6
\$\begingroup\$

I don't have an LCR meter and wanted to learn how to use a fast edge pulse generator and an oscilloscope to do capacitance measurements as an academic exercise.

In W2AEW's videos, he builds a simple fast edge pulse generator using a 74AC14 hex Schmitt trigger (inverting) and demonstrates how to use it for capacitance and inductance measurements as well as a time-domain reflectometer.

W2AEW's Schematic

I've built this up on a breadboard, but am getting some different results.

I am measuring the input and output of the first Schmitt trigger with a 100MHz digital oscilloscope, with the same value components (6.8kΩ resistor; 0.047μF capacitor) at 5V DC.

FEP at 200us per div

First, the frequency is about 2.5 kHz, not 6 kHz as in W2AEW's example. I'm not sure what is causing the substantial difference. But for this application, the frequency is not as important, so I haven't worried about it.

Second (and what this question is about), when I look closely at the rise time of the square wave output, I am measuring something like 60ns, instead of <5ns as expected.

FEP at 20ns per div

I realize that measuring something as fast as 2ns with a 100MHz scope is not ideal (I assume I would want something like 500MHz to obtain more detail), but I didn't expect such a difference.

W2AEW's edge, measured with a 350MHz analog scope is quite different looking:

W2AEW at 10ns per div

Is it possible there's something wrong with my circuit, or do I simply not have a fast enough scope to measure such fast edges?


Edit/Update:

Per Oli Glaser's answer, the reason for the delay is that the probes were set to 1x rather than 10x, and limiting bandwidth, possibly as low as 10MHz. By setting the probes to 10x, the waveform immediately starts to resemble what it should (~13ns). The extra delay and capacitance is probably because I am using a breadboard and some unshielded jumpers to connect to the oscilloscope probes.

Here's the resulting improved display:

FEP at 5ns/div and probes 10x

\$\endgroup\$
  • 3
    \$\begingroup\$ Do you have your probe on x10? \$\endgroup\$ – Oli Glaser Aug 17 '13 at 22:01
  • 3
    \$\begingroup\$ In the pic with the 60 ns rise time, at the bottom, it says Ch2: 2.00VB" (roughly). Does the "B" in that mean that the bandwidth limiting function for that input channel is turned on? \$\endgroup\$ – user3624 Aug 17 '13 at 22:02
  • 2
    \$\begingroup\$ Also, I noticed you say you built it on a breadboard - a breadboard is not good for high speed stuff such as this, "dead bug" or a similar technique would be better (or even stripboard) The extra breadboard/wiring capacitance may account for at least some of the frequency difference. \$\endgroup\$ – Oli Glaser Aug 17 '13 at 22:16
  • \$\begingroup\$ 0.047μF is 47nF ... the schematic shows 4.7nF. That could account for the unexpected frequency you are seeing. \$\endgroup\$ – aja Aug 17 '13 at 23:41
  • 1
    \$\begingroup\$ @JYelton - yeh, things get a bit more fiddly at high risetimes/frequencies, so even tiny bits of stray capacitance/inductance can make a big difference. At RF a lot of previously "invisible" components appear, R turns into RLC, lumped into distributed, etc... ;-) \$\endgroup\$ – Oli Glaser Aug 18 '13 at 1:51
4
\$\begingroup\$

Set your probe to x10.

At x1 setting, there is a lot more capacitive loading, and the bandwidth is severely reduced.

See the specs for this typical passive probe below, and look at the difference between the risetime/bandwidth at x1 and x10 settings.

enter image description here

\$\endgroup\$
  • \$\begingroup\$ This is exactly the problem. At 10x setting, the waveform is much closer to the expected (13ns or less). The breadboard and unshielded little test jumpers I am using are certainly responsible for the extra delay/capacitance. \$\endgroup\$ – JYelton Aug 18 '13 at 1:18

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.