modern RS232 ports

I thought it was "What's the frequency, Kenneth?".

essage

ors still

ata if it

I just find those usb to 232 converters a crap shoot. I had a great one (keyspan), but no win7 drivers. My Airlink 101 sort of works, but not with everything. The voltages are decent (+-8v), so I don't get the compatibility problem. For a desktop, I just buy a damn serial board and be done with it. I've got 8 ports on my linux box. But for a notebook, only some of my socketcom cards work in win7. I can't run DOS with the socketcom due to Dell not having a 16 bit mode on the cardbus. Socketcom stopped making serial cards, even at their obscene prices.

Basically making someone do 232 these days is a pain. I suppose if the customer like it but really, it is time to bury that 232 technology.

On a sunny day (Sun, 20 Mar 2011 20:34:36 -0700 (PDT)) it happened " snipped-for-privacy@sushi.com" wrote in :

I disagree. RS232 wil probably outlast USB3,4,5,6,7,8,...,99

Just because it is simple, can go long distances, can be used for other things, and almost all mobos have it, especially the passively cooled cheap small Atom boards. And it is very easy to implement on the embedded side. I like it. Just that Larkin does not know how to connect it, and puts 12V on its OUTPUTS is no valid argument against it.

Or even Request To Send.

boards.

is no valid argument against it.

What in the world are you talking about?

John

On a sunny day (Mon, 21 Mar 2011 08:16:19 -0700) it happened John Larkin wrote in :

boards.

is no valid argument against it.

One of your previous problems of blowing up a RS232 port. Amnesia?

things,

Atom boards.

is no valid argument against it.

I blew up a _receiver_ by applying +12, not a driver. RS232 requires

John

hings,

Atom boards.

PUTS is no valid argument against it.

A good 232 chip will take +/-12V on the output. Better yet, it should be able to start up with all the transmitters tied to a hard rail. [A potential hairy situation in products with charge pumps due to such chips being more prone to substrate current issues.] This is a standard apps test at any decent company. Nobody puts this in writing, but you risk a pissed off vendor if your chips blow up easily because nobody likes RMAs. Any good vendor would do such a test at the design in phase.

things,

Atom boards.

OUTPUTS is no valid argument against it.

_Including_ the drivers, FWIW.

Best regards, Spehro Pefhany

things,

Atom boards.

OUTPUTS is no valid argument against it.

That's quite odd. The receiver is resistive input, in a floating isolation island.

Which was no problem whatsoever in my original design... the lateral PNP structures could hold off at least 50V :-) ...Jim Thompson

On a sunny day (Mon, 21 Mar 2011 15:08:17 -0700 (PDT)) it happened " snipped-for-privacy@sushi.com" wrote in :

Lots of mobos have chipsets with build in RS232 running from 3.3V or so. If you put plus or minus 12V DC on a serial output of those, what do you think will happen? Any idea of the circuit that will allow that safely?

Supply rail, yes, but not an external voltage that is higher, AFAIK RS232 specifies an input voltage range, and an output short current protection, but not a protection against external high voltage sources with low impedance connected to an output. Or does it?

On a sunny day (Mon, 21 Mar 2011 14:50:51 -0700) it happened John Larkin wrote in :

things,

Atom boards.

OUTPUTS is no valid argument against it.

Inputs yes, are you absolutely sure you did not touch an output from the same package? That could destroy the chip and then input would no longer work either. I have never seen a RS232 input die from +12V ever.

On a sunny day (Mon, 21 Mar 2011 18:12:05 -0400) it happened Spehro Pefhany wrote in :

I think that is only valid for a current limited source, so basically if you accidently swap pins. But if you are sure, connect a lab supply with +25 V and 10 A current limit to the serial output of your PC, Let us know if it meets your 'spec'. Just curious.

Wrong, idiot.

Curious what ten amps across your chest would do.

No, they actually have to withstand an indefinite short to +/-25V.

"Accidentally" swapping pins 2/3 was the standard way of getting serial interfaces going, IIRC, and in that case, the driver is in the other device and out of your control.

There's not much difference anyhow, as the device you're connecting it to could (in theory anyway) supply stupidly high current and still conform. IIRC, something like 500mA IIRC. So the driver could have to dissipate 25W if shorted to the opposite voltage extreme. Obviously no IC drivers supply anything like this current, but if you're really trying to conform (I know, you're not!).

The fragility of those 3V RS-232 drivers is probably due to some kind of (too) easily triggered latchup condition.

things,

Atom boards.

OUTPUTS is no valid argument against it.

I blew up two HP ProLiant server serial ports in a row, applying +12 to receivers. They used a TI version of a MAXxxx chip, on the motherboard. I just swapped the boxes, hot-plug RAIDs, with other machines and got back in business.

Really, they could have added a resistor pack on the motherboard and protected all the lines, for 5 cents or something.

John

will happen?

protection,

connected to an output.

I've told you, several times, that it was an INPUT that was blown by

John

I know this is obvious, but were both machines properly grounded ?. If not, switcher leakage would most likely fry devices as you plug in the db9's.

Speaking from experience :-)...

Regards,

Chris