Jet lagged in NY I was laying in my bed and then the obvious idea occurred to me that every power amp I know has always done the power supply all wrong. How come?
Power amps draw most current in the lowest frequencies. Furthermore, in the lowest frequencies, both stereo channels typically are recorded in phase. Not to do so would lead to cancellation in play back, and usually the bass is just recorded in mono and mixed to both channels.
In all power amps I have ever opened, built, read about or seen schematics of, both channels are operating in the same phase and this is not the best way to do it, probably.
The reason is that in the lower frequencies, this draws current from the positive or the negative rail at the same time. Let me clarify this with a little sim I did. The load of the opamps consists of the feedback loop.
I took two simple opamp circuits connected to the same positive and negative rails. In the first option, U1 is connected inverted, and U3 non-inverted to the same rails. In the second option, U2 and U4 are both connected non-inverting to the same rails. The difference this makes in the current drawn from the rails is glaringly obvious. In the first option, current drawn from each rail is half of that in the second option, at double the frequency.
When looking at the FFT of these currents, it becomes clear that in the first option, the high peak at 1KHz you can see in the second option just isn't there. Furthermore, because the wave shape is much more gentle, the hash between the other peaks is lower too in the first option. Because they overlap it is difficult to see, but all the other peaks at higher frequencies are identical between the two options. But, as said before, the grass is a bit lower.
I think the implications are easy to understand. We should no longer build our power amps with both channels in the same phase; one should be reversed. Since speakers are a floating load, it is easy to accommodate this phase reversal by reversing the colours on one of the loudspeaker outputs. Plus, of course, we need an inverting stage preceding one of the channels.
Your reservoir caps will become four times more effective by doing so, just to mention one advantage (half the current is drawn at double the frequency). Also, maximum current drawn from each secondary is halved, potentially leading to less need for copper. Finally, the current drawn from the rails looks cleaner, which is no disadvantage either.
I am sure that I am not the first one to come up with this idea, but since it was new to me, it might be new to you too.
Have fun,
Vac
Power amps draw most current in the lowest frequencies. Furthermore, in the lowest frequencies, both stereo channels typically are recorded in phase. Not to do so would lead to cancellation in play back, and usually the bass is just recorded in mono and mixed to both channels.
In all power amps I have ever opened, built, read about or seen schematics of, both channels are operating in the same phase and this is not the best way to do it, probably.
The reason is that in the lower frequencies, this draws current from the positive or the negative rail at the same time. Let me clarify this with a little sim I did. The load of the opamps consists of the feedback loop.
I took two simple opamp circuits connected to the same positive and negative rails. In the first option, U1 is connected inverted, and U3 non-inverted to the same rails. In the second option, U2 and U4 are both connected non-inverting to the same rails. The difference this makes in the current drawn from the rails is glaringly obvious. In the first option, current drawn from each rail is half of that in the second option, at double the frequency.
When looking at the FFT of these currents, it becomes clear that in the first option, the high peak at 1KHz you can see in the second option just isn't there. Furthermore, because the wave shape is much more gentle, the hash between the other peaks is lower too in the first option. Because they overlap it is difficult to see, but all the other peaks at higher frequencies are identical between the two options. But, as said before, the grass is a bit lower.
I think the implications are easy to understand. We should no longer build our power amps with both channels in the same phase; one should be reversed. Since speakers are a floating load, it is easy to accommodate this phase reversal by reversing the colours on one of the loudspeaker outputs. Plus, of course, we need an inverting stage preceding one of the channels.
Your reservoir caps will become four times more effective by doing so, just to mention one advantage (half the current is drawn at double the frequency). Also, maximum current drawn from each secondary is halved, potentially leading to less need for copper. Finally, the current drawn from the rails looks cleaner, which is no disadvantage either.
I am sure that I am not the first one to come up with this idea, but since it was new to me, it might be new to you too.
Have fun,
Vac
Last edited:
I think you are proposing to invert one channel of a stereo signal prior to entering a stereo amplifier, and then to modify the speaker for that channel so that the driver(s) has reversed polarity compared to the other channel's speaker driver(s). The benefit then being that if the stereo amplifier used a common power supply then the signal induced ripple voltage on the main supply capacitors would be a lower level for most common music material.
Yes Tim indeed, but it is even simpler. Just reverse the speaker output plugs for one channel on the amp. No modification of the speaker required.
Very interesting...maybe I can get a little extra out of my p.a. subs with this trick. Easy enough to try.
What an interesting idea. Perhaps there's something come-back-and-bite-you-in-the-*** about it, but I can't think what it might be right now.
I think Harmon Kardon did this in the 80s with some amplifiers. Not sure if there is a patent anywhere or why it is not done more. One reason is some decided that each amplifier side needs it's own filter capacitors and just share a transformer.
Inverting one amplifier is fine. Just make it obvious that one speaker output positive is ground.
Inverting one amplifier is fine. Just make it obvious that one speaker output positive is ground.
Very interesting...maybe I can get a little extra out of my p.a. subs with this trick. Easy enough to try.
For simple valve amps, two preamp stages that pull anode current in anti-phase are often connected to the same decoupled supply rail for that very benefit.
Some amplifier companies, like Dynaco, used to specify distortion with out of phase inputs for this reason.
Otherwise, the full power low frequency distortion shot up alarmingly.
Last edited:
Not everyone has been doing it wrong; those making bridged amplifiers have always been doing it right.
I seem to recall B&O ? doing something like this way back yonder. It seemed an intriguing idea.
Nice work 🙂
Nice work 🙂
Correct but not new at all Grimson does that in some models regardless that their bad implementation of the input buffer produces horrible noises on shut down ...
Opamps tend to have +1 amplification in non invert mode and many ppl can hear difference between inverted an non inverted mode. So there are cases where this is not a good idea.
Good to hear that it does indeed make a difference! And for such a simple thought, I was convinced about not being the first one to have it. But perhaps it is useful to bring it up again. I know for sure that the next amp I make will be connected in this way; will keep you posted.
If ' many people' can hear that difference, imagine what they will hear when the supply ripple disapperas mostly and the power supply get loaded by only a quarter! Audio nirvana! What a great idea!
jan
QSC and others in the high power amplifier market have been doing this for a while. Others use two transformers with the primaries wired differently.
Won't be an issue until someone patents it!
Won't be an issue until someone patents it!
The guy that mixes The Red Hot Chillie Peppers live shows has been advocating using this technique for years:
AV: Fighting For Power - A Way To Increase The Actual Power That Amplifiers Deliver To Loudspeakers - Pro Sound Web
AV: Fighting For Power - A Way To Increase The Actual Power That Amplifiers Deliver To Loudspeakers - Pro Sound Web
- Status
- Not open for further replies.