This question may not be valid, please excuse my ignorance. I am having difficulty measuring a decent square wave at 500kHZ. The image below shows a 4 volt CMOS output with a frequency of 500kHz connected directly to the scope using a 2' BNC cable.

I am an engineering student so both pieces of equipment are older and on the used side. Kinda like my first car. The function generator is made by BK Precision and is labeled as a 5MHz generator. (Model: 4011). The scope is a Tektronix 465B that is rated to handle at least 5MHz. The function generator has a TR and TF <= 120ps.

Scope Manual

Function Generator Manual

Obviously the output is nowhere close to a square wave. The horizontal scale is set to .5us. The same output generates a square wave at much smaller frequencies. Is there anyway to tell from the image below which piece of equipment is ancient and should be replaced first? I can't imagine a world at 500KHz.

Any help would be much appreciated! Thanks for reading.

Output Image http://itssimplydesign.com/outputwave.jpg

======UPDATE====== 8:30 pm 2012.11.03

I attempted to calibrate the oscilloscope, but all of the tests turned out to be ok. The Trace Rotation screw was about a quarter of a degree off so I adjusted the rotation a tiny bit. The ASTIG screw was already aligned for optimal focus. In the original post I was using BNC to alligator clips to probe the circuit. I just went to the lab and picked up a 100MHz Tektronix probe to replace the clips. The output of the function generator is the same...


When the function generator is attached to a CMOS inverter (ZVP4105A and ZVN2110A) with a 1pF capacitor the output wave is a perfect square wave. The circuit is shown below:

CMOS Inverter http://itssimplydesign.com/inverter.jpg

All of the parameters remained the same. The function generator is still producing a 4 Volt 500 kHz* input wave. The input wave looks exactly like the first picture above (sawtooth) Yet the output is now perfect. (See image below). I have no idea how this could be possible.

CMOS Output http://itssimplydesign.com/OutputWave_CMOS.jpg

How can a function generator attached directly to a scope produce such an under-compensated wave but when the same generator is attached to a network it turns out normal? TEACH ME!!

  • Edit fyi: Mhz -> kHz.

======Update #2======== 3:07 am 2012.12.03

After pages and pages of my final EE252 lab report I have reached several conclusions. Unfortunately none of the pain staking research has to do with the actual report content. But I am a perfectionist the has to know why things do what they do. Here are the four different scenarios I put together regarding the damping shown in the earlier posts.

Four http://itssimplydesign.com/scope.png

All of the waveforms shown in the image above represent the input waveform only. Refer to the CMOS circuit in the first update for reference. Waveform (a) was produced by placing a BNC to alligator clip directly to the shared gate. Ugly, I know. Waveform (b) was formed by placing the x10 probe directly to the shared gate node. It is almost perfect. Waveform (c) introduces a 50 Ohm termination resistor (to balance the function generators 50 Ohm impedance) between the input and shared gate. This measurement was also taken with the clips placed directly after the terminating resistor. I assume that the continued damping effect is caused by faulty resistor values in addition to a floating output impedance. The final waveform, waveform (c) was created by using the combination of a 50 Ohm resistor and the probe. It was interesting to discover option (b) provided the best results. I assume this is where the probe adjustment would come into play. I will be sure to take a look in the near future.

  • 6
    \$\begingroup\$ Neither should be replaced -- a decent analog scope, like this 100MHz dual-trace, is a great reality check in circuit analysis; any function generator is always useful. Rather, compliment them with better equipment as needed. If working with high-fidelity audio, get a function generator with lower THD in the audible range; if measuring higher than ~10MHz signals, upgrade the oscilloscope. \$\endgroup\$
    – tyblu
    Mar 11, 2012 at 20:17
  • 2
    \$\begingroup\$ Are you using a probe or do you just have a bnc cable connected directly from the function generator to the scope? Your question implies one, but your response to Russ' answer implies the other. \$\endgroup\$
    – The Photon
    Mar 11, 2012 at 22:47
  • 2
    \$\begingroup\$ Nice scope. Hold on to it. Download manuals for your scope: ko4bb.com/manuals/index.php?dir=Tektronix/… bama.edebris.com/manuals/tek/465b \$\endgroup\$
    – markrages
    Mar 12, 2012 at 2:28
  • 1
    \$\begingroup\$ @KevinVermeer - I apologize for the misunderstanding. Yes, the cable is BNC to BNC in the original image above. It is BNC to clips in the updated images. \$\endgroup\$ Mar 12, 2012 at 4:12
  • 2
    \$\begingroup\$ As others have said: Don't throw out any of the two. The thought alone hurts my soul badly. The scope is perfect! (readingjimwilliams.blogspot.com/2012/02/…, readingjimwilliams.blogspot.com/2012/02/…, readingjimwilliams.blogspot.com/2012/02/…). The function generator may not be the most precise piece of equipment, but once you learn about its limitations, you will find it useful. I own a similar one and a fancy machine, and I still like the simple one for quick work. \$\endgroup\$
    – zebonaut
    Mar 12, 2012 at 8:05

2 Answers 2


500kHz is a ways past the maximum specified frequency for square waves for this generator: 100kHz. It appears to have a single-pole RC of about 3.5µs, which would work great for a 0.35/(3.5µs)=100kHz square wave. The output may have a LPF for slew limiting. Also, it is a 50Ω source, so it should be terminated properly to avoid ringing. Try using the TTL and CMOS waveforms, too. B&K have put together this document: Function & Arbitrary Waveform Generator Guidebook .

The external CMOS inverter is not a 50Ω source -- it's source impedance is only a few ohms at most (for low currents) due to VCC and ground impedances and FET RON equivalent resistances. Notice that the output duty cycle isn't 50%, and the edges are ringing.

  • \$\begingroup\$ Thanks for the guidebook. I will definitely put that to use. I had experimented with a 47 Ohm resistor in series with the CMOS inverter (as mentioned in the manual), but did not see much improvement with the ringing. All of the waveforms above are generated with the TTL/CMOS output. \$\endgroup\$ Mar 12, 2012 at 4:19

Ancient but superb oscilloscope!

  • Oscilloscope probe may need adjusting but othewise the scope is well capable of handling this signal.

  • The waveform shows what would probably be expected from a digital generator running near the top end of its frequency range.

Try adjusting probe used to give a correct square wave response. You will find a small adjustment screw on probe, accessed with an about 1/8" staight bradedscrew driver. Turn slowly, do not force - less than 1 turn total travel. Observe results on this waveform and adjust for best rise time shape and then use calibration output waveform on scope. This is at lower middle of picture.

The waveform does appear to have genuine steps in it of about 50 ns duration, which is about what you'd expect from a piece of digital equipment clocked at about 5 Mhz. Note that this is not necessarily "bad" and that modern equipment will no necessarly be substantially better if the spec is also not better. Depending what you are doing this may be very acceptable. As long as you know the limitations of your equipment you can often accept using them near the limits and make suitable allowances.

Report back.


You have substantially changed your question - and also made an observation that needs commenting on:

The input waveform is definitely not a sawtooth. It appears to be a nominal square wave wity digital steps in it and some slew rate limiting somewhere.

The addition of the inverter changes the question. The inverter output is designed to be high or low and to slew rapidly. When the stepped input signal is applied the output stays low as the signal increase in level until a trigger level is reached and then the output goes low "rapidly". This eliminates the stepped effect.

The signal shown is not a perfect square wave. It has non vertical transitions and some "stuff happening" at the transition points. IF that scope has a bandwidth limiting feature )it may have) there will be a button or switch to switch it on and off. Check - if found, ensure it is off.

Oscilloscope probe calibration:

From this very good oscilloscope calibration page

enter image description here

  • \$\begingroup\$ Fantastic, thank you. I will give it a shot. I will let you know how it goes, or at least the year I am transported to. \$\endgroup\$ Mar 11, 2012 at 21:22
  • \$\begingroup\$ I tried fine tuning the scope using the ASTIG and Trace rotation screws but they were already in check! I added an update to the original post that involves a CMOS circuit that produced a perfect wave at the same frequency. \$\endgroup\$ Mar 12, 2012 at 0:41
  • 3
    \$\begingroup\$ The screw Russell is talking about is on the probe itself, not the astigmatism or trace rotation screws. It adjusts the [seres] input capacitance of the probe to compensate for the [shunt] input capacitance of the 'scope at high frequencies. \$\endgroup\$
    – tyblu
    Mar 12, 2012 at 2:21
  • \$\begingroup\$ @tyblu - I see what you mean now. Until about 4 hours ago I had never used a probe. I borrowed one from the school to test the circuit above. Shows how 'hands on' my school is. I am almost a junior and we are still using BNC to clips in labs. \$\endgroup\$ Mar 12, 2012 at 4:14
  • \$\begingroup\$ I used the probe to view the 1kHz square wave generated by the scope calibration bar shown in the image. I set the probe and scope to x10 and viewed the wave at 10mV. I also attached the ground clip to the scope ground input shown in the image above (far left). The square wave seems normal. There isn't any room for adjustment. \$\endgroup\$ Mar 12, 2012 at 4:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.