Help determining if oscilloscope is functioning properly - TDS5054B-NV-AV

[newtekuser]

The scope passes calibration and displays a square wave when hooking up the probe to the compensation signal on the front of the scope. I also probed a few crystals and it shows sine waves (i.e. 4.6Mhz) properly but for the life of mine I can't get it to show a square wave. Shouldn't I see square waves when probing address/data pins off ICs and LCD boards, even DC voltage?
Once I hook up the probe I can see the signal jumps from 0 to several mV (Volts if I probe a Vpp pin) but when I dial in the scale and resolution to zoom in on the signal the wave form is not square, more like sinusoidal with some jagged edges.

Is there some setting I'm missing? Bad acquisition board maybe? (I get the same result on all four channels)

Some context: I was troubleshooting issues with an old PC/XT keyboard (basically it's not detected by the computer) trying to determine if some ICs are dead and was only able to determine that one chip is alive. I only came to this conclusion because the clock was functioning at the correct frequency and reset was not asserted. However, the waves on other pins were not square as I was expecting (i.e.: data/address pins)

Thanks in advance!

[RoGeorge]

Fast digital signals are never sharply square.

A quartz oscillator (the crystal itself when measured) will produce a sinusoidal wave, unless its a can oscillator with square output on the upper right pin.  Also the bandwidth of the oscilloscope can make the signals look "rounded".  For example, a 100 MHz perfectly square wave measured with an 100 MHz oscilloscope will look like almost like a sinusoidal wave.  Even with a 1 GHz scope will still appear rounded.  Then there are impedance mismatches and reflections, ringing in the ground alligator clip+wire (use the tip spring of the probe), etc.  Another thing is that 3-state digital lines tend to look very noisy and dirty because of the fast switching between well defined 0/1 levels and 3-state.

I hope the probe are 10x (not 1x), and properly calibrated against the internal square wave signal provided by the oscilloscope.

It's hard to say without knowing the exact circuit you are measuring, the probes you are using and the signal you see on the screen.  Maybe attach a small photo (1-2 megapixel ~100kB, please do not attach huge/tens of megapixel and many MB pics).

Most probably your oscilloscope is working properly.

[newtekuser]

Thank you for your reply RoGeorge!

I am using a cheap Chinese (P6100) 100MHz passive probe at 10X and calibrated to the scope. Attached is a collage of screenshots:

- top left: signal from the 4mhz crystal
- top right: compensation signal / calibration at scope front panel
- bottom left: signal without probe connected
- bottom right: signal with probe connected to one of IC (PIC16F887) output pins that connects to a hitachi LCD (data pin)

Initially I was using a Tektronix P5050 500MHz probe which was acquired at an eBay auction but was not even able to see a square wave when hooking it up to the scope compensation (learned that these often go bad), so I went with the cheap $9 probe thinking it will be more than enough since at the moment I'm prototyping using the PIC16 family and won't need too much bandwidth.

Edit: my scope is a Tektronix TDS5054 500MHz model

[kcbrown]

Thank you for your reply RoGeorge!

I am using a cheap Chinese (P6100) 100MHz passive probe at 10X and calibrated to the scope. Attached is a collage of screenshots:

The issue is neither with your scope nor your probe.

The square wave from the calibration signal of the scope proves that the scope is perfectly capable of displaying a square wave.  The timebase used for that is 400 us, and the wave requires 2.5 divisions, which translates to 1 ms, to complete a cycle, making it a 1 kHz square wave.

The sine wave from the crystal is being shown with a 2 ms timebase, and a cycle takes 0.5 divisions, which translates to 1 ms, to complete a cycle, so it, too, is a 1 kHz signal.

No, what's being generated by the crystal is clearly a sine wave.


Note that for anything more than about 1 MHz, you should be using a 1:10 probe.  Most cheap probes have a switch that lets you select "1x" or "10x", the latter of which being what you want to use for anything of any reasonable frequency, and is also what you need to have selected when you're performing compensation adjustments to your probe.  You need to tell the scope that you're using a 1:10 (i.e. "10x") probe under those conditions so that it will display voltages properly.

[newtekuser]

Thank you for your reply RoGeorge!

I am using a cheap Chinese (P6100) 100MHz passive probe at 10X and calibrated to the scope. Attached is a collage of screenshots:

The issue is neither with your scope nor your probe.

The square wave from the calibration signal of the scope proves that the scope is perfectly capable of displaying a square wave.  The timebase used for that is 400 us, and the wave requires 2.5 divisions, which translates to 1 ms, to complete a cycle, making it a 1 kHz square wave.

The sine wave from the crystal is being shown with a 2 ms timebase, and a cycle takes 0.5 divisions, which translates to 1 ms, to complete a cycle, so it, too, is a 1 kHz signal.

No, what's being generated by the crystal is clearly a sine wave.


Note that for anything more than about 1 MHz, you should be using a 1:10 probe.  Most cheap probes have a switch that lets you select "1x" or "10x", the latter of which being what you want to use for anything of any reasonable frequency, and is also what you need to have selected when you're performing compensation adjustments to your probe.  You need to tell the scope that you're using a 1:10 (i.e. "10x") probe under those conditions so that it will display voltages properly.

Thank you for your response kcbrown!

I looked at scope settings and could not find how to set probe attn, only these settings  (unless they are burried deep in other menus)

See attached pic.

[tautech]

Staring right at you: Probe Atten ! Screaming press me press me.........

[RoGeorge]

From the photo it all looks OK to me.

It's expected for a Quartz crystal to show a sinusoidal oscillation (often with some distortions).  PIC16F887 oscillator does not produce a nice square wave.  There are other gates inside the PIC that will turn the sinusoidal oscillation into a nice square for the microcontroller to use internally, but on the 2 outside crystal pins it's not nice and square.  Also, it is not unusual for a crystal to stop oscillating at all when probed with an oscilloscope, because of the parasitic capacitance of the probe.

Search for "probing crystal oscillator" to see what waveform to expect, for example at minute 31:00 in this video https://youtu.be/I4bAfDu6F1k?t=1852
https://microcontrollerslab.com/oscillator-types-microcontrollers/

In your pic at bottom right it looks like there is no data, just some noise over a DC voltage.  All 4 pics looks OK.

Bottom line, the oscilloscope seems OK.

[newtekuser]

Thank you everyone! Yes, the attenuation button was hidden in plain sight (lol).
I did set it to 10 from 1.0 and now I can see the signal jumping correctly one division when the scale is set to 5v/div and probing Vpp pin on the IC.

The only thing that I'd like to know is why I'm only seeing voltage noise when probing data pins? I.e. I'm pressing a push button which makes the lcd display something while the proble is connected to data pin (I tried all pins and seeing the same) and only see noise similar to the bottom right screenshot I attached previously.

[tautech]

Thank you everyone! Yes, the attenuation button was hidden in plain sight (lol).
I did set it to 10 from 1.0 and now I can see the signal jumping correctly one division when the scale is set to 5v/div and probing Vpp pin on the IC.

The only thing that I'd like to know is why I'm only seeing voltage noise when probing data pins? I.e. I'm pressing a push button which makes the lcd display something while the proble is connected to data pin (I tried all pins and seeing the same) and only see noise similar to the bottom right screenshot I attached previously.

Your most valuable tool is the trigger, understand how it works, all the different types and how use them correctly.
First assign it to the channel in use and adjust it within the waveforms vertical.
Set trigger and channel inputs to DC coupling and positive going Edge for the vast majority is use.

Turn OFF shit you don't need and ON stuff that helps until you become more familiar with the scope.
Cursors OFF unless you need them, Measurements ON like frequency and Vp-p.
KISS until you have your head around scope use.

Remember, your very first sanity check is always the probe compensation output.
Have fun, we've all be where you are. 

[RoGeorge]

why I'm only seeing voltage noise when probing data pins

Probably it happens too fast to observe the data with the oscilloscope in auto sweeping mode.

Writing a text to the LCD can be very fast.  Once the text is sent, there is no other (continuous) signals sending until something else is written again to the LCD.  The LCD has it's own electronics that pulses the electrodes on the glass with the LCD characters, so it's not the PIC who keeps pulsing the LCD glass electrodes.

You need to set the oscilloscope to trigger on the signal you want to see, and to memorize it if your oscilloscope has a memory/record/single trace option (name of this feature might be called different for different oscilloscopes brands).

Often memorizing the trace happens (depending on the oscilloscope brand) simply changing the sweeping mode from "Auto" to "Normal".  You will also need to set the trigger on the signal you want to observe.

Find a user manual for your oscilloscope model and read about how to capture a transient, non repetitive signal.

Alternatively, if may try to change the PIC program in such a way that it will keep writing to the LCD without looking at the button press, so the writing signals to the LCD will happen much often, often enough that they might be visible on the oscilloscope without the need of single trace/record/memorize.

[newtekuser]

Thank you all for putting me on the right track!
I modified the pic program to cycle some text on the lcd every few seconds and set an edge trigger on 500mV level and I do see a signal captured every now and then (attached) even if it's not a square wave. In my case I just want to see something other than noise (ie data) to establish whether something is alive or not. At least now I know my scope is working, so thank you!

Btw, do I need to do the probe calibration before I set the probe attenuation to 10x or after?

[tautech]

Btw, do I need to do the probe calibration before I set the probe attenuation to 10x or after?

Compensation only works on 10x mode....try both and see. 

[kcbrown]

Btw, do I need to do the probe calibration before I set the probe attenuation to 10x or after?

To be clear, there is a massive difference between setting the attenuation value on the scope to 10x versus setting the attenuation of the probe itself to 10x.

Unless the probe has a fixed attenuation value, it'll have a slider switch somewhere on it (most have this on the probe body itself) that allows you to change its attenuation.  Since 1x on the scope gets you correct display values, that means the probe is currently set to 1x.

For proper compensation adjustment, you need the probe to be set to 10x, and the scope should be set to that as well so that displayed values are correct.

[egonotto]

Hello,

"- top left: signal from the 4mhz crystal"

two errors: 4 mHz = 0.004 Hz. I think you mean 4 MHz.
For 4 MHz you need a much higher sample rate. Chose 200 ns / div.

In your picture you got a wrong alias signal.

Best regards
egonotto

[newtekuser]

Hello,

"- top left: signal from the 4mhz crystal"

two errors: 4 mHz = 0.004 Hz. I think you mean 4 MHz.
For 4 MHz you need a much higher sample rate. Chose 200 ns / div.

In your picture you got a wrong alias signal.

Best regards
egonotto

That's right, I meant MHz and not sure anymore if that pic was indeed from the 4.6MHz signal but I do know the 'measure' function (measure frequency) said 4.6MHz.

[egonotto]

Hello,

here is an example. A 4 MHz sinus correct in picture signal.png and wrong with bad adjustment in error_alias.png.

Best regards
egonotto

[newtekuser]

Hello,

here is an example. A 4 MHz sinus correct in picture signal.png and wrong with bad adjustment in error_alias.png.

Best regards
egonotto

Probed the 4.6MHz signal again using 200ns timing and 1.25GS/s, pic attached