In simulation with 24 volts +/- the front end will swing around 36 +/- peak to peak
which will give you around 25.45 volts rms. So around 80 watts at 8 ohms and around 150 watts at 4 ohms. this does not take into account the drop in the out put stage so around 75watts into 8 ohm and 145 watts into 4 ohm. 5 pair of output should be able to give you the 145 watts into 4 ohm. Once again class AB not sure about pure class A.
which will give you around 25.45 volts rms. So around 80 watts at 8 ohms and around 150 watts at 4 ohms. this does not take into account the drop in the out put stage so around 75watts into 8 ohm and 145 watts into 4 ohm. 5 pair of output should be able to give you the 145 watts into 4 ohm. Once again class AB not sure about pure class A.
I've never seen anyone report peak to peak swing with +/- voltages. Did you mean peak to peak of +45 V? Isn't peak to peak an absolute value, basically? Also, Vrms should be 15.91 V for a 45 V peak to peak signal. Besides, I don't see how you can possibly get more than 22.63 Vrms from a +/-32 V supply. The only way you could get 31.82 Vrms is with a square wave or something close; certainly more harmonics than a pure sine.
If there is something that I am missing, then my apologies: please explain.
Jim has never learned how to manipulate voltage units.
Jim's explanation of voltage is completely wrong and as a result his predictions of power are worthless.
+-32Vdc supply rails cannot allow a BA3 to output +-45Vpk at the speaker.
It could allow an output of 45Vpp.That is very different.
Let's assume the prediction is 45Vpp then the sinewave equivalent is 45Vpp / 2 / sqrt(2) = ~15.9Vac.
the predicted sinewave power is 15.9Vac^2 / Rload = ~32W into 8r0.
However I suspect that the 15.9Vac prediction based on 45Vpp is too low.
I would expect (32+32) / 3 as the likely Vac into a high resistance load. Into an 8r0 load expect ~ a couple of volts less and into 4r0 another couple of volts below the 8r0 value.
Prediction into 1k0 ~ 20Vac
Prediction into 8r0 ~ 18Vac+-1
Prediction into 4r0 ~ 16Vac+-2
And all I need is a pencil and paper.
One MUST be able to check the outputs of any software based prediction for accuracy before one can rely on the results.
That "MUST" demands that the user must also be able to determine manually from a reasonable model very similar results. That requires a deep understanding of what the software is doing for the user.
Jim's explanation of voltage is completely wrong and as a result his predictions of power are worthless.
+-32Vdc supply rails cannot allow a BA3 to output +-45Vpk at the speaker.
It could allow an output of 45Vpp.That is very different.
Let's assume the prediction is 45Vpp then the sinewave equivalent is 45Vpp / 2 / sqrt(2) = ~15.9Vac.
the predicted sinewave power is 15.9Vac^2 / Rload = ~32W into 8r0.
However I suspect that the 15.9Vac prediction based on 45Vpp is too low.
I would expect (32+32) / 3 as the likely Vac into a high resistance load. Into an 8r0 load expect ~ a couple of volts less and into 4r0 another couple of volts below the 8r0 value.
Prediction into 1k0 ~ 20Vac
Prediction into 8r0 ~ 18Vac+-1
Prediction into 4r0 ~ 16Vac+-2
And all I need is a pencil and paper.
One MUST be able to check the outputs of any software based prediction for accuracy before one can rely on the results.
That "MUST" demands that the user must also be able to determine manually from a reasonable model very similar results. That requires a deep understanding of what the software is doing for the user.
Last edited:
I don't really care if some people haven't learned some things but, any one who says "Jim's explanation of voltage is completely wrong and as a result his predictions of power are worthless." should be forced to 6 months, minimum, Moderator duty. To learn how to cumunicate 
How to comunicate in an appropriate way to others who may not be so experienced...
I know, it goes on and on... In 6 months there will be all these newBs who havent read anything but car stereo crap and know almost everything the rest of the world does???
I'm sure there could have been someone who would be making you feel like Jim right now, maybe your mother?, but... The content was probably correct. But I don't F$%^& give a %^&* when I read that BS!
How to comunicate in an appropriate way to others who may not be so experienced...I know, it goes on and on... In 6 months there will be all these newBs who havent read anything but car stereo crap and know almost everything the rest of the world does???
I'm sure there could have been someone who would be making you feel like Jim right now, maybe your mother?, but... The content was probably correct. But I don't F$%^& give a %^&* when I read that BS!
I don't really care if some people haven't learned some things but, any one who says "Jim's explanation of voltage is completely wrong and as a result his predictions of power are worthless." should be forced to 6 months, minimum, Moderator duty. To learn how to communicate
I agree 100%
Ok let's see if we can agree on some things or if I'm all wet. In simulation with a +/- 32 volt supply in simulation one side of the front end will swing 24 volts peak or for a total swing of 48 volts peak to peak. As I understand the balanced circuit it acts the same as a bridged amp so V1(side 1 ) - V2(side 2) equal to 2 times the voltage swing of one side.
Therefore the output at the speaker would be 48 volts peak or 2 times that for 96 volts peak to peak at the speaker. I understand this does not include any loses in the output stage. If this is not correct please let me know where I am going wrong.
Therefore the output at the speaker would be 48 volts peak or 2 times that for 96 volts peak to peak at the speaker. I understand this does not include any loses in the output stage. If this is not correct please let me know where I am going wrong.
You were going along just fine until the "Therefore the output at the speaker would be 48 volts peak or 2 times that for 96 volts peak to peak at the speaker. I understand this does not include any loses in the output stage."
In the previous sentence you said "one side of the front end will swing 24 volts peak or for a total swing of 48 volts peak to peak" Good enough. While this is happening, the other side is doing the opposite. -24V peak
At a point during the positive peak, you have +24V on one side and -24V on the other. Some 180 degrees later, you might have that same side doing a - peak, while the other side is doing a + peak. Measureing across those 2 outputs will measure 48Vp-p. Thats all your gonna get
In the previous sentence you said "one side of the front end will swing 24 volts peak or for a total swing of 48 volts peak to peak" Good enough. While this is happening, the other side is doing the opposite. -24V peak
At a point during the positive peak, you have +24V on one side and -24V on the other. Some 180 degrees later, you might have that same side doing a - peak, while the other side is doing a + peak. Measureing across those 2 outputs will measure 48Vp-p. Thats all your gonna get
I think you're right. An unbalanced amplifier would ground one terminal of the speaker, and vary the other between +24 V and -24 V, for a total of 48 Vpp. A balanced amplifier does not ground either terminal of the speaker, so when one terminal is at +24 V the other is below ground at -24 V, thus, the speaker sees a potential that ranges between +48 V and -48 V, for a total of 96 Vpp.
Apart from the fact that you originally said 45 Vpp and now you're saying 48 Vpp, that would nevertheless be upwards of 30 Vrms measured across the speaker terminals, not across a single output pin versus ground.
EDIT: In other words, the maximum voltage that a speaker sees from an unbalanced amplifier (half the BA-3b) is +24 V, the minimum is -24 V, and the peak-to-peak is 48 V. But, you have to take into account that the balanced BA-3b exposes the speaker to a maximum voltage of +48 V, and minimum of -48 V, and a peak-to-peak of 96 V. It doesn't quite make sense, at first, that a 64 V power supply (+/- 32V) can create 96 Vpp, but it really can because the reference voltage for the speaker is constantly moving rather than being fixed at 0 V. I've missed this detail on paper when dealing with DAC circuits, so I totally understand how others in this thread got confused by thinking of everything in terms of ground-referenced voltages without remembering that the speaker is not referenced to ground.
P.S. You do have to make sure that the +/-32 V power supply can provide enough current.
Last edited:
Now is when we need pictures
Or maybe, now is when you go build it and try and measure it. You probably would have to use rms to measure it, and I don't think I'm willing to go there just yet...
The BA-3 your trying to engineer, not the balanced thingy, produces +and- 24Volts out with an input of 1V (Hypothetically). 24Volts maximum, in any language 24V. No 2X B.S. 24Volts peak!
Stop making me bang on my kybd!
You could call that output from 1 BA-3 48 Volts Peak to Peak. But it is still 24 Volts peak. The difference between the 2 terms is one does not take into account the polarity. It is an absolute value of 24 Volts peak! 48 Volts peak to peak is between the 2, as in, the difference.
You could put 2 BA-3s together...
When feeding them out of phase inputs. Did you read that? Go reread it again. When feeding them out of phase inputs, you will still get 24V peak out of each. But then, in differential, or Ballanced mode, if one has a peak 24 Volts positive, and 180 degrees later, 24V negative, it also has the other side producing 24V Peak. The same as 2 seperate amplifiers. Since it has a + and a - input, at the same time, we can, as you did, treat them like a balanced or bridged amplifier. And, call the output 48V peak to Peak. Don't be adding your car stereo buddies 'ol addage of it makes 2X the power or whatever. I'm tellin you, 1 BA-3 produces 24V out, two of'em in balanced/bridged mode would produce 24 Volts out of each, but, out of phase that can be called 48 Volts out. If you feed 2 BA-3s both an in phase signal you get 0, Nothing, Nota balanced or diferential out. Measureing between them will give you paratically nothing! Measureing 1 output to ground would be 24 Volts Peak. Measureing the other output would be the same 24 Volts Peak. In phase they are only in parallel and only current adds.
I left the club after two Martini's I think I needed another...
Or maybe, now is when you go build it and try and measure it. You probably would have to use rms to measure it, and I don't think I'm willing to go there just yet...
The BA-3 your trying to engineer, not the balanced thingy, produces +and- 24Volts out with an input of 1V (Hypothetically). 24Volts maximum, in any language 24V. No 2X B.S. 24Volts peak!
Stop making me bang on my kybd!
You could call that output from 1 BA-3 48 Volts Peak to Peak. But it is still 24 Volts peak. The difference between the 2 terms is one does not take into account the polarity. It is an absolute value of 24 Volts peak! 48 Volts peak to peak is between the 2, as in, the difference.
You could put 2 BA-3s together...
When feeding them out of phase inputs. Did you read that? Go reread it again. When feeding them out of phase inputs, you will still get 24V peak out of each. But then, in differential, or Ballanced mode, if one has a peak 24 Volts positive, and 180 degrees later, 24V negative, it also has the other side producing 24V Peak. The same as 2 seperate amplifiers. Since it has a + and a - input, at the same time, we can, as you did, treat them like a balanced or bridged amplifier. And, call the output 48V peak to Peak. Don't be adding your car stereo buddies 'ol addage of it makes 2X the power or whatever. I'm tellin you, 1 BA-3 produces 24V out, two of'em in balanced/bridged mode would produce 24 Volts out of each, but, out of phase that can be called 48 Volts out. If you feed 2 BA-3s both an in phase signal you get 0, Nothing, Nota balanced or diferential out. Measureing between them will give you paratically nothing! Measureing 1 output to ground would be 24 Volts Peak. Measureing the other output would be the same 24 Volts Peak. In phase they are only in parallel and only current adds.
I left the club after two Martini's I think I needed another...
Yes - that is correct; and you can never have a peak higher than the rail voltage. Being on 'one side' has nothing to do with it.
No. Perhaps that equation is true concerning current, but you can't have a voltage peak higher than the rail.
No. It's impossible to have a voltage higher than the rail.
The losses are real, of course, but that's not actually important in this case. The thing to remember is that you can't have an output voltage higher than the rail voltage. There voltage can't be created out of nowhere -- the voltage has to come from the rail.
Remember that an amplifier is nothing more than an AC oscillator for your DC power supply.
This is not the whole amplifier. This is the driver portion of the amplifier. To get the 100 watts you need a BA- output stage, (it would be easiest to use the Push-Pull), a bunch of power mosfets, a big power supply, and an enormous amount of heatsinking.
And, for this discussion, we are talking about 2 BA-2 output stages, per channel. Yep, 2 more for the other channel. Hmmm, did Nelson say these output stages draw in excess of 300W for 2ch? O.K. this amp is sucking 600W of power even without a signal...
Back to the 24V out of this BA-3X. N.P. ussually talks in 8 ohm outputs. 24V on a bridge loaded 8 ohms is (48V across 8 ohms) 6A (first approximation easy math)??? 6*6*8=288 Watts peak! Is that what you have in mind?
Oh yea, the question was How did N.P. get 100W at the begining of the thread? Were those rms watts or peak watts?
100W/8ohms=(12.5)sqrt=3.535A He got 3.535A through an 8 ohm speaker probably. Might have needed almost 28.3 Volts on that 8 ohm speaker.
Back to the 24V out of this BA-3X. N.P. ussually talks in 8 ohm outputs. 24V on a bridge loaded 8 ohms is (48V across 8 ohms) 6A (first approximation easy math)??? 6*6*8=288 Watts peak! Is that what you have in mind?
Oh yea, the question was How did N.P. get 100W at the begining of the thread? Were those rms watts or peak watts?
100W/8ohms=(12.5)sqrt=3.535A He got 3.535A through an 8 ohm speaker probably. Might have needed almost 28.3 Volts on that 8 ohm speaker.