Trying to probe RS232 on an analog oscilloscope - need some help. Thx

[Electro Fan]

Question:  Anyone have any advice on how to get a string of hyperterminal characters to trigger on an analog oscilloscope?

Background:

I can connect two Microsoft PCs to one another.  One PC is connected via a RS232 serial port and the other PC uses a USB port.  The two are connected via a startech.com cable that has a RS232 Null Modem connector on one end and a FTDI USB connector on the other end.  The StarTech cable came with a USB driver that works fine on the USB PC.  (There are some other threads here about people trying to get similar connections working - this cable was easy to setup and was delivered very rapidly via Amazon in case anyone is interested.)

I can definitely send characters a key stroke at a time, back and forth in both directions - or I can send an entire text file in either direction - so I know the overall send/receive process is working between the two PCs.

Next I have been able to remove the RS232 connector from the PC with RS232 port and plug it back into the RS232 port and the data still flows in both directions when I resume typing or sending a text file, so I know just removing the cable doesn't send everything into lala-land.  I can also continue to type on the USB PC and see the characters echo on the USB PC screen even when the RS232 connector has been removed from it's PC.

Next I probe the ground (Pin 5) on the StarTech female RS232 connector and the RxD (Pin 3) on the StarTech female RS232 connector with my Tektronix analog scope.  Then I try sending data from the PC (which has the USB port) and I'm just trying to see the voltage go high and low on the scope.  I don't seem to get anything resembling the +/-13 or so volts that I would expect to see from the RS232 connector (or even anything like the 0/+5 volts or so I would expect from a USB connector) - but maybe I have those specs wrong.  I've tried various settings on the scope from millivolts to whole volts and various settings around milliseconds (including slower and faster).  I've tried all three trigger settings NORM, AUTO, and Auto Level.

It's possible that I don't have the right pins probed (but I'm pretty sure I have them correct and I've tried other combinations), or it's possible that maybe the data really isn't flowing despite my step by step process, or it's possible that I don't have the scope set right for triggering (my second most likely guess), or it's possible that the world is telling me it's time to buy a digital scope so I can more easily trigger on a single shot event.

For what it's worth, I found this:


Part way through video the demo shows the pins being probed (it looks like the same pins I'm using but my connector happens to be female).

Any suggestions would be appreciated.  Thanks! 

[c4757p]

Next I probe the ground (Pin 5 on the StarTech female RS232 connector) and RxD (Pin 3 on the StarTech female RS232 connector) with my Tektronix analog scope.  Then I try sending data from the PC (which has the USB port) and I'm just trying to see the voltage go high and low on the scope.  I don't seem to get anything resembling the +/-13 or so volts that I would expect to see on a RS232 connector (or even anything like the 0/+5 volts or so I would expect on a USB connector) - but maybe I have those specs wrong.  I've tried various settings on the scope from millivolts to whole volts and various settings around milliseconds (including slower and faster).  I've tried all three trigger settings NORM, AUTO, and Auto Level.

What do you see? Just zero? (And no, you won't see 0/+5V on USB, it's lower-voltage differential - and RS232 doesn't require the full voltage either, you're more likely to see about +- 8V)

I wouldn't use auto level, because it will probably try to adjust the trigger level when the data is transmitted, and you won't see anything while it adjusts. Set the level manually to 0V.

[Electro Fan]

What do you see? Just zero? (And no, you won't see 0/+5V on USB, it's lower-voltage differential - and RS232 doesn't require the full voltage either, you're more likely to see about +- 8V)

I wouldn't use auto level, because it will probably try to adjust the trigger level when the data is transmitted, and you won't see anything while it adjusts. Set the level manually to 0V.

Thanks.  I see signals that look like noise (fuzzy and diagonal) and nothing over two volts; mostly either 1-2 volts or in the millivolts.  Nothing that resembles square waves. It might be that I don't have a really good understanding of setting the triggering.  For what it's worth I have a Tektronix 2247A.   (The photo isn't my unit - just a photo that sort of shows the trigger layout toward the right side of the scope.)  I'm guessing "manually" means use the "NORM" trigger setting?

[c4757p]

NORM or AUTO - NORM will only display if there is a signal to trigger on; AUTO will trigger by itself if the signal is absent, so there's always something on the screen.

Hmm... I suspect the probing setup. How exactly are you connecting the probes? You have connected ground, right?

Once you have a display, you can play with the holdoff control to see if you can get the whole signal to display. Failing that, you can probably hack together a simple 555 circuit to isolate the start bit, then trigger on that.

[AG6QR]

To get started, set the time base very slow, try something in the neighborhood of a second per division.  This won't let you pick out the ones and zeroes of individual bytes, but it will let you see the range of voltages present.   Set the trigger to "free run", where it just continuously sweeps without waiting for a trigger event.  (Auto could work, but it will wait a while before sweeping when no trigger is present).  This way you can concentrate on figuring out the vertical range without worrying about triggering.

Set the vertical scale at something like 2V/div.  Ground the probe and center the trace.  You should see it sweep slowly from left to right across the time axis line of the screen, repeating the scan continuously.

Now probe the circuit, and watch the trace both while you're sending characters and while the line is silent.  You should get varying voltages.  Adjust your vertical scale and position as appropriate so that the trace mostly fills the available vertical space.

Now speed up the sweep to something like 0.1 s/div and adjust the triggering so that it triggers on a rising or falling slope in the range of voltages you're seeing.  You should be able to get it so that it triggers when you send a character.  Finally, speed up the sweep so that you can see the individual bits.

[Electro Fan]

NORM or AUTO - NORM will only display if there is a signal to trigger on; AUTO will trigger by itself if the signal is absent, so there's always something on the screen.

Hmm... I suspect the probing setup. How exactly are you connecting the probes? You have connected ground, right?

Once you have a display, you can play with the holdoff control to see if you can get the whole signal to display. Failing that, you can probably hack together a simple 555 circuit to isolate the start bit, then trigger on that.

I have the male end of the Tek probe inserted into the female RxD and the probe ground clip attached to the female GND with a short piece of wire (into the female GND).  Might be something flakey about the connection but I've tried pretty hard to make good connections.

I've also payed with the holdoff control - I'm not 100% sure but I don't think this is the issue.

I happen to have recently done a simple project with a 555 but I'm inclined to think my skills with the 555 would make this approach to solving the RS232 issue more complicated rather than less complicated - but I appreciate the creativity 

At about 1:53 in the video the demo shows how the AC and DC signals can viewed - what the demo shows is fairly close to what I've been seeing on my scope.  I've been just using DC coupling - any chance I should have been using AC?  (Duh??)  Thx

- Are we pretty sure the characters are viewable as clean rising and falling square waves on an analog scope?  I thought the trigger just slowed down a sweep of a continuous signal to make it look like the signal is stationary?  On my analog scope I don't see how I can actually freeze a one shot signal, but maybe I'm not understanding something (VERY possible). 

[Electro Fan]

To get started, set the time base very slow, try something in the neighborhood of a second per division.  This won't let you pick out the ones and zeroes of individual bytes, but it will let you see the range of voltages present.   Set the trigger to "free run", where it just continuously sweeps without waiting for a trigger event.  (Auto could work, but it will wait a while before sweeping when no trigger is present).  This way you can concentrate on figuring out the vertical range without worrying about triggering.

Set the vertical scale at something like 2V/div.  Ground the probe and center the trace.  You should see it sweep slowly from left to right across the time axis line of the screen, repeating the scan continuously.

Now probe the circuit, and watch the trace both while you're sending characters and while the line is silent.  You should get varying voltages.  Adjust your vertical scale and position as appropriate so that the trace mostly fills the available vertical space.

Now speed up the sweep to something like 0.1 s/div and adjust the triggering so that it triggers on a rising or falling slope in the range of voltages you're seeing.  You should be able to get it so that it triggers when you send a character.  Finally, speed up the sweep so that you can see the individual bits.

Thanks - this is very close to what I have been doing (I think ).

Just to confirm, when I set the sweep to very slow I can get it so slow that I see the dot go from left to right very slowly, and then I have gradually increased the speed.  But I will go back and retrace your suggested process and see if I can get it to show the data signals.  Thx again

[c4757p]

It won't freeze a one-shot signal, but keep in mind that a digital signal is repetitive. It won't see a transmission as one event, it will trigger on each edge. If you want to see the whole transmission, you could set up the computer to transmit repeatedly, as well.

I don't see a problem with the mechanical way you are probing. But... have you checked the pinout?



RXD is pin 2. What end are you probing at? The signal should be going into the receiving computer on pin 2, or coming out the transmitting computer on pin 3.

[Electro Fan]

It won't freeze a one-shot signal, but keep in mind that a digital signal is repetitive. It won't see a transmission as one event, it will trigger on each edge. If you want to see the whole transmission, you could set up the computer to transmit repeatedly, as well.

I don't see a problem with the mechanical way you are probing. But... have you checked the pinout?



RXD is pin 2. What end are you probing at? The signal should be going into the receiving computer on pin 2, or coming out the transmitting computer on pin 3.

c4757p - Thanks, very nice of you to post that pinout diagram.

It is similar to the diagram supplied in the StarTech manual but a bit different.

In the StarTech manual it also shows Pin 5 as the Ground - so we are good on that.

The StarTech manual shows Pin 3 as RxD - which I would think is the pin on which the RS232 connector is receiving data.   So, Pin 3 (the middle pin on the long side) is what I have been primarily trying to get working with the scope probe.  But just in case of some confusion, I have also tried Pin 2 (which is labeled by StarTech as TxD).

Your diagram might be the official/conventional pinout; maybe the StarTech is their proprietary way of doing things - possibly because of how the USB end works but I don't see why that would be the case.  For now, I'm trying both Pin 3 and Pin 2 in concert with the GND on Pin 5.

I'll keep "plugging". 

- Just to confirm, the scope should be on DC coupling, right?

[alm]

A continuous stream of capital Us (eg. UUUUUUUUUUUUUUU...) is an excellent test signal. In 8N1 (8 bits, no parity, 1 stop bit) format this produces a square wave.

[c4757p]

Yep, DC coupling.

A continuous stream of capital Us (eg. UUUUUUUUUUUUUUU...) is an excellent test signal. In 8N1 (8 bits, no parity, 1 stop bit) format this produces a square wave.

How the hell did I never think of that?

[Electro Fan]

A continuous stream of capital Us (eg. UUUUUUUUUUUUUUU...) is an excellent test signal. In 8N1 (8 bits, no parity, 1 stop bit) format this produces a square wave.

Cool,  Thanks, alm.  I will try Capital Us next (already have 8 bits, no parity, 1 stop bit).

[AG6QR]

- Are we pretty sure the characters are viewable as clean rising and falling square waves on an analog scope?  I thought the trigger just slowed down a sweep of a continuous signal to make it look like the signal is stationary?  On my analog scope I don't see how I can actually freeze a one shot signal, but maybe I'm not understanding something (VERY possible).

Even if you're not getting the triggering right, as long as the sweep is running, the worst that will happen is that you'll get a confused mass of multiple digital signals on top of each other.  You should still be able to see that the signal spends most of its time at the "high" and "low" voltage range of the RS-232 signal lines (supposed to be +-12V or so, might be 0V to 5V or some other range).

The trigger doesn't "slow down", in the sense that, once the sweep starts, it progresses across the screen at the selected speed no matter what.  But if you meant "slow down" to mean that the trigger causes each sweep to wait until some starting condition is met before it will start the sweep, you've got it.

An analog scope isn't the best tool for viewing a one shot digital waveform, especially not if a digital scope is available.  But there are ways to do it on an analog scope anyway.

One is to turn the "one shot" digital waveform into a repetitive waveform.  Autorepeating a "U" as alm suggests is great.  Setting the trigger holdoff will be important to trigger on only the first bit of the "U".

Another way is to use the screen persistance to see a true single shot.  Darken your room.  Turn the trace brightness up.  Send one character.

Sometimes people would use photographic means instead of just relying on the phosphor persistance.  If you can set a camera on a tripod to watch the screen, in a darkened room, with a long shutter speed, you might be able to capture a single trace.  Scope cameras have been designed to make the process easier.  Many of them would use polaroid film.  They're pretty much obsolete these days.

Even the lowest of today's hobby grade digital scopes makes observing one-shot digital signals so much easier than it is on a top notch analog scope.

[Electro Fan]

Thanks everyone.

Here's the report.  I've got the good news and the bad news.

The good news is that I have found the cap lock key and the U key and without a doubt I can send a capital U or as many as I want in either direction from computer to computer

The bad news is that not much else seems to be working 

I have tried setting the trigger around 1.5 volts (and higher and lower, even at -1.5 volts in case something is inverted) but nothing seems to trigger in any mode (including Norm or Auto).

I'm sure it's user error.  I haven't tried the dark room persistence idea yet but I'll revisit everything as best I can.

I'm pretty sure this is a cosmic message that it's getting close to time for a DSO 

[dfmischler]

When no data is flowing and everything is cabled up you should see a negative voltage on Rx and Tx.  Do you see that?  What voltage do you see there?

[Electro Fan]

When no data is flowing and everything is cabled up you should see a negative voltage on Rx and Tx.  Do you see that?  What voltage do you see there?

Hey, hey, some progress!

When I leave out the StarTech cable and I connect the scope (or a DMM) directly to the male pins on the RS232 connector (on the other computer) I get 12 volts if I probe the Ground pin with the plus (red) on the DMM and the I probe the middle (3 Pin what StarTech calls RxD) with the negative (black) on the DMM.

(ie, you were right - I was seeing negative voltage which is what caused me to put the red/positive DMM probe on the GND and the black/negative DMM probe on the RxD); that was my way of getting postive voltage - which I figured would help me set the trigger on the scope above the 0/Gnd line, but I realize I can trigger south of the line too.)  The good news is that I have something to measure 

I'm going to try redoing the sweeping and triggering by probing the male pins on the RS232 port of the other computer and quit trying to probe the female connector on the StarTech cable; I'm getting the StarTech cable out of the equation.  It works fine for data transfer but for now I'm confident the computer with the RS232 is working so I'll focus on that as the source of the data.  I'll try both Pin 3 and Pin 2 to make sure I'm not confusing the send and receive.

Thanks

[alm]

You can't reverse the scope leads (attach ground lead to pin 2/3 and probe tip to the ground pin / shell) like you did with the DMM. The scope's ground lead is grounded, and so is the computer (probably), so you would be shorting out the output. Not likely to be harmful, but it won't allow you to measure anything.

[Electro Fan]

Back at it again.

I took some color female to male breadboard jumper wires and inserted all the female jumper ends over the male pinouts on the RS232 port.  Then I inserted all the male jumper ends into the female RS232 cable connector and ran the cable back to the USB port on the other computer.  The text transfers nicely back and forth between the computers so I have confidence that my jumper wires attached to the 9 pins of the RS232 port on the computer are making good connections.

I can use the DMM to probe Pin 5 GND (with DMM negative) and Pin 3 RxD (with DMM positive) and I get -12 volts (actually -11.98).  When I connect the DMM probe to Pin 2 TxD (with DMM positive) I get +11.33 volts.

When I replace the DMM probes with the scope probe the voltage readings on the scope correspond to the DMM voltage readings but when I push the keys I can't detect any waveforms and no combination of triggers reveals anything. 

That's about it so far......

The silly thing is that this is sort of fun

Go figure 

[c4757p]

Is the scope interfering with the actual transmission of data, or do you get a proper transfer with the scope connected?

[alm]

Do you get a decent square wave if you use the same probe, channel and front panel setup to probe the probe calibration signal available on the front panel of the scope? This should be a 1 kHz square wave (a period should be one div wide on 1 ms/div) and is somewhat similar to an RS-232 signal. My guess that there is something wrong with the trigger settings, like triggering set to line, or HF reject, or something else. I think W2AEW has some useful videos available. This one might be relevant:

[Electro Fan]

Is the scope interfering with the actual transmission of data, or do you get a proper transfer with the scope connected?

That's a good question, but I don't know how to determine the answer.  It sure doesn't seem like anything other than static (unchanging) voltage is getting through to the scope.  Right now the scope is connected to the GND and TxD and it reads about 11.25 volts.  When I push a key on the computer I can't see any change in the voltage (numerically or with a waveform).

I'm wondering if despite what the demo showed in the video if I'm really sending any voltage change over just the two wires?  When I connect all nine wires/pins the text transfers but maybe something about my setup is not really signalling anything up and down?  It's kind of challenge because I can't figure out if it's the pin connection/assignment, or if it's something I'm doing that is not enabling the triggering to occur.  I need to find a way to isolate which of these is the issue.

Thanks

[c4757p]

OK. Don't worry about triggering. Set up your computer to emit a constant square wave (UUUUUUU........ as above). Then put the scope in auto trigger mode, with the trigger level anywhere, perhaps 1ms/div 5V/div, and probe around on the cable. In auto mode, it will still plot the signal even if it can't lock onto it, so you'll get a wide illuminated region. Get that working first. And do it spliced into the cable while the other computer is receiving, so you can verify that the transfer is taking place.

And like alm said, check the cal signal to make sure the scope and probe work.

[alm]

If you're unsure about pin assignment, then try connecting as few wires as possible. You only need RX, TX and ground if you're not using flow control (turn this off on both sides). You can even disconnect RX on one side, and that side is only able to send. If you're still receiving data on the other side, you can be sure that that one remaining wire is TX. RS-232 uses a voltage swing of at least +/- 3 V or so, so there will be voltage signals on that line.

Of course you would do this after ensuring you can display a known-good signal on the scope, see my earlier post for some hints. If you're unable to get the probe cal signal to display even after watching W2AEW's video, then a good picture of the scope's front panel showing all of your settings (i.e. your scope with the settings you're using) might be useful to troubleshoot any issue with scope settings.

[Electro Fan]

Do you get a decent square wave if you use the same probe, channel and front panel setup to probe the probe calibration signal available on the front panel of the scope? This should be a 1 kHz square wave (a period should be one div wide on 1 ms/div) and is somewhat similar to an RS-232 signal. My guess that there is something wrong with the trigger settings, like triggering set to line, or HF reject, or something else. I think W2AEW has some useful videos available. This one might be relevant:

Yes, I can get a very nice 1 kHz square wave using the front panel test feature and I can easily move the trigger into the path of the square wave to stop the wave.  Maybe this shows that my triggering system and my basic triggering skills might be ok in which case the issue starts to look like maybe although the static voltage is making it to the scope from the computer the key presses aren't toggling the voltage up and down?

If I send one character from the keyboard (like a capital U) that should theoretically send 9 bits (8 data bits and a stop bit) in the form of 9 1s and 0s, ie 9 ups and downs, right?  And given the voltage of RxD and TxD the voltage should toggle between roughly +13 volts and roughly -13 volts.  So, if I start out high with no key presses, at +13 volts, then maybe I should set the trigger between 0 and -13 volts, or should I set it on the positive side just slightly above 0 volts?

Thanks again

[c4757p]

Like I said, if you set the computer to transmit continuously, you don't have to predict where to set the trigger, because you'll have a constant square wave. Just set it, probe, and play with the controls.