Hello,
I am trying to view a 11 MHz square signal on my PCS500 scope and the wave form is not really square. The signal is from an 11.0592 mhz “metal can” oscillator. Is my scope having a problem?
This is what I’m getting:
Have you observed the signal on any other 'scope (100 MHz, and 100 MS/s if digital, or better would be best)–if not then there is no way knowing what it “should” look like. It is also doubtful that a 11 MHz square wave oscillator module is generating an entirely accurate square wave output.
What sort of coupling or probe are you using, do you know the output impedance of the oscillator?
=================================================================
Also, let’s take a look at what a square wave is?
A true square wave is an amalgamation of all the odd harmonics of the fundamental frequency, to observe even a reasonably close approximation of a square wave requires that the 3rd through 9th odd harmonics be passed through the signal path to the 'scope–for your 11 MHz signal that would mean that the oscilloscope would have to have a 100 MHz band width (11 MHz * 9 = 99 Mhz).
Here is a square wave “built” from the 3rd through 9th harmonics (33, 55, 77, and 99 MHz for your 11 MHz fundamental):
But as the PCS500 is specified as having a 50 MHz bandwidth, those 77 and 99 MHz components will never make it through the 'scope’s input–here is a square wave “built” from the 3rd through 5th harmonics (33, and 55 MHz for your 11 MHz fundamental):
[b]However that 50 MHz bandwidth is measured as the point at which the oscilloscope’s response has dropped to -3 dB relative to a “full amplitude” display. So that 55 MHz harmonic would likely be at least 3dB down (-3 dB) from it’s actual amplitude when displayed.
Here is a square wave “built” from the 3rd through 5th harmonics, but with the amplitude of the 5th harmonic reduced by 3 dB (I.e. at 70.8% of it’s “real” value):[/b]
Is this (above) close to what you are observing?
===============================================================
In the last, just for completeness sake, here is a “square” wave constructed of only the fundamental and 3rd harmonic:
It occurred to me just now lying in bed that I only addressed the analog domain issues of displaying an 11 MHz square wave on a PCS500.
With a 50 MS/s maximum sample rate (Nyquist frequency = 25 MHz) there should be little if any (actually none) of the 3rd (33 MHz) harmonic making it to/through the ADCs. If it does get to the ADCs it will be aliased to 17 MHz.
The 11 MHz square wave trace could well be reduced to just an odd looking IM distorted wave?
OP, can you post any images of what you are seeing?
Thank you for your answer.
As you can see in the captured image, the scope is working at 1Gs/s.
I changed the probe with a Hameg HZ 154 and thing are getting worst (see the image below). I don’t have any data about the oscillator but I tried with other oscillators with freqencies between 10MHz and 20MHz and it looks the same.
I will try to observe it on a better scope as soon as I find one.
Thanks cliffyk, once again a very clear insight in the theory behind square waves and digital scopes.
For high frequency measurements you have to use the x10 setting of the probe.
This setting reduces the capacitive load of the probe.
[quote=“Catalin C.”]Thank you for your answer.
As you can see in the captured image, the scope is working at 1Gs/s.
I changed the probe with a Hameg HZ 154 and thing are getting worst (see the image below). I don’t have any data about the oscillator but I tried with other oscillators with freqencies between 10MHz and 20MHz and it looks the same.
I will try to observe it on a better scope as soon as I find one.
[/quote]
That is what I would expect to see when observing an 11 MHz signal on the PCS500.
The 1 GS/s mode is an Equivalent Sample Rate (ESR) mode, which makes twenty 50 MS/s “passes”, each slightly offset in time¹, to record twenty sample sets of a repetitive waveform. These are combined to produce a more accurate composite “picture” of the repetitive signal. The 1 GS/s rate is an equivalent rate is as it combines twenty acquisitions, each acquired at 50 MS/s; 50 MS/s * 20 passes yields an ER of 1 GS/s.
I.e. the ESR mode does not alter the basic sampling rate of 50 MS/s, nor magically bypass the 25 MHz Nyquist limit of a 50 MS/s sample rate. It is often referred to as RIS, Repetitive Interleaved Sampling.
¹ - Each “pass” is begun at a slightly different time after the trigger event, and is made at the instrument’s maximum real time sample rate, 50 MS/s for the PCS500. However each acquisition is offset in time (horizontally) from the others. The acquisitions are then combined to produce a more accurate representation of he waveform–but with all the bandwidth limitations of the 'scopes input and real time sample rate.
Here is a good illustration of how it works:
You are welcome, that’s what happens when retired engineers have insomnia…
This is what happens when they wake up, once again I hope this can shed more light on the effect of bandwidth on square wave signals.
I do not have a function generator capable of producing a clean 11 MHz square wave (I.e. with a < 1.0 ns rise time)–few people do, however the calibration output of my Lecroy 9354L scope can generate a 2 MHz square wave with a 1.0 ns rise. Using that signal I made the following observation:
figure 1
The top trace in figure 1 is the 2 MHz input, captured at 1 GS/s, 10k points (it was saved to internal memory to allow reconfiguration to capture the lower trace).
The bottom trace is the same 2 Mhz signal, but with a 4.0 MHz digital filter applied–I.e. leaving only a very small amount of the 6 MHz 3rd harmonic.
Figure 2 below presents an FFT spectrum analysis of the two time domain traces shown above. In each trace a cursor, highlighted in yellow, marks the 6 MHz 3rd harmonic of the fundamental frequency:
figure 2
The top trace is the spectrum of the original 2 MHz square wave, with each of the odd harmonic components (out to 80 MHz or so) clearly visible.
The bottom trace is the spectra of the filtered signal, the lower trace in figure 1. It is clear that the 3rd harmonic has been reduced in amplitude, and that subsequent odd harmonics have been eliminated via the filtering.
Now I really understand what are the limitations of my scope.
On the other side, I’m happy that’s nothing wrong with it.
Thank you very much!
[quote=“Catalin C.”]Now I really understand what are the limitations of my scope.
On the other side, I’m happy that’s nothing wrong with it.
Thank you very much![/quote]
Understanding the capabilities and limitations of any tool are the essential elements of making best use of it.