oddball televideo problem

From: Tony Duell <ard_at_p850ug1.demon.co.uk>
Date: Thu Jan 15 17:08:21 2004

>
> > I've had problems in the past with voltage levels on RS-232.
>
> > One device runs -12/+12V, the other 0/+5V, and they refused to talk
> > bi-directionally. Putting in a homemade level shifter (forced the
> > 0/+5 to -9/+9) made it work.
>
> I think a sending device is required to swing at least below -3V and
> above +3V for it to be RS-232 (though I may have the voltages wrong,
> it's been a while - but I'm sure they're symmetric about 0). I ran

I beleive you are correct. One state is a -ve voltage between -3V and
-25V, the other is a +ve voltage between +3V and +25V. Anything between
-3V and +3V is undefined.

This means that devices using +/-12V levels meet the spec. So, actually,
do +/-5V levels. But TTL levels _don't_.

However, the well-known 1489 receiver chip used a lot of RS232 ports has
a threshold above 0V, and will normally accept a TTL input (on the RS232
side). So some badly-designed devices output 0 and +5V (or thereabouts)
and hope the device they're conencted to uses a 1489 or similar. Of
course this doesn't meet the official spec, but then so few modern
devices do. The problem comes when you connect such a device to a serial
port which meets the RS232 spec, but which has a threshold at, or just
below, 0V (I believe Apple Mac serial ports are like this).

[Actually, I'll admit to having fed the output of an HCT04 (or similar)
into a serial port once. My excuse it that it waas a one-off test board,
and I had the schematics of the serial card it was driving (which showed
a 1489 reciever). No way would I do this in a device to be used by others!)

-tony
Received on Thu Jan 15 2004 - 17:08:21 GMT