RS-232 levels to computer

Indeed, one could argue that practice in the PC and other industries, with their huge installed base, trumps whatever 232 spec is current, and any new implementation that acted only minimally compliant with the spec (eg. thresholds at +/-2.6V nominal) would be 'broken'.

One major advantage of the way the receivers generally work** is that the output state is known for a disconnected cable.

Current practice, in effect, imposes *tighter* constraints than the standard. This is not an uncommon situation.

Best regards, Spehro Pefhany

That's why the damned things act up when the peripheral and computer are plugged into different line circuits with significant voltage difference between GNDs- not enough noise margin. For high reliability, the bipolar drive should be used, and it should be as large as possible, +/-12V if feasible.

Yup. A lot of the time at most one of the devices (the computer) has the common earthed. Laptops (if you can find one with a serial port) and most peripherals are generally isolated from earth. USBRS232 converters are generally not isolated (the cheap ones never are).

Best regards, Spehro Pefhany

There's a specification reason for that... input float has to provide a "1".

(I designed both the 1488 and 1489 ;-)

...Jim Thompson

About once a year or so I have to deal with a broken so-called "232" interface. You can design anything you want, but if it doesn't work in accordance with the spec (I'll accept any version, not just the latest) please don't call it xxx-232, please specify what its real input and output limits are, and please don't try to sell it to me.

That's why I deliberately used the terms 'high' and 'low' rather than 0 and 1.

There are two elements here - what the spec mandates a device must do, and what a device must do to meet the spec requirement.

In the no-man's-land between, the spec does not mandate a specific response. But to meet the spec, a device needs to have both positive-going and negative-going thresholds (assuming hysteresis has been used) within that no-man's-land. Clearly those thresholds aren't going to be at +/- 2.999999v.

As Spehro's footnote asked:

I for one certainly don't. JimT, as the 1488/1489 man, would know about this, and a quick viewing of the data sheet for those early devices will say the same.

Crikey! Back in 1969 my degree honours project had us running RS232 at 9600 baud between two campus buildings that were several hundred metres apart. I shudder to think what the data ground line was carrying.

Was that possibly synchronous?

I don't think you correctly understood what I wrote. The poster to whom I was responding said many poorly designed interfaces will work swinging from ground to +5V. I said that many well-designed interfaces will work, with that swing, too, but that the spec doesn't guarantee it.

Is there some part of that that you disagree with? Is it your contention that the spec DOES guarantee operation with a swing from 0 to 5V?

--Mac

those

No

I believe I did.

Quite the opposite, BUT ...

The reality lies in what the spec - and I would argue quite intentionally - does NOT guarantee or state.

It is years since I actually had a copy of the 232 tome before me, so these numbers are from memory. The four pivotal requirements for interfacing were:

1. A transmitter (driver) had to deliver a mimum of 5V (either polarity) into IIRC a 1K load.
2. A transmitter had to withstand an indefinite short to ground or any other signal line (i.e. includes driver-driver contention). This is the one that made trial-and-error interfacing non-destructive in compliant hardware.
3. A receiver had to respond correctly to +/-3V applied.
4. A receiver had to withstand +/-25V.

Armed with those, one went forward into the jungle. Note that the spec doesn't state where in that +3v-3v zone the receiver threshold should lie.

Compliant real world receivers invariably used a positive threshold, responding to 0v as low and +3v as high. As a result, TTL swings would ALWAYS drive a real world 232 receiver.

The spec doesn't guarantee it. That is true. But the real world devices do.

I'm open to someone showing me documentation of a receiver IC with a negative threshold. But as an example of a receiver, have a look at the input circuit of the ubiquitous 1489:

those

It couldn't possibly have been that weak TTL RS-232.

those

However I have seen a 60V difference between two GenRad mid-rise buildings in Phoenix. I opted for transformers and Manchester.

...Jim Thompson

In article , Spehro Pefhany wrote: [....]

IME: this does not apply to laptops. Many of them have a 1/0 threshold that is just about exactly zero.

If this is something you just want to use for debuigging, you can solder a battery in the able to subtract about 3V from the 0-12V swing. The drain on the battery is very small so it lasts quite a while.

You can also build a full RS-232 swing circuit into some external box that is part of the test set for the product. You put a connector that isn't a DB-9 on the product. This way no-one[1] can plug a RS-232 straight in.

In article , Richard Henry wrote: [....]

LT1081 and MAX-232, I believe both put the boundary between 0.5V and 1V.

into IIRC a 1K load.

other

That is materially the same as ANSI/EIA-232-D (1986) and -E (1991), though some weasle-words are different.

This thread goes on and on to nowhere in particular.

But in the end, If a manufacturer builds equipment requiring + and - 3V minimum in their EIA232 interface (think Garmin for example), and some "think different" group of Bozos (think Apple) designs a MODIFIED EIA232 interface using only the negative voltage, and the equipments won't talk to each other, who is wrong? All at once now..... say "APPLE."

When a manufacturer builds equipment with a non-standard interface they should be shot with cream pies at 3-feet.

But in the privacy of our homes, shops and labs you can mix, match and modify to make equipment do all sorts of unnatural things, or to do the natural things companies like Apple screwed-up.

Don

DE-9.

those

No, from memory it was either +/-12 or +/-15v. The driver/receivers were strung between the op amp rails of the then analog computer.