Inverting audio signal with ne5532

Thread Starter

Dean Rantala

Joined Sep 27, 2018
34
I have an ne5532... Imagine that I am using the following schematic (this image is from Elliot Sound Products)

gain.gif

I want to drive two amp modules (yes, they are bridgeable). So, I need to invert the signal of ONE module... What I plan to do is: duplicate the above circuit, but reverse the -/+ inputs (leaving the rest of the circuit as-is). Is this doable? Or do I have to do something else here? Is the feedback resistor going to be an issue when feed back to the positive vs negative input?
 

Thread Starter

Dean Rantala

Joined Sep 27, 2018
34
So would something (literally) like this work?

invert_unity.jpg

Should I just add one of these between the input and one of the bridged amps? Or should I buffer the other amp as well?

In other words... Let's just consider a single channel (using two amps). Should I send the signal directly to amp #1 and run the same signal through the above and send the output from that opamp to amp #2? Or... should I buffer both amps? A unity buffer to amp #1 and a unity inverter to amp #2?

Thanks!
 

AnalogKid

Joined Aug 1, 2013
10,987
The circuit in post #1 has a gain of 2, while that of post #2 has a gain of -1. Rather than try to have the audio source drive two amplifier circuits and try to match the gains between them, better to have the audio drive only the non-inverting stage, and drive the inverting stage with the output of the non-inverting stage. That way, if you want to change the gain (or filter the input, or whatever) you need modify only one circuit. Also, I would increase the resistors in #3 to 10K. this will reduce the load on the opamp output stage and improve the accuracy of the signal inversion.

ak
 

Thread Starter

Dean Rantala

Joined Sep 27, 2018
34
Oh yeah....

The circuit in post #1 has a gain of 2, while that of post #2 has a gain of -1. Rather than try to have the audio source drive two amplifier circuits and try to match the gains between them, better to have the audio drive only the non-inverting stage, and drive the inverting stage with the output of the non-inverting stage. That way, if you want to change the gain (or filter the input, or whatever) you need modify only one circuit. Also, I would increase the resistors in #3 to 10K. this will reduce the load on the opamp output stage and improve the accuracy of the signal inversion.

ak
So.. I breadboarded this up and wired in inline to the actual stereo... oh yeah.. Almost no audible noise (the original "inverter board" for this was super noisy, had a bad power-on DC surge and just plain did not sound as transparent). Gonna whip this up onto a PCB after I finish playing the Hell Freezes Over concert.. it just sounds too damn good....
 

BR-549

Joined Sep 22, 2013
4,928
May I ask....What are you trying to do? I do not understand you. Can you draw a labeled picture?

What are amp modules? Are you working with 1 channel or 2?

Are you looking for an inverter or phase splitter?
 

Thread Starter

Dean Rantala

Joined Sep 27, 2018
34
I am restoring (and I use that word loosely) a Harman Kardon 75+ stereo/quadraphonic receiver.

This is a mid-1970's stereo receiver and was back when HK built some really good stuff. These receivers were over-engineered and have:

-> Two transformers, one for left channel, another for the right channel
-> Modular design - many of the functions are broken into boards that can be removed like computer cards
-> The ability to operate in 4 channel discrete mode (4 independent channels)
-> The ability to operate in 2 channel stereo mode (front+rear left amps bridged, front+rear right amps bridged)
-> The 4 channel decoding circuitry is even a removable (upgradable!) module that was meant to be upgraded if desired

So.. internally, there are 4 individual amplifier modules. Each "amp module" has it's own heat sink and is truly discrete from one another. 2 amp modules use one transformer, the other 2 use the second transformer. On the back, there is a switch that switches from 4 channel mode to stereo mode. When you flip this over to stereo, this switch (a really big switch that has like 20 freakin poles) essentially bridges the amps into a 2+2. So rather than 20 watts RMS x 4, you get 40 watts RMS x 2. Given the build of the amplifier modules and the insane power supply.. the actual power output is almost double the advertised specs (mentioned above). But they were built like this for extreme headroom and dynamic range. The official frequency response of this is 5hz - 50khz.

To put it simple: the amplifier section on these is absolutely amazing. The pre-amp section, however... that is another story. This includes the inverter board.

I am a bit of an audiophile and so I am rebuilding the entire pre-amp section of this.

The point of this thread was to figure out how I could eliminate the original "inverter board". This is the board that - when the stereo/quad switch is flipped - sends positive phase to the front amp modules and negative phase to the [normally rear] amp modules. Each amp module is a simple push/pull design (+/- 36 volt with center tap ground). When bridged, while one amp is pushing, the other module is pulling. So rather than a 36 volt max potential, you end up with a 72 volt total potential (if I understand the topology correctly). More power.

I have bread-boarded an 5532 as a unity inverter and was able to successfully replace the existing inverter board with the 5532 circuit.

Here is the original inverter board schematic:

Inverter.jpg

So.. that is the long-winded story as to what I am working on lol. That old board was really noisy (hiss), had a horrible DC offset during start-up (I smoked a couple cheap bench speakers before I realized what was happening) and even after replacing all caps and the transistors.. it just did not help it.

Replacing the above circuit with a single 5532 as an inverter did the trick.. and sounds better.
 

BR-549

Joined Sep 22, 2013
4,928
Thank you. I see what you mean. My comment on it wouldn't help you. I would repopulate those boards with high quality, high tolerance components and match the transistors. Then listen to it.

I am not an audio tech. Sounds like a fine amp. Good luck to ya.
 

AnalogKid

Joined Aug 1, 2013
10,987
That type of circuit is called a phase splitter. It is based on the fact that the signal voltages on the collector and emitter are 180 degrees out of phase (one is inverted with respect to the other). If you get the resistor values balanced for whatever is loading the two outputs, the two voltages are exactly the same amplitude. They cancel nearly perfectly when summed.

The circuit is the basis for many hands-free intercom and telephone hybrid designs. The patent for the original tube version was granted in the late 1920's or mid-1930's.

ak
 

Thread Starter

Dean Rantala

Joined Sep 27, 2018
34
Would someone mind giving me a pointer here? If I were working with an OPA, I would be able to figure this out, but..

I found this: https://www.electronics-tutorials.ws/amplifier/amp_2.html BUT... the original circuit seems a bit more complicated than that and I want to be sure..

Basically, this is what I need to know: what is the (give or take) aprox gain of the *original* inverter / phase splitter circuit? I ran into a couple (not very big issues) where now.. most my gain is in the pre-amp section (about 12db) and virtually no gain exists in the inverter board.. and since there are a couple passive circuits BETWEEN the pre-amp board and the inverter, these passive circuits are not behaving as they used to. This is not a real big deal - I have re-arranged the factory layout so that these items (balance control, for example) are now BEFORE the pre-amp board. But I just kinda wanna understand things a bit more before I button this guy up. The voltage to the above inverter board is supposed to be 30VDC.

I have a feeling most the gain (originally) was in the inverter board and the pre-amp board more/less maintained unity. Strangely, the volume control for this receiver is literally the first thing the signal hits after the input selection.. before even the pre-amp section. It was (originally) like this:

[INPUT SWITCH] --> [50K VOL POT] --> [PREAMP BOARD] --> [HI/LOW CUT SW] --> [50K BALANCE POT] --> [INVERTER] --> [AMP]

The balance pot was not working until completely right or left - and then the signal completely cut out. I moved the balance control to be after the vol pot and it now works correctly. The part that still does not work correctly is the HI/LOW cut switch function. Mainly, the high cut switch. It really has no affect on the signal when activated. This switch basically shunts a 15nF cap to ground when active (cutting the upper frequencies).. I NEVER use this feature, but it is the fact of the matter. Again, I think this is because the gain is too high by the time it leaves the pre-amp and runs through these passive components. Much like the balance control was. Interestingly, the low-cut works (which - when activated - routes the signal through a 47nF cap to cut low frequencies).
 
Top