Convert Push-Pull Output to SE with Passive Components

[pnielsen]

I am using a stock home stereo amplifier to output a continuous audio test signal into a resistive load. The amp input is from a portable MP3 player.

The amp has a push-pull output stage. IOW the current reverses once each "cycle", and there is no earth reference. For experimental purposes, I would like to convert this to single ended, meaning no current reversal.

What is the simplest way this can be done, preferably using passive components, external to the amp and prior to the resistive load?

 

[dknguyen]

Your terminology is incorrect. I think you mean you want to change the bipolar signal from your amp into a unipolar signal?

In which case, couple the amp output to a resistive divider through an AC coupling capacitor. Bias your resistive divider at 1/2 the max negative peak of your amp output so it can never drop below zero volts to produce current reversal.

The part that consists of C1, R1, and R2. Amp output enters from the left and exits into MP3 player on the right:
https://www.experimentalistsanonymous.com/ve3wwg/lib/exe/fetch.php?media=ef_ac.png

But your MP3 player also probably has a bias that can't mismatch. I'm not an audio guy so I may be missing something specific about your MP3 player input (I suspect this might already be accomodated for in the MP3 player circuitry).

 

[ChrisP58]

The only passive way that I can think of is with a transformer.

 

[pnielsen]

Yes, unipolar. I think SE was an old vacuum tube term. Showing my age.

Can ChrisP58 please explain what type of transformer to use and how it would be wired? I have a few PA line transformers on hand.

Thanks.

 

[ChrisP58]

The outputs of your MP3 player are probably 2 single ended, ground referenced outputs that are 180° out of phase with each other. The term is Bridge Tied Load. This is frequently done to increase the output power capacity with low voltage supply rails.
As such, you might be fine with just using one output with respect to gnd. You will need to AC couple it though.

If you do need a transformer, you want something with a primary impedance greater than that of the output of the MP3 player, with an acceptable turns ratio. The primary to secondary voltage will be be affected by that turns ratio.
By PA line transformer do you mean a 70 Volt speaker transformer? They should probably be OK, but you'll probably want to use the highest 'power' tap on the primary. (which will be the lowest P to S turns ratio)
If you want something smaller, look at some of these: https://www.mouser.com/Xicon/Passiv...ignal-Transformers/_/N-5gbg?P=1z0z5h6Z1z0zls8

 

[pnielsen]

As mentioned in my OP, the MP3 player feeds into a domestic type stereo amplifier. So it is the output from the latter that I need to convert to a unipolar signal. The MP3 will not provide sufficint current on its own for my application.

I have uploaded an image of the type of transformer I was referring to.

If I understand your latest rely correctly, I would wire the amplifier output across the primary common and 60W tap, which is the highest power available on this particular transformer. The unipolar output then appears across the common and 8 ohm secondary.

Can you please explain how this removes the push-pull current reversal that would be present at the amp's output? Do the common tags need to be tied to ground?

On related point, it appears wiring the transformer as above will reduce the output voltage of the amp. I assume it would be OK to wire the transformer the other way around, i.e. amp output into the 8 ohm tap. That way I can step the voltage up instead of down.

 

[Pommie]

Can you scope one of the outputs (of the power amp) to see if it ever goes negative relative to ground?

Mike.

 

[pnielsen]

Sorry, I am away from my bench until next week. But I thought most domestic hi-fi amps used a BTL circuit that swings positive-negative at the output.

However, as I understand, the external transformer suggested by ChrisP58 would remove the DC on its "secondary" side so I am unsure what you are getting at.

Would you be able to explain?

 

[ChrisP58]

My apologies for misreading your original post. I missed that there was a separate power amplifier. Please disregard the first paragraph in post #5

Since audio is an AC signal, the output of an audio amplifier needs to reverse polarity.

But I thought most domestic hi-fi amps used a BTL bipolar circuit that swings positive-negative at the output with respect to ground.

Internally, the amplifier is usually a complimentary circuit that pulls the output pin positive and negative as it follows the AC input signal. This complimentary circuit uses a bipolar power supply with balanced positive and negative voltage rails. With respect to ground,

As for BTL in hi-fi amplifiers, that is typically a feature is some stereo amps that allow the two channels to work as one. It is usually a switch that ties the two inputs together with a 180° phase reversal in front of the second channel. Then you connect your speaker load, usually a big mono subwoofer, directly across the two amplifier output terminals, ignoring the two ground terminals.

I would suggest that you double check your amplifier. Unless it's something rare, you can probably use it's output directly, without the need for a transformer.

 

[Nigel Goodwin]

I am using a stock home stereo amplifier to output a continuous audio test signal into a resistive load. The amp input is from a portable MP3 player.

The amp has a push-pull output stage. IOW the current reverses once each "cycle", and there is no earth reference. For experimental purposes, I would like to convert this to single ended, meaning no current reversal.

What is the simplest way this can be done, preferably using passive components, external to the amp and prior to the resistive load?

Your entire premise is most likely wrong, as already mentioned audio amplifiers aren't 'single ended' - some (mostly cheaper ones) might bridge two amplifiers to give more power to a higher impedance load, but the vast majority don't. There's a VERY simply way to tell, simply measure the DC voltage on the output pins - FROM CHASSIS TO EACH PIN IN TURN - a normal amplifier will be zero volts (or very close), a bridged amplifier will be a substantial voltage, 6V upwards on both pins.

 

[pnielsen]

You appear to be saying my premise that most domestic hi-fi amps produce a bipolar output current is incorrect. This implies to me that they would then be "single ended", i.e. referenced to ground at the output.

I will not be able to perform the test you suggest to confirm this unitl the weekend.

But for the moment, let's imagine the amp output is bipolar and involves current reversal. My original question remains unanswered. How can the output of such an amp (push-pull, BTL or other) be made unipolar using passive components?

For example, what if I wired the primary of a transformer with a split secondary across the amp output, took the signal off the "top" leg of the secondary and grounded the center tap? Since there would be no ground across the output pins, I assume I would have to use the amp's power supply ground.

Comments?

 

[Nigel Goodwin]

You appear to be saying my premise that most domestic hi-fi amps produce a bipolar output current is incorrect. This implies to me that they would then be "single ended", i.e. referenced to ground at the output.

No, single ended refers to a single output device, most commonly used with valve amplifiers, although there have been occasional transistor ones as well.

The term has no bearing on referencing to ground.

Almost all amplifiers are push-pull, with the speaker referenced to ground (or at least chassis) with a tiny number (usually cheap ones) been bridged, where both outputs float at half the HT rail.

I will not be able to perform the test you suggest to confirm this unitl the weekend.

But for the moment, let's imagine the amp output is bipolar and involves current reversal. My original question remains unanswered. How can the output of such an amp (push-pull, BTL or other) be made unipolar using passive components?

For example, what if I wired the primary of a transformer with a split secondary across the amp output, took the signal off the "top" leg of the secondary and grounded the center tap? Since there would be no ground across the output pins, I assume I would have to use the amp's power supply ground.

Comments?

That's a vastly complicated and expensive way to do it, assuming you could even source a suitable transformer? - which is extremely unlikely (why would anyone make a transformer with no market?).

The answer is trivial, just an electrolytic from one of the outputs to the speaker, and the speaker to ground - this uses only ONE of the two bridged amplifiers, halves the available power, and doubles the minimum speaker impedance.

But what are you trying to prove - if you simply ask, we could probably tell you?.

 

[pnielsen]

Yes, I must have a sentimental attachment to valve terminology. I still call my scope a CRO. My younger colleagues think I have a pest problem.

The electrolytic solution is the one I was looking for. Before applying I will check the amp output on my scope to see if the speaker is not already ground referenced.

As I understand, most car amps can operate in BTL mode, so the technique you describe would be useful there as well.

The reason for my OP was simply that a need an output that does not reverse in current, since this would double the frequency in my particular application.

Thank you for you assistance.

 

[ChrisP58]

The reason for my OP was simply that a need an output that does not reverse in current, since this would double the frequency in my particular application.

How about a single series diode?

But, depending on the power, the amplifier may not be happy driving such an unbalanced load. You could use a full wave bridge, but that would double your frequency.

 

[pnielsen]

OK. I will separately try a diode as well as a cap to see which works best. But I am concerned the diode may alter the signal's waveform or DC level.

If an unbalanced loading is a problem for the amp, I can artificially load the other output terminal with an 8 ohm resistor to ground.

Still, I am hoping the amp I have is not BTL and will already be driving directly to ground. If not, I now have a few solutions.

 

[ChrisP58]

A standard silicon diode will drop about 0.65 Volts. Depending of your frequency, you'll probably want a fast recovery or a switching diode. Schottky diodes are fast, and have a lower forward voltage, but have a higher reverse leakage current.

A capacitor (alone) will strip off any DC offset that may be present on the amps output. But will still leave an AC voltage that will produce an Alternating Current (changing direction) on it's output.

 

[pnielsen]

Thank you for clarifying that point. I was thinking though that the diode would clip any part of a BTL amp output that is negative with respect to ground. See uploaded image.

With regard to your comment on the series capacitor, if it shifts the DC level entirely above ground reference, then it would appear the current does not reverse relative to it. That is actually what I was looking for, and what I originally meant by no current reversal.

 

[Nigel Goodwin]

With regard to your comment on the series capacitor, if it shifts the DC level entirely above ground reference, then it would appear the current does not reverse relative to it. That is actually what I was looking for, and what I originally meant by no current reversal.

I suggest you tell us exactly what you're trying to do, as your idea of 'no current reversal' probably isn't what you think.

There's also no issue with 'unbalancing', as is obvious when you understand the way bridged amplifiers work.

 

[pnielsen]

My understanding of "current reversal" is when the direction of the current flow reverses within a conductor. For example, this would occur (referenced to ground) if the ouput of an amplifier swings between a positive and negative rail, as opposed to solely between positive and ground.

I had incorrectly assumed in my OP that the former occurs at the output of a home stereo, push-pull amp. For reasons of experimentation, I was asking here for a way to modify that to a non-reversing current using passive external components. IOW to shift the bipolar signal entirely above ground potential.

Thanks to the responses so far, I now know that the speaker output of this type of amp almost always swings between positve and ground. Therefore, it is in fact not a problem for my intended use. I have uploaded a typical schematic.

With regard to "unbalancing the amp", I was referring to Chrisp58's earlier post, which I had assumed applied to the BTL type, not push-pull.

Thank you for your patience.

 

[Nigel Goodwin]

The amplifier you posted uses a split supply, so the output varies either side of zero, i.e. swings both negative and positive. An older amplifier, with a single supply, uses an output coupling capacitor - and while the output on the transistors varies negative and positive with respect to half of the HT, the speaker output (after the capacitor) swings negative and positive with respect to ground, exactly as the split supply one does.

As I've said before, if you tell us EXACTLY what you're trying to do, we can make sensible suggestions - as it is your entire premise may be wrong?, as your questions make little sense.