Hi,
using a ground or power plane on double sided PCB will ADD to the stray capacitance.
Experts know how to avoid or use to advantage this problem. eg. a single trace above a ground plane has a 110ohm characteristic impedance at high frequencies when correctly proportioned.
Amateurs, like us, wade in without our galloshes (wellingtons) on.
using a ground or power plane on double sided PCB will ADD to the stray capacitance.
Experts know how to avoid or use to advantage this problem. eg. a single trace above a ground plane has a 110ohm characteristic impedance at high frequencies when correctly proportioned.
Amateurs, like us, wade in without our galloshes (wellingtons) on.
gaetan8888 said:Hello
I read few time in the forum that there is stray capacitance in audio circuit.
I was assuming that you have that problem mostly in RF circuit.
Wen I do a pc board I leave large space between lines, is it enough to avoid stray capacitance ?
Thank
Gaetan
One square cm = 3 pF so it's very seldom this has a really big influence. A trace over a groundplane has a fraction of a pF. Did you have any special circuit in mind?
It does depend on the circuit though. If you are working at high gain (even at 'low' frequencies), or using FET-input opamps then a handful of pF stray at the inputs of opamps can lead to unwanted interactions, even oscillation.
You can even get oscillation in unity-gain buffers (ie non-inverting pamp with output linked to -ve input) if, for some reason, you use a resistor instead of a 'xero ohm link/track) if that resistance, and that tiny stray, forms enough phase-shift at HF. This is really easy to achieve if you are also using video opamps or other high-bandwidth parts because diy audio lore 'demands' it
The best thing you can do read is Application Note 47 from Linear Technology: 'High Speed Amplifier Techniques' by Jim Williams. Pure 100% gold becasue it is exhaustive, readable, well-illustrated, and conveys an enormous amount of good information discovered the hard way...
You can even get oscillation in unity-gain buffers (ie non-inverting pamp with output linked to -ve input) if, for some reason, you use a resistor instead of a 'xero ohm link/track) if that resistance, and that tiny stray, forms enough phase-shift at HF. This is really easy to achieve if you are also using video opamps or other high-bandwidth parts because diy audio lore 'demands' it
The best thing you can do read is Application Note 47 from Linear Technology: 'High Speed Amplifier Techniques' by Jim Williams. Pure 100% gold becasue it is exhaustive, readable, well-illustrated, and conveys an enormous amount of good information discovered the hard way...
I'd say in low noise high gain designs with low signal impdenaces, this is not a big problem if the pcb is good.martin clark said:Does dpened on the circut though. If you are working at high gain , even if low frequency, or using FEP-input opams then a handful of pF stray at the input sof opamps can lead to unwanted interactions, even oscillation.
I agree completely, it's a long way down the list of things worth worrying about.
But it's not always easy to convey what 'good' means. A few 10s of pF stray shouldn't be a problem in a preamp; but it could be , very easily, in a badly-thought-out line stage, or a DIY MC phono stage
But it's not always easy to convey what 'good' means. A few 10s of pF stray shouldn't be a problem in a preamp; but it could be , very easily, in a badly-thought-out line stage, or a DIY MC phono stage
Stray capacitance in audio circuit
In reading this thread I saw mention of distributed vs lumped element representations for audio interconnect. At audio frequencies, over the distances encountered in home audio equipment, it is safe to say that a lumped element representation is good enough. Distributed (transmission line) modeling becomes necessary when the propagation delay through a channel becomes a significant percentage of the reciprocal of the highest frequency. Even at 20 KHz, this does not occur until the interconnect reaches lengths of ~1000 feet.
There was also mention that use of a ground plane increases stray capacitance. This statement partially true. Adding a ground plane may increase the capacitance between a signal path and ground. There may be instances where one wishes to maintain a very high imput impedance, for example, where this additional capacitance poses a problem.
However, a ground plane also serves to decrease the effective coupling between two signal traces, and it this coupling that is a potential problem in audio circuits. It is fairly easy to see how a ground plane works. Consider a pair of signal traces with a per unit length coupling capacitance of C1. Now consider the same pair of lines with a ground plane. There will be a per unit length capacitance between each trace and ground of C2. Typically C2 >> C1, so the addition of the ground plane introduces a capacitive voltage divider between the two signal traces.
In reading this thread I saw mention of distributed vs lumped element representations for audio interconnect. At audio frequencies, over the distances encountered in home audio equipment, it is safe to say that a lumped element representation is good enough. Distributed (transmission line) modeling becomes necessary when the propagation delay through a channel becomes a significant percentage of the reciprocal of the highest frequency. Even at 20 KHz, this does not occur until the interconnect reaches lengths of ~1000 feet.
There was also mention that use of a ground plane increases stray capacitance. This statement partially true. Adding a ground plane may increase the capacitance between a signal path and ground. There may be instances where one wishes to maintain a very high imput impedance, for example, where this additional capacitance poses a problem.
However, a ground plane also serves to decrease the effective coupling between two signal traces, and it this coupling that is a potential problem in audio circuits. It is fairly easy to see how a ground plane works. Consider a pair of signal traces with a per unit length coupling capacitance of C1. Now consider the same pair of lines with a ground plane. There will be a per unit length capacitance between each trace and ground of C2. Typically C2 >> C1, so the addition of the ground plane introduces a capacitive voltage divider between the two signal traces.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
- Home
- Amplifiers
- Solid State
- Stray capacitance in audio circuit ?