# How to estimate stray capacitance

Discussion in 'General Electronics Chat' started by daskywalker, Apr 23, 2010.

1. ### daskywalker Thread Starter New Member

Mar 28, 2010
13
0
Consider his circuit:

http://www.electro-tech-online.com/...16914481-help-adc-0804-self_clocking_0804.jpg

I was to observe the signal at CLK IN and CLK R with the oscilloscope. And I saw the triangular curve (similar to charging and discharge curve) at CLK IN and a square wave with same frequency at CLK R. Anyone know why it is like that?

I also need to estimate the stray capacitance...is there a formula or method I can use?

2. ### mik3 Senior Member

Feb 4, 2008
4,846
67
CLK R switches between Vcc and ground as to charge/discharge the capacitor. CLK in senses the voltage across the capacitor and at some threshold levels it switches CLK R. Thus CLK R depends on CLK in, thus why they have the same frequency.

What stray capacitance are you talking about?

3. ### daskywalker Thread Starter New Member

Mar 28, 2010
13
0
I guess its the stray capacitance between the CLK in and CLK R where u have the resistor and the capacitor to ground.... I've heard there is a formula like
T (tao)=RC, but honestly I hav no idea what stray capacitance in general means.....

4. ### SgtWookie Expert

Jul 17, 2007
22,194
1,764
It's not really "stray" capacitance; it's parasitic capacitance.

In an ideal world (simulation), capacitors have only capacitance, inductors have only inductance, resistors have only resistance.

Real-world capacitors also have parasitic resistance in series (fairly low) and in parallel (very high), inductance (small amounts).

Real-world resistors also have inductance (wire-wound can have quite a bit) and capacitance (small amounts).

Real-world inductors can have a fair amount of capacitance if they have a lot of turns. They also have resistance.

In a Spice simulation, a wire is a perfect conductor; zero Ohms. It can carry an infinite amount of current with no voltage drop whatsoever, and has no parasitic properties.

In a real-world PCB, a trace 10mm long has an inductance of 15nH at 10MHz.
It also has resistive properties, as copper is not a perfect conductor.
It also has capacitance, depending upon what is near the trace on the same and other layers of the board, and what the dielectric constant of the board material is.

You might find this paper on the dielectric property and loss tangent of FR4 board material interesting:
http://emclab.mst.edu/documents/TR00-1-041.pdf

On a side note; when you touched the pins with your o'scope probe, you added a significant amount of capacitance to the circuit; roughly 20pF to 60pF. If you hold your probe close to, but not touching the pins so that you get a little bit of capacitive coupling, you will see the waveform but not significantly disturb the timing of the circuit.

5. ### daskywalker Thread Starter New Member

Mar 28, 2010
13
0
Oh i forgot to mention one thing. We also observed the signal at the scope when no external capacitors were present. The curves are pretty much the same but with a lot higher frequency and the edges of the waves are curved. Does that help me to estimate the stray capacitance?

6. ### mik3 Senior Member

Feb 4, 2008
4,846
67
Yes, you can estimate the parasitic capacitance by looking at the triangular waveform and estimating the time constant of the circuit (RC). You know R and thus you can calculate C. Bear in mind to include the probe capacitance in the calculations.