dy...@deedums.com wrote:

[T Gerbic]

I might be a bit different from most but if I designed a serial card, it would have full MODEM control.  I have a Cisco VoIP/PBX in my house and a telco line simulator, plus at least a couple dozen desktop, rack and custom, sync/async Racal-Vadic MODEMs. Got a few other vendor MODEMs as well.  Would be interesting to set up and play with some 202/V.23, full duplex and sync MODEMs again. At least for a little while. Don't think I remember any bisync or SDLC software for CP/M, but that would be interesting as well. 

Just thinking about the good old days...

[Doug Jackson]

I fed my family for YEARS by connecting íncompatable devices using RS232.  Devices ranging from modems to printers to lab gear, through to traffic counters.

The only thing that RS232 did well was specify the connector dimensions.

The starting point for a simple cable was

2 - 3

3 - 2

7 - 7 

4 linked with 5 on either side

6,8,20 linked with each other on either side - and a 3 wire cable

Then when you wanted hardware handshake, you tied 4 - 5 and 5 - 4 across the cable.

:-)

Kindest regards,

Doug Jackson

em: do...@doughq.com

ph: 0414 986878

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/retro-comp/52d9f900-e62c-4945-9a93-6d45985a7545n%40googlegroups.com.

[Jaap van Ganswijk]

Rs-232 is a mess as regards hardware. Later we used to rely on the xon/xoff system. When a system can't handle more input it sends an xoff character and when it can handle it again it sends an xon. Of course these characters need to be handled with priority and not be pulled through the normal buffering stages.

To view this discussion visit https://groups.google.com/d/msgid/retro-comp/CAM_9E62u3NQpZVMpGrk3-fxRfBHdRtkq2kSdh0yqSdx8FW%3DWJw%40mail.gmail.com.

[Dylan Hall]

Thanks everyone for your comments.

A follow up question. If I'm to connect more of the signals I'll end up needing more receivers than drivers. If I have a MAX232 where I'm only using the receivers can I omit all the caps related to the charge pumps?

Dylan

To view this discussion visit https://groups.google.com/d/msgid/retro-comp/CADNzunzxNWg3vbOEnWarYJB-kEo9_Y55jXk4bZxKkBYKZtmNuw%40mail.gmail.com.