>> It is true because the receiver chips are made that way. I do not make
>> receiver chips, only use them. You will see in the spec sheets that the
>> hysteresis can be (and is) deleted or reduced. I think that they do
>> this to prevent edge ringing and cross-coupling from unmatched cheap
>> serial lines at high speed to cause glitches.
>
> Can you give me an example of such a device?

MC1489 has an explicit 'response control node' available for each
comparator. With this the Schmitt effect can be reduced to zero if
necessary. The same pin controls the threshold using an external resistor
and can be used as an input.

Normally these features are left unused, which is not so good. The
difference between the schmitt levels should be chosen to be 33% of the
input signal for maximum suppression (Schmitt thersholds + and - 5V for
+/-15V sending RS232). The 1489 does not meet this requirement w/o
external parts but,

The level at the end of a long RS232 line can be +/- 5V due to capacitive
and resistive load, so 33% of it it is close to 1.6V which is within the
Schmitt setting capability of the mc1489.

The glitch part is from the low slew rate of rs232 signals (30V/us) which
means that the device spends a long time in the 'forbidden area'). So if
there is any small noise of about the size of the schmitt levels on the
line during this time it will cause a glitch at the output. The mc1489
datasheet suggests using capacitors to slow down the response. But this
would make the part unusable for low level (+/-5V), high speed signals.

Other parts that are usable as RS232 receivers are RS485 receivers
used single-ended fyi.

So I think that the saying 'practice is more complicated than theory'
applies fully and that the the facto standard choice of about 2.5V Vih and
0.8V Vil for RS232 receivers seems to be a good compromise, and especially
that ymmv.

Peter

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body