0
\$\begingroup\$

Please see the screen shot of the 74LS161 wave form. the yellow line is the clock input. the blue line is the QA counter output. I am surprised to see such a large negative voltage spike of the counter output. The negative spike last about 80ns.

My question is that: is it normal?

btw, nothing is connected to the counter, just one 74LS161 on the breadboard, input clock, enable T,P, disable load, disable clear, and measure the output. I added a 0.1uF despike cap, doesn't help on the output spike.

thanks!

enter image description here

\$\endgroup\$
  • \$\begingroup\$ Was it 74LS or 74HC? Was it a x1 probe? How long was the probe ground lead? \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Mar 18 '17 at 23:49
  • \$\begingroup\$ Classic measurement error with poor probing methods, resulting in high Q undershoot 60 ~ 100% of previous "1" level due to cable capacitance of 100pF/m and ground lead inductance. This depends on Ic=CdV/dt getting stored in the coax capacitance and resonating in the ground lead wire shifting the signal conductor. "ground shift" 74LS has "1" impedance much higher than 50 Ohm , less overshoot. but "0 impedance much less. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Mar 19 '17 at 0:39
  • \$\begingroup\$ One test I often do to see if these types of artifacts are real is to grip the scope probe cable tightly in my hand - if you see a change in the trace it is probably not real and caused by bad grounding as described by Tony. \$\endgroup\$ – Kevin White Mar 19 '17 at 0:57
  • \$\begingroup\$ Quite frankly I would only care about the undershoot if that output were the clock input to another device. Otherwise, it does not matter. \$\endgroup\$ – Misunderstood Mar 19 '17 at 4:03
1
\$\begingroup\$

Prototyping board uses to have a fair amount of inductance. Long wires too.

Any output that switches inductive load tends to ring and show voltage spikes.

Keeping components as close as possible in your prototyping board and short wires usually helps in reducing this effect.

If you then use perfboard and solder the components of your prototype, you should see further improvements.

EDIT:

Also, check that your scope probe is well compensated. As Trevor says, they can easily show effects that aren't actually happening.

\$\endgroup\$
0
\$\begingroup\$

If that's all you have connected then all you have left is the scope probe and cable. What is your probe rated at?

Such scope induced measurement errors are common and irritating. Heisenberg uncertainty principal kicks in.

Proto-board has some interesting loading effects too.

ALSO Do you have a decoupling cap on that chip?

\$\endgroup\$
  • \$\begingroup\$ you are right, I just realized that CH2 is 10x prob, CH1 is 1x prob. After I switched to 1x for CH2, the spikes disappeared. Is it due to 1x is not fine enough to detect those spikes or 10x has issue with high freq. by the way, I calibrated/compensated 10x probe using on board 1KHz square wave. \$\endgroup\$ – Ale Mar 19 '17 at 1:02
  • \$\begingroup\$ Different probe and probe settings add different inductances and capacitances... The ringing you see is a phantom being introduced by attaching the probe itself. Like I said, it can be really irritating, esp. when working up in that ring frequency range :) \$\endgroup\$ – Trevor_G Mar 19 '17 at 1:06
0
\$\begingroup\$

It seems the probe issue, after I switched the CH2 to 1x probe, the spikes disappeared, here is the screen shot. Thanks all the inputs!

enter image description here

\$\endgroup\$
  • 1
    \$\begingroup\$ This tells you, oddly enough, that you need to go back to using 10x probes, and learn how to use them properly. 1x probes are notorious for missing short transients, as their uncompensated capacitance filters out such artifacts. Your original traces with the ringing were clear evidence of classic improper ground connection, with a large probe loop area. What you need to do is connect the ground right at the ground pin of the counter, and keep the ground lead as short as possible. See pdfs.semanticscholar.org/dd01/… particularly Appendix A. \$\endgroup\$ – WhatRoughBeast Mar 19 '17 at 1:28
  • \$\begingroup\$ when I shorten the ground probe wire to less than 1 inch, with 10x CH2, the negative spikes indeed reduced more than 60%. \$\endgroup\$ – Ale Mar 19 '17 at 21:40
0
\$\begingroup\$

Here is scope-probe ringing: 100nS, 2.8pF, Q=1000; input edge is 0.2 nanoseconds: enter image description here

If you want to change the simulation params (there are 13 canned set ups, editable), you can download SWE from robustcircuitdesign.com

Merely click on topleft "examples", select the "scope probe ringing" example, and then back on toplevel window go to bottom left and click 'run". No SPICE skills needed.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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