Close
0%
0%

'scoping FAILs

I'm no newb when it comes to 'scoping... but sometimes I just FAIL at it.

Similar projects worth following
NEW FAIL: Single-Trace Additive(subtractive) viewing of a voltage-difference.

OLD FAIL: X-Y 'Scoping - digital signals
Experiments and learning to interpret x-y with my 'scope

The x-y feature on my 'scope has gone mostly-unused... basically just switched-to by accident, or for the random moment of curiosity.

Don't get me wrong, I get the *general* idea, but--since I mostly work with digital signals--I haven't actually used it much as a diagnosis-tool.

Today was one of those "random moments of curiosity" which led to some sudden realizations of the *meaning* hidden in this plot.

Read the details, and follow along, and maybe you'll learn something too (or correct me!).

This experiment/learning-experience began as my not-so-uncommon accidental-switch to "x-y mode" while trying to zoom in on a brief signal beyond my 'scope's time-scale zoom-ability.

(Obviously, I need to get out the ol' rubbing-alcohol and cotton-swabs!)

I needed to determine the time-delay between the rising/falling-edges of two signals.

Even at maximum-zoom (0.2us/div), these signals appear to be nearly-simultaneous.

I had imagined they should be; these two outputs are two bits on the same port, toggled simultaneously by the same port-write instruction in my microcontroller (an ATmega8515).

e.g.:

while(1) 
{ 
   PORTE = 0x03; 
   PORTE = 0x00;
}   
The two channels on the oscilloscope are connected to port output pins PE0 and PE1 (bits 0 and 1 on port 'E').

After "over-zooming" into X-Y Mode, I saw the following:

And immediately noticed that *if* the two signals were, in fact, simultaneous, then I should've gotten the two bright dots at the lower-left and upper-right corners, and a straight-ish line between these two points; the graph of y=x (where x is from 0V to 3.6V)... (or maybe a skinny slanted diamond/trapezoid?)

Instead, we're seeing nearly a square.

(NOTE: It just occurred to me that I have two different brands/quality of 'scope probes, which *may* have some bearing on this? The remainder of my discussion will most-likely disregard this new tidbit... until later?)

I also noticed some other characteristics that seem to stand-out:

  • [1] The very bright dots at the lower-left and upper-right corners
  • [2] The "Square" connecting these two dots
  • [3] The "Horizontal" lines appear to be *brighter* than the verticals
  • [4] The other corners seem to be significantly "brighter" than the lines, almost a *dot*
  • ([5] I didn't notice that the "horizontals" were actually as slanted as they are... this, too, will likely be disregarded early in this discussion...)

It got me to thinking what might *cause* these sorts of effects...?


Before reading-on, know that:
I'm not claiming that my hypotheses--nor even my observations--are *correct*.

As I said before, I've barely ever used the X-Y mode for anything other than some entry-level analog "labs" nearly a decade ago... This is *hypothesizing* and *experimenting*, and some of my observations seem to be corroborated by some loose-interpretation of the data-sheets and maybe a little bit of my experiences with digital-signals/circuitry, and also some (now) glaring-oversights.

For the purposes of this discussion, (at least, early-on) I'm going to assume the signal was *square* and that the 'scope-probes were identical. Thus, meeting the first *four* characteristics, listed above (not the fifth).



Wherein I do some hypothesizing!


The first characteristic is the bright dots at the lower-left and upper-right corners.

These are where both signals are Low (lower-left) and High (upper-right) at the same time.

If this were an ideal digital circuit, with no "rise-time" and no "fall-time", the 'scope would show nothing but these two dots, as all time would be spent either high or low.

//Just Dots:
//
//
//            B
//           *
//           
//           
//           
//   *
//  A      
// 
//            .       .   
//     H      .¯¯¯¯¯¯¯.
// x:  L______.       ._____
//            .       .
//            .       .
//     H      .¯¯¯¯¯¯¯.
// y:  L______.       ._____
//            .       .
//        A   |   B   |  a...

Since digital circuitry is not magical, there must be *some* time between these transitions. Thus, in a slightly-less ideal circuit, I'd expect to see a line connecting these two points. This line might be quite dim, possibly even invisible on a 'scope, if the rise-time and fall-times are very quick in comparison to the time actually spent in the high or low state.

The above idealization assumes that the rise-times of the two outputs are identical, as well as the fall-times (the line: y(t) = x(t)). And, considering that they are from the same port on the same piece of silicon, it seems a reasonable assumption.

//Line:  
//(Rise-times needn't be equal to fall-times,
// but rise-times must be equal
// and fall-times must be...
Read more »

  • FAIL: Voltage Difference Measurement

    esot.eric10/23/2015 at 08:37 0 comments

    UPDATE: FAILED SIMPLIFIED SETUP. REVISED.

    Unfailed: Understanding/explanation of original fail, at the bottom.

    Simplified Setup:

       5V
       ^
       |
       |    
      _|_   +
     /   \ 
    | Vs  | V_Source (to Measure)
     \___/
       |    -
       |
       |    +
       \
    R2 /   V_R2 
       \
       |    -
       v
      GND
    We want to measure V_Source, with an oscilloscope. So, attach Channel 1 to V_Source+, attach Channel 2 to V_Source-
      Vin
       ^
       |
    V1 *---------------> 'Scope Ch1
      _|_   +
     /   \ 
    | Vs  | V_Source (to Measure)
     \___/
       |    -
    V2 *---------------> 'Scope Ch2
       |
       |    
       \
    R2 /    
       \
       |    
       v
      GND
    Set your 'scope to "Add[itive]" Mode. Set Channel 2 to "Inverted."

    Then you get a single trace on the scope which shows:

    V_Trace = V1 - V2 = V_Source

    Excellent!

    ------------------------------------------------------

    Set both 'scope channels to 0.5Volts/Division.

    V_Trace appears to be AC: 0.1V

    Trace (V/Div = 0.5V):
    ---|------------------------|---- 1.0V
       |                        |
    ---|------------------------|---- 0.5V
       | _   _   _   _   _   _  |
    ---|/ \_/ \_/ \_/ \_/ \_/ \_|---- 0V
       |                        |
    ---|------------------------|---- -0.5V
       |                        |
    ---|------------------------|---- -1.0V

    EXCELLENT!

    ------------------------------------------------------

    Adjust 'scope channels to 50mV/Division.

    (Leave everything else the same)

    V_Trace appears to be AC: 0.01V

    Trace (V/Div = 0.05V):
    ---|------------------------|---- 0.10V
       |                        |
    ---|------------------------|---- 0.05V
       | _   _   _   _   _   _  |
    ---|/ \_/ \_/ \_/ \_/ \_/ \_|---- 0V
       |                        |
    ---|------------------------|---- -0.05V
       |                        |
    ---|------------------------|---- -0.10V

    WTF????????

    WAS EXPECTING:

    Trace (V/Div = 0.05V): 
    (Image zoomed in Time/Div, 
     but NOT V/Div, 
     for ASCII Art)
    ---|----__----------------_|---- 0.10V
       |   /  \              / |
    ---|--/----\------------/--|---- 0.05V
       | /      \          /   |
    ---|/--------\--------/----|---- 0V
       |          \      /     |
    ---|-----------\----/------|---- -0.05V
       |            \__/       |
    ---|-----------------------|---- -0.10V

    Again... V_Source was already measured to be 0.1V in the 500mV/Div setting, but now appears to be 0.01V in the 50mV/Div setting. W T F.

    That was *measured on the same 'scope* just seconds prior...

    So why are we seeing V_Source=0.01V when we saw 0.1V before?!

    -----------------------------------------

    Think about it, verify your connections, check for broken solder-joints, swear at the universe, etc. for an hour like I did... (Remember, I simplified this experience for yah; my circuit's a bit more complex... and I didn't just switch the 'scope's knobs, I also switched to a duplicate, but not identical, circuit inbetween steps... which resulted in checking the new circuit is right *over and over*).

    Finally:

    Adjust those 'scope signals to make sure you saw what you saw at 0.5V/Div...

    Whoa, wait... it is the same as it was last time! Yep, we are in fact seeing the same signal appear as 0.01V when viewed at 50mV/Div, and 0.1V when viewed at 0.5V/Div... Something's wrong with the 'scope? I don't use 50mV/Div very often...

    ----------------------------------------

    Didja figure it out yet....?

    -----------------------------------------

    Spoiler Alert!

    -----------------------------------------

    Here's a hint...

    Vin = DC: 5V

    The 'scope's set to 50mV/division...

    The scope *shows* 10 divisions, that's 500mV total shown on the screen...

    The two channels are set in Subtractive-mode...

    So, we have one 'scope channel at something like 5VDC + 0.05VAC... and the other at something like 5V - 0.05VAC.

    The 5V offset-voltage is feeding into the two probes...

    But is unseen in the trace due to being in subtractive mode.

    The 50mV/Div setting means that we're amplifying that 5V offset, on each probe, and likely hitting the amplifiers' rails. Duh

View project log

Enjoy this project?

Share

Discussions

DeepSOIC wrote 01/05/2016 at 17:18 point

I think this might be not because of the probes, but because of scope's X and Y amplifiers being different. AFAIK, in Y channel, sometimes a delay line is added, to be able to see the edge one is triggering off. If there is a same delay in X amplifier, it shouldn't be a problem.

  Are you sure? yes | no

esot.eric wrote 01/08/2016 at 02:22 point

Interesting, I was unaware! I'll have to check the ol' manual and see what it says. Seeing the trigger-edge would be quite handy in other situations... if it's not already on one axis of my 'scope, I might see what can be done with the delay-lines in my collection (maybe from an old VCR?) and the external-trigger input. Thanks for that idea! (LOL "vcr delay line" first result: http://hackaday.com/2014/01/07/acoustic-delay-line-memory/) Always wondered how to make use of those weird things.

  Are you sure? yes | no

Jacob Christ wrote 02/28/2015 at 15:50 point

Yes, my god man, clean your scope.

  Are you sure? yes | no

esot.eric wrote 04/01/2015 at 09:43 point

Good laughs all'round now that it's the banner for the "Lists" -> "Oscilloscope Projects". Good thing that particular image is blurry, otherwise every 'scope-project-list viewer would be greeted by little bits of last month's dinner ;)

  Are you sure? yes | no

Dan K wrote 02/17/2015 at 04:09 point

WARNING!

Repeating this experiment as before may very well earn you a spot on the 'Fail of the week' column! A scope with a CRT display should only be operated in X-Y mode if the inputs are constantly changing or the display brightness is greatly reduced. Those bright spots at the two corners of the square corresponding to both inputs being the same value? As you concluded the CRT's electron beam stays fixed at those two points most of the time. This may cause the phosphor to burn-in very quickly, leaving you with two permanent dots on the screen! I recommend that you reduce the trace intensity for subsequent experiments.

  Are you sure? yes | no

esot.eric wrote 02/17/2015 at 06:17 point

Good warning. Indeed.

That level of brightness was only to capture the dimmer lines on camera. It hurt the rods and cones in my eyes, so who knows what it's doing to the phosphor ;)

I think I'll throw your warning in the description, if you're cool with that.

I figured "Fail of the week" was destined *before* you pointed this out ;)

  Are you sure? yes | no

Dan K wrote 02/19/2015 at 02:49 point

Sure, paraphrase it as you wish.

I actually referenced "Fail of the Week" because you did. There's nothing like telling yourself "that was dumb of me to do," and then committing another error which, in addition to further embarrassment, causes physical damage to hardware :P

  Are you sure? yes | no

DainBramage wrote 02/19/2015 at 03:02 point

Been there, done that, felt the instant regret.

  Are you sure? yes | no

DainBramage wrote 02/16/2015 at 17:10 point

I didn't know that Goldstar made scopes!

I'm far from an expert on these things, but as I was reading I found myself wondering if you had adjusted the compensation on the scope probes.

  Are you sure? yes | no

esot.eric wrote 02/16/2015 at 23:54 point

Only after I wrote this ;) 

When I started turning that pot/cap it occurred to me that I've FOTW'ed a few times due to forgetting about the probes' characteristics.

That extra foot of length on one might have something to do with it, as well... 1ns/6in is the rule of thumb? So there very-well could be 2ns propagation-delay difference in the probes, themselves. Much too fast to see in normal scope traces, on this machine, but maybe those op-amps are fast enough to show it in X-Y mode.

That adjustment definitely affected the plot... Good-Eye.

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates