How drive LEDs for RS232 input 3v to 30v + and -

Do you need accurate voltage detection, or are you just trying to determine what is on the line? [High, low, pulled to ground via input resistor]

Maxim's first generation auto-shutdown just detected if input was within a VTH of ground to classify the signal as ground.

If the situation has a line at hard ground and lines at ground due to the input resistors on the 232 device, it will take some sophistication to find the real (hard) ground.

try this:

+--->|----(in1)---->|----+ | | +--->|----(in2)---->|----+ | | +--->|----(in3)---->|----+ | | +--->|----(in4)---->|----+ | | +--->|----(in5)---->|----+ | | +--->|----(in6)---->|----+ | | +--->|----(in7)---->|----+ | | +--->|----(in8)---->|----+ | | +----[10K]--+--[10K]-----+ | +-- ref diodes are 1N914 or 1N4148 (or similar in an 8-pack, etc)

that'll be less succeptible to signal changes, as long as one signal is low ans one is high it'll give an output that's close to ground (as long as it's not passing much current).

you'll need battery power or a well-isolated supply.

some op-amps come 4 per chip...

--

Bye.
   Jasen

It sounds perfectly practicable to me. Practical is another question. There is nothing impractical about using eight op amps - you can get quad packages.

Farnell lists more than a page of bipolar input quads, half a page of FET-input quads and nearly a page of CNMOS-input quads. Some of them are quite cheap.

--
Bill Sloman, Nijmegen

No, the voltage doesn't need to be too accurate, though +/-3V would be ideal. Yes, one line coming in from the rs232 port would be expected to be an a signal ground, though it would not be obvious before testing which one. That is one point of the test. The signals will be presented on a (possibly incorrectly wired) RJ45 and this is to be a breakout box for the eight lines.

snip

That's the kind of thing I was thinking of but using, say, 10k resistors fed to a common point. I don't think I can see what the diodes would do. Say, for example, only in1, in2 and in3 were connected and that they were at +12v, 0v, and -12v. Would the +12v raise the RHS to +11.3, and the -12v drop the LHS to -11.3?

Great! I'm just looking at op-amps and they look to be really useful for this application. It looks like any one op-amp will drive its output positive or negative depending on the greater of the non-inverting and inverting inputs. I should they be able to use the single output to drive two LEDs, one for +ve and one for -ve. The problem is, I need a gap of about 6v in between when both LEDs should be off. I'll read up some more.

If you are looking at actual RS232 circuits, why not use an out- of-the-(made in China)-box tester?

6 or 8 bicolor leds to show you the status of each line. (They just use a resistor and a bicolor led per line.)

Ready made, must set you back at least 2 pound :-)

--
Kind regards,
Gerard Bok

snip

I have a couple of these. The problem is they require you to know which line is ground. This is OK for a 9- or 25-pin standard serial cable but the 8-pin RJ45-type serial connectors seem to vary. Many are Yost-standard pinouts with the two central pins as ground, but not all. Hence this idea.

Wouldn't it be possible to use an external ground and just 'test' all 8 leads for possible RS232 signal ? Those pins that don't light a LED could well be ground :-)

--
Kind regards,
Gerard Bok

snip

You mean test them in isolation? I don't think that would work. Each of the signals is only what it is relative to its own ground. Just as the "voltage" of a line is only relative to another line and birds can perch on power lines etc......

The idea is to take all eight lines to a common point with high value resistors. Then measure the difference between that point and each line. Because the potential difference between the reference point and each line will be high impedance I will need an amplifier to drive LEDs.

. . . . . .

yeah,

now suppose in4-8 were all +12V

with the diodes ref (the output) would still be about 0V

but with resistors you'd get the average of all the inputs

1x0 + 1x-12 + 6x12 ------------------ = 60/8 = 7.5V 8 the diodes and two resistors above get you the voltage half-way between the highest and lowest voltages.

Bye. Jasen

Well, I don't know your actual situation. If your source is 'true RS232' one of the wires should be pretty close to ground level. Your problem would be, to find out which one is.

My suggestion is, to connect all 8 wires to a detector that is connected to 'real ground'. That may not be a reliable situation for datacommunications but it may well be an easy way to detect signal levels on each wire. And to detect, which of them is RS232-ground.

Anyway, if you have the gadgets already available, how much time would it take to probe the idea? What would it save you if it just works :-) ?

It is not uncommon for RS232 ground to shift several volts from 'real ground'. But it is not common to float. If you are in doubt, fuse the wires.

But you are quite right if you already know that your RS232 source will be floating.

--
Kind regards,
Gerard Bok

jasen wrote: snip

Cool!