# Coupling

Discussion in 'General Electronics Chat' started by vindicate, Jul 6, 2010.

1. ### vindicate Thread Starter Active Member

Joined:
Jul 9, 2009
158
0
When coupling a circuit(like audio) with a capacitor, do you have to have your AC voltage go negative?

For instance if you had a a Sine wave that goes from +5v to 0V and coupled it with a capacitor, wouldn't the capacitor just fill to 5v and block all current?

2. ### SgtWookie Expert

Joined:
Jul 17, 2007
22,176
1,709
It depends. A capacitor blocks DC, but passes the effects of an AC signal. For instance, if you were going to connect your electret mic preamplifier to an audio amplifier that used a bipolar supply, you would likely want the signal to be centered around 0v.

If an AC signal is offset from 0v enough so that it does not traverse 0v, it is more correctly called rippled DC.
If the output side of the cap has a current path to ground via a resistor or high-value choke (inductor), then the average value of the output will be 0v - however, you will still see the AC signal.

Try it in your simulator. Use a high-value resistor (10k or greater) or a high-value inductor (say, 10mH) and see what your output looks like.

3. ### vindicate Thread Starter Active Member

Joined:
Jul 9, 2009
158
0
Ok, your right. If I have a sin from 0-5V the output after the coupling is 2.5 to -2.5.

It thought I tried that before but I guess not.

Again Sgt.Wookie you are the man. Thanks alot.

4. ### GetDeviceInfo Senior Member

Joined:
Jun 7, 2009
1,543
223
all you really need is a change in signal, and this can take place at any opertaional voltage of the cap. The value of the cap and the rate of change will imply an impedance which will attenuate the signal.

5. ### vindicate Thread Starter Active Member

Joined:
Jul 9, 2009
158
0
Look at this example I made:

I'm trying to simulate a microphone as simply as possible. So the DC sawtooth generator is the mic. It's ranges from 500mv to 0mv.

On the right side of the capacitor(before the resistor divider) you get +250mv to -250mv. You add in the resistor divider and get 2.75 to 2.25v(roughly). That's all expected.

I would assume you would then pump that into an opamp to get amplification. But my question is, don't you want to go down to 0V not some arbitrary +V number? If I put what I have into a non-inverting amp with feedback, I would end up getting like 5.5V to 4.5V or something like that.

Am I completely wrong on this, or am I on the right track.

File size:
25.9 KB
Views:
83
6. ### SgtWookie Expert

Joined:
Jul 17, 2007
22,176
1,709
Your resistors are too low in value. Use 47k to ground and 33k to +V, or 100k to ground and 68k to +V. Otherwise you'll start getting distortion at lower frequencies.

If you are using dual supply rails, then you would want the signal to be centered around 0v. However, you are likely going to be using a single supply, so you must bias the signal somewhere near Vcc/2.

Since you have already stated in another thread that you intend to use a TL081, you need to bias the input at Vcc/2 +1.5v to give you the best chance of avoiding clipping.

You will not be able to use a TL081 with a 5v supply, as it would only have a 0.5v output swing. You might as well use 9v in your simulations, because that's what you will be using.

7. ### vindicate Thread Starter Active Member

Joined:
Jul 9, 2009
158
0
I wasn't really trying to simulate the mic circuit that I'm going to build, I was trying to simulate a mic circuit in general just so I can get a better handle on the concepts.

In the simulator all they have is the "ideal" opamp so that's what I'm working with for now.

So you were saying that you want to bias the input at Vcc/2. In my example that would be 2.5V. So technically mu input is biased at VCC/2.

When you amplify that though, how do you make it so the "peaks" are amplified more, and not amplify the bias?

Can you amplify that from a 2.25 to 2.75(a vcc/2 bias) to a 5v to 0v(with a vcc/2 bias). The bias stays the same, just the min and max and amplified. Maybe it's not possible, to me that seems logical though.

Last edited: Jul 7, 2010

Joined:
Jul 9, 2009
158
0
Anyone?
..,

9. ### SgtWookie Expert

Joined:
Jul 17, 2007
22,176
1,709
I see.

Why don't you download a more serious simulator? LTSpice is free, and very serious.

You reference one input to Vcc/2.

One way to do that is to use a single resistive divider between Vcc and GND to establish the reference voltage, and then use much higher-value resistors to the cap and the opamp input. In this case, you could use 10k resistors in the voltage divider, and two 100k resistor to establish the bias on the opamp side of the cap, and the opamp itself.

You don't want to do that with audio. You want the average ref level and the input level to be very close to the same voltage.

10. ### vindicate Thread Starter Active Member

Joined:
Jul 9, 2009
158
0
So you would use 2 resistor dividers? If thats what you mean I'm not quite sure where they would be placed in the circuit

As for LTSpice, I'm working on it. I haven't gotten used to it yet.

11. ### SgtWookie Expert

Joined:
Jul 17, 2007
22,176
1,709
That's not what I said.

12. ### vindicate Thread Starter Active Member

Joined:
Jul 9, 2009
158
0
I guess I assumed the 2 100k resistors would be used as a divider.

13. ### Audioguru New Member

Joined:
Dec 20, 2007
9,411
888
Bias the input of the opamp at near half the supply voltage. It should be filtered so maybe 3 resistors are needed.
Make the opamp circuit have AC gain but a DC gain of 1 by using coupling capacitors.
Use a TL071 which is a TL081 that is selected for low noise.

14. ### vindicate Thread Starter Active Member

Joined:
Jul 9, 2009
158
0
Any chance you would like making a quick drawing of that for me?......Please haha

15. ### Jony130 Senior Member

Joined:
Feb 17, 2009
3,628
1,002
Maybe something like this will do the trick:

File size:
11.7 KB
Views:
21
16. ### Audioguru New Member

Joined:
Dec 20, 2007
9,411
888
I would combine the biasing and filtering of the two circuits into one circuit with fewer parts like this:

File size:
10.2 KB
Views:
22
17. ### SgtWookie Expert

Joined:
Jul 17, 2007
22,176
1,709
Vindicate,
You obviously don't understand these circuits.

Sorry that I've been getting impatient with you.

You need to start reading our E-books, at the beginning, and work through them at your own speed until you understand the basics. Without a good "handle" (understanding) of the basics, you will continue to be hopelessly lost. Start with Chapter 1, and work your way through. If you're really "stuck", first try hard to figure it out yourself, and then ask questions.

A much-more in-depth course is here:
http://www.phy.davidson.edu/instrumentation/NEETS.htm
If you study for eight hours a day, five days a week, you might make it through the course material in under a year. It's a fast-track course oriented towards making people with no prior knowledge of electricity/electronics capable technicians on advanced electronics systems.

18. ### vindicate Thread Starter Active Member

Joined:
Jul 9, 2009
158
0
Yeah, I guess I have a weird way of learning. Sorry..

I have 2 questions for the circuit that audioguru posted.

1: What is the purpse of the resistor on the power rail?

2: Are the capacitors drawn as 2 black lines implying they are ceramic caps and the other electrolytic?

File size:
10.8 KB
Views:
30
19. ### Audioguru New Member

Joined:
Dec 20, 2007
9,411
888
Instead of the RC filter filtering only the DC to the microphone, I moved the RC filter so it also filters the bias voltage to the opamps.

I simply copied the capacitors used by Jony.
I personally try not to use ceramic or electrolytic capacitors to couple audio signals because they cause distortion. Instead I use metalized plastic film capacitors that have a tight tolerance and no distortion.
A ceramic capacitor is an excellent high frequency supply bypass and an electrolytic is a good low frequency supply bypass.