Is this standard tolerance for square wave generator output?

[alsetalokin4017]

@MarkF:  Your trace is plotted at 5 V/div vertically, is this correct? 
Also I am surprised that my Fast Edge result (2.35 ns risetime) is so much better than yours (3.20 ns), since my construction technique and connections are not as good as yours and I am using the HC version of the chip rather than the AC version. Is your scope unlocked to full bandwidth?  (see my reply #14 above for the trace from my scope and FEO)

[pwnell]

I tried again, this time using a 74HC14N.  It did not make a real difference.  Tried 50ohm termination directly attached to the scope - then it got worse - 5ns, remove 50 ohm termination then it bounces back to 3.2ns to 3.7ns depending on the position of my tongue.

However I think the point is, the scope shows a fast rising and falling edge just fine.  So my issue is definitely with the output of the square wave generator built in to the MSO5104.  I also tested it again with a 50ohm terminator (I use a BNC T adaptor with a 50 ohm terminator on one end and the other connected to a BNC 58 cable connected to the output of the AWG on the same scope.  The AWG output is set to 50 ohm, the scope is HighZ but with the attached 50 ohm terminator.  (I have not yet received my through terminators).

Here is the comparison (second image):

Yellow = without terminator, red is with terminator.  Since this did not affect the rise time which is quite high for a square wave that should have a half period of 33.33ns, I guess it is just the quality of the built in AWG.  A rise and fall time of ~12ns means the wave lost 12ns so there is only 21ns "on" and 21ns "off.  Interesting.

[srb1954]

Yellow = without terminator, red is with terminator.  Since this did not affect the rise time which is quite high for a square wave that should have a half period of 33.33ns, I guess it is just the quality of the built in AWG.  A rise and fall time of ~12ns means the wave lost 12ns so there is only 21ns "on" and 21ns "off.  Interesting.

The AWG waveform would appear to be slightly better than the datasheet specified performance of 15ns rise and fall time. I would say that there is nothing particularly wrong with the operation of the AWG.

If you need higher performance then you will probably have to go for an external dedicated AWG or function generator. However, be prepared to spend quite a bit of money if you want to get significantly better performance.

[pwnell]

I have a DG972 on order that will ship out in a couple of days... Will compare them once it arrives.

[gcewing]

If your scope has a "sinc interpolation" mode or some such you might like to make sure it's turned off. It can make high-frequency waveforms look more curvey than they really are.

[pwnell]

I looked for it but do not believe the MSO5104 has that option...

[Bud]

Ordered one, will test. Might also try and build that circuit if I can find the components. Thanks for the help.

Сan't the scope itself connect a 50\$\Omega\$ resistor? In the generator output settings, there is an output type.

The output type is only to properly show the output signal level . Nothing is physically switched in the output.

[MarkF]

@MarkF:  Your trace is plotted at 5 V/div vertically, is this correct? 
Also I am surprised that my Fast Edge result (2.35 ns risetime) is so much better than yours (3.20 ns), since my construction technique and connections are not as good as yours and I am using the HC version of the chip rather than the AC version. Is your scope unlocked to full bandwidth?  (see my reply #14 above for the trace from my scope and FEO)

I had a 50Ω coax cable and 50Ω thru terminator connected from my PCB to the scope.

You need to ignore the vertical values. 
The scope thinks it has a 10x probe connected instead of a 1:1 cable.
I didn't bother to change the vertical settings.

YES.  My scope is an unlocked DS1074Z to 100MHz bandwidth.

If I'm thinking correctly here (i.e. a triangle waveform with these rise and fall times):

  bandwidth = 1 / (rise_time + fall_time) = 1 / (3.2ns + 3.4ns) = 151MHz

as my measured bandwidth of the scope.  Assuming, the edges are faster than the scope can measure.

By my same calculation:
  Your 2.35ns rise time with the same fall time would result in a 212MHz bandwidth.
Are you summing four outputs or five?  It looks like you amplitude is 500mV higher than mine.

I must be looking at this totally wrong???   

[radiolistener]

If I'm thinking correctly here (i.e. a triangle waveform with these rise and fall times):

  bandwidth = 1 / (rise_time + fall_time) = 1 / (3.2ns + 3.4ns) = 151MHz

for 10%/90%: bandwidth = 0.35 / rise_time

for 20%/80%: bandwidth = 0.22 / rise_time

in your case bandwidth = 350 / max(3.2, 3.4) = 102 MHz

[alsetalokin4017]

@MarkF:
My FEO uses 5 summed outputs of the Texas Instruments SN74HC14N chip and the sixth gate for the oscillator. I think when I built it I selected the resistors to match closely, fwiw. As far as amplitude goes it's a matter of power supply setting between 3-6 volts, but this doesn't seem to affect the risetime.
I know I have some AC versions around here somewhere that I can try when/if I find them. Like I said, I would expect yours to be faster since you used the BNC connector and your circuit is nice and tight, whereas mine is spread out more on the perfboard, no ground plane, and I used a socket for the chip. Maybe I just got lucky and got hold of a chip that exceeds tolerance, it does happen.

[MarkF]

My chip is in a socket also.

I added the BNC external input to buffer the square wave output from my DIY Function Generator.
In buffering the BNC input, I used one of the five gates.
This gate buffers both the oscillator and the external input.

Therefore, I only have four gates driving my output.  You have a little more drive power.
Maybe my rise time is slower because the four gates don't have the power to drive a 50Ω load?

[radiolistener]

Maybe my rise time is slower because the four gates don't have the power to drive a 50Ω load?

Try to measure amplitude for on two different load. For example for 25 Ohm and 100 Ohm. What is amplitude? It allows to determine exact output impedance.

[pwnell]

So I went a bit crazy.  Since I do not have the tools to build a properly laid-out circuit board for this level of performance, I picked up one of these - a (certified) 40ps pulse generator.

So I attached it directly to the 1MOhm input of my scope, triggered on it and got a poor 2.299ns rise time.  According to 0.4/2.299e-9 that gives 173MHz bandwidth (my scope should do 350MHz).  The shape of the square wave is beautiful though.  See below.



So I got a brief moment of insight and decided to use my newly acquired 50ohm through terminator.  With that in the pathway between the generator and the scope, I got the signal below - 860ps rise time.  That yields 0.4/860e-12 = 465MHz or 407MHz if you use 0.35 instead.  The signal however does have more overshoot.



If I change the scale to 1ns per div I get a rise time of 752ps (532MHz).  Not sure if at this scale it is still accurate.



And comparing the overshoot with 50ohm terminator



and without