EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: PixieDust on April 07, 2019, 05:48:36 am
-
Hi guys,
I'm working with some ~32Mhz signals at the moment and can't figure out whether I have a problem with my oscilloscope being too low in bandwidth or whether there is something wrong with the signal itself.
I have a Tektronix 2431L scope. According to link #1, it's a 40Mhz real time bandwidth scope (so obviously this is woefully insufficient for my purpose) in real time mode, but in equivalent mode, it should allow me to see up to the 9th harmonic as per link #2:
http://w140.com/tekwiki/wiki/2432 (http://w140.com/tekwiki/wiki/2432)
http://mustcalculate.com/electronics/harmonics.php?f=32M (http://mustcalculate.com/electronics/harmonics.php?f=32M)
Link #3 is just another spec of my scope.
https://www.eurekaspot.com/sp.cfm/DIGOSC/TEK/2431L.html (https://www.eurekaspot.com/sp.cfm/DIGOSC/TEK/2431L.html)
What I'm seeing is pretty much identical to what the guy from this thread is seeing:
https://electronics.stackexchange.com/questions/334082/why-is-30mhz-clock-signal-distorted-on-oscilloscope (https://electronics.stackexchange.com/questions/334082/why-is-30mhz-clock-signal-distorted-on-oscilloscope)
I tried measuring the signal with the spring, but got identical results to when I used a ground lead.
The only problem I'm seeing is that we're both using a Tektronix scope :-DD. Anyway...
I set up my scope to be in repetitive mode, which according to the manual does the following:
"When ON, repetitive sampling for NORMAL, ENVELOPE and AVG mode acquisitions is enabled. At SEC/DIV settings of 200ns and faster, the time base is sampling the incoming waveform at its maximum rate of 250 megasamples/second. With REPET OFF, data points that fall between the actual digitized points on the waveform are interpolated to obtain their displayed positions. Interpolation allows expansion of the acquired data to a SEC/DIV setting of 2 ns on a single-event acquisition up to a Useful Storage Bandwidth of 100 MHz. At SEC/DIV settings of 100 ns and faster, REPET ON invokes random sampling (also called equivalent-time sampling) of the incoming signal, extending the bandwidth of the instrument to 300MHz for repetitive signals."
So 300Mhz might not be drastically more than the other guy's 200MHz but I'm not sure.
The other signals that are derived from clock appear to behave as expected, so I have a feeling my clock signal is fine and is a square wave, but my device isn't working, so I'm currently investigating the clock signal.
Would love to hear your opinions on what I'm experiencing.
EDIT: Forgot to mention that I have these probes:
https://www.caltestelectronics.com/ctitem/20-probe-oscilloscope-passive-voltage/CT3133RA (https://www.caltestelectronics.com/ctitem/20-probe-oscilloscope-passive-voltage/CT3133RA)
-
Here are the waveforms I'm getting. The first one is without 50 Ohm off, the second one is in 50 Ohm mode.
-
Don't use 50Ω mode, it is only for signals that are carried via coax from other instruments. (It also has a much lower power tolerance than the normal scope input: you can burn your scope's inputs by frobbing the 50Ω setting just to see what happens.)
The waveform on the display looks like ringing. Did you compensate the probe to your scope?
-
What sort of probes are you using for your scope and how does the GND connection look like?
The 10cm GND lead normally used with scopes completey wrecks the signal integrity.
Just try to hold the probe with one hand, and then fumble with the GND wire with the other (Make it an "O" ring, fold it in a ball, etc.)
-
Don't use 50Ω mode, it is only for signals that are carried via coax from other instruments. (It also has a much lower power tolerance than the normal scope input: you can burn your scope's inputs by frobbing the 50Ω setting just to see what happens.)
The waveform on the display looks like ringing. Did you compensate the probe to your scope?
Ok thanks. Yep LF compensated. They don't appear to have HF compensation.
What sort of probes are you using for your scope and how does the GND connection look like?
The 10cm GND lead normally used with scopes completey wrecks the signal integrity.
Just try to hold the probe with one hand, and then fumble with the GND wire with the other (Make it an "O" ring, fold it in a ball, etc.)
I used the ground spring method.
These are the probes I have:
https://www.caltestelectronics.com/ctitem/20-probe-oscilloscope-passive-voltage/CT3133RA (https://www.caltestelectronics.com/ctitem/20-probe-oscilloscope-passive-voltage/CT3133RA)
-
How do you generate the 32MHz? What's the output impedance? What is the next stage, does it happen to be an inverter with a resistor in - out? Schematic please.