diyAudio (
-   Solid State (
-   -   Constant input impedance polarity switcher (

Steven 13th October 2003 08:53 PM

Constant input impedance polarity switcher
2 Attachment(s)
Some records sound better when the abolute polarity (on both stereo channels) is reversed, maybe because somewhere in the recording and production chain absolute polarity (or phase) was not maintained. Often the change has to do with the positioning of vocals or instruments in the stereo image; is the singer behind the speakers or in front of, etc.
I don't want to go into a discussion whether it should be absolute polarity or absolute phase; I think absolute polarity is more correct, but both are used in practice. Anyhow, you know what I mean :yawn: .
Some commercial amplifiers have a dedicated polarity switch on the front panel to make it easy to switch polarity. Far easier than to walk to your loudspeakers and reverse the wires if you change the disk.
Some years ago I designed a circuit that can do just that. It is built around an opamp, but it could be any other balanced in, single ended output amplifier stage as well. I know that there are plenty of amplifier stages that have a true balanced output that you can use straight or crossed, or with just one of the outputs used. But in this case I would like to present an alternative.

It uses a single opamp that is switchable between inverting and non-inverting mode. The absolute gain is one, when driven from a low source impedance. However, the nice thing about this circuit is that it has the same input impedance in both modes, so it is not necesssary to drive it from a low impedance. You can use it as the final buffer stage of the preamp, just after the volume control potentiometer. Depending on the chosen resistance of the potentiometer you get some additional attenuation but that is the same in both modes. Input bias currents of the opamp result in an offset at the output that is the same in both modes, so no clicks. The feedback loop around the opamp is never interrupted during switching, so no problems with break before make or make before break switches.

The input resistance Ri=R1=R2=R3. The value itself depends on the output impedance of the previous stage, the required noise performance, driving capabilities of the opamp, etc. The output impedance is low. The Zout in the schematic diagram is the output impedance of the previous stage and was used for simulation.

:cool: I haven't seen this circuit before, so I think it is original, but I'm interested to know whether the same thing has already been described earlier. I have seen it before in a version without R3, but then you won't get a constant input resistance.


millwood 13th October 2003 08:57 PM

looks interesting but I don't see how it was switched to a non-inverting configuration.

Nelson Pass 13th October 2003 10:20 PM

Well, I haven't seen it before, and it would appear to
work as stated. Imperfections in the performance of
the op amp could theoretically be a factor, but I imagine
are negligible as a practical matter. Nice job. :cool:

mrfeedback 14th October 2003 02:41 AM

2 Attachment(s)
I posted these two pics a while back - similar but not quite the same.


mrfeedback 14th October 2003 02:42 AM

2 Attachment(s)
And this........

mrfeedback 14th October 2003 02:50 AM

Hello Steven,
Ummm, upon closer look at you circuit you have two different input impedances.
With the switch as shown the input impedance is R1 in parallel with R3 - ie 10K//10K = 5K ohms.
With the switch in the down position, the input impedence is R1 = 10K ohms.
Somebody correct me if I am wrong.

A while back I built a polarity switcher box using latching relays as the switching element and a wired remote control (tired of swapping polarity at the back of the speakers).
I encountered the problem of differing input impedance causing level changes, and then slipped a meter probe at a regulator and blew up the op-amps - ok it was late at night and poor lighting and too many Cab Savs.
The box is still in my rack awaiting further attention but lately too many other things to spend time on - like getting repairs out and money in.

I fully agree that in room acoustic polarity control is mandatory hence my quick circuit lash up, but alas I am still stuck with swapping speaker wires for now.


Nelson Pass 14th October 2003 03:42 AM

Alas, you are right.

Doh! :clown:

No, wait a minute. The negative input tracks the positive input
ergo, the input resistor to the - input sees no current and
does not "exist"

Ah well, either way, only wrong once.

:clown: :clown: :clown: :cool:

Steven 14th October 2003 06:10 AM

Nelson, thanks for answering mrfeedback; I just woke up (Europe).

Mrfeedback, Nelson is right, when the + input of the opamp is driven, the - input follows (feedback) and there is no current through R1, so R1 seems to be infinite.
In inverting mode the input impedance is equal to R1, in non-inverting mode it is equal to R3. As long as R1=R3 you have a constant input impedance. To have also equal (absolute) gain in both modes, R2 should be equal to R1. Then R1=R2=R3.
Maybe I should have added a little more of a description to the circuit.


Steven 14th October 2003 08:32 AM

circuit of mrfeedback
Hi mrfeedback,

Your first circuit is definitely different. I suppose that the two switches form a SPDT switch (both open makes no sense, both closed neither).
In case the upper one is closed you have a normal inverting amp with gain is -1 and a shunt from the - input to the ground which has no function other than a small increase in noise. There is no voltage across that resistor; the - input is already a virtual earth. Input resistance is 100k with the given values.
In case the lower switch is closed, you get a gain of +1, because you first attenuate the signal with a factor of 2 and then amplify wih a factor of 2. But the input resistance is now 200k with the given values.
Disadvantage of this circuit is the different input resistance for positive and negative mode. Also a break before make switch is required otherwise you get a nasty plop at the output when both opamp inputs are short circuited.

I prefer my circuit. :cool:


Steven 14th October 2003 08:39 AM

2nd circuit of mrfeedback
The 2nd circuit of mrfeedback is more or less the same as the first, only now the gain can be controlled from +1 to -1 with a potentiometer and somewhere in the middle it mutes (gain is 0). The disadvantage of different input impedance for the extremes +1 and -1 are the same as in the first circuit, still 200k vs. 100k. Still it is a nice circuit.


All times are GMT. The time now is 06:55 PM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright 1999-2017 diyAudio

Content Relevant URLs by vBSEO 3.3.2