I like your white board presentation/explanation. It is often difficult to explain complex circuits and to be able to visualize the complex circuitry involved is very helpful. Thanks Paul.
Thanks, Paul. This was great! Tube output stages are a different animal altogether; they output *current*, meaning that the frequency response will *follow* the speaker's impedance curve. I assume this is why you prefer transistor power amps!
In short: You want push pull class A(B) topology, lots of parallel output devices with low source resistance and a big power supply with low resitance everywhere. Usually this means big and heavy and very hot, if its supposed to also sound good.
Florian, not true. Just use one channel on any 2-channel power amp and you can drive a 2-Ohm speaker. You just need 2 of those identical amps. They can be small and low budget like the NAD C275 BEE.
@@raffiequler7510 Bridging gives you double the voltage swing but not generally more current capability. From the amps channel perspective speaker impedance is then cut in half. If the amp was voltage limited before you can play louder though. Of course class D can also give you lots of cheap current, but who wants that.
@@florianhofmann7553 I'm not talking about bridging. I'm talking about using only one channel on an amplifier. Stereophile measured the NAD C370 with just one channel driven and it gave 305 Watts into 2 Ohms continuously.
The explanation given in this video also highlights the reason why the pre-amp or voltage gain stage is where the amplifiers characteristics or colour is imparted in the weak signal coming in. Guitar amps are a good example of this. Weak voltage from the pick ups is increased by the pre-amp. The power amp stage should just increase the current or power so that the speakers can be driven effectively. The power stage shouldn’t impart colour onto the signal.
Alternately, you can do what Bob Carver did with his Sunfire amps by tracking the input signal and maintaining the transistor rail voltage to 6 volts above whatever is required to reproduce the sound output. This results in huge amounts of power while producing little heat.
The output voltage level is still the representation of the signal thus a traditional power amp is still also a voltage amp, but of course with low output impedance.
In most speakers happen also nonlinear drops of impedance which are accompanied by reversal of impedance character. In linear range it is kept near by neutral side. But with apparent mechanical or electrical resonances it may be pushed to capacitance side. At this moment amplifier is in trouble because feedback may lose part of negative feedback and it amplifies affect of nonlinear reproduction of speaker - amplifier. That is why also I do not advice to use too small impedance speakers to even way above needs amplifiers. As example 2 Ohm speaker to 4 Ohm amplifier. It won't burn oversized end stage but will work unpleasant. When used 16Ohms speaker to 8 Ohm amplifier it also affects peformance but while the feedback is increased it's up to user. Only my conclusions
The direct answer, IMHO is pretty much all the above. The type of transistor/tube, the topology, and the power supply for the output stage of the power amp determine its ability to drive a difficult load. A little more complicated in class D.
Which, of course, is the whole point of the final amp - to drive a complex load with such a low impedance that regardless of the connected impedance, the speaker sees a voltage that's a nearly perfect representation of the input signal.
@@glenncurry3041 As you said, tubes are a different beast, having high output impedance. Unlike transistor amps, a high-Z load will damage the output Tx of a tube amp, so Z matching is critical. Negative feedback can help to dynamically improve damping factor but only up to a limit.
so a speaker must be designed so that it responds to a certain p2p voltage varied over frequency with identical output SPL, regardless of frequency, and the amp needs to live with the complex impedance varying between low and high (reflex port resonance frequency), capacitive and inductive loads? and feedback loop compares in and output voltages, correct?
Paul: I now sort of understand this. But the manual for my Revel Salon 2 speakers reads: Impedence: 4 ohms (nominal) and 3.7 ohms (minimum @90 Hz). So how do these impedence specs fit into this video explanation? Also, the 3.7 ohms (minimum) is confusing. Does this mean the speakers can present to the amps a load of 3.7 ohms but not less? Is this a warning that I need real beefy amps to drive these inefficient speakers. Or does it means the amps must be able to drive at least a load of 3.7 ohms? Paul is probably too busy to answer this "minimum" question, so if anyone out there cares to give me an explanation please do so. Thanks and thanks Paul for this tutorial.
Impedance is sort of the resistive value for ac. And it is not presented as normal resistance, as the value changes with different frequencies. Why is it like that… well an example of this would be that a speaker with dual drivers , will have different speaker driver for different frequencies to reproduce them best. A small one that can move fast for high frequencies and a bigger one for the lower. These two drivers should not try to reproduce both high and low - sometimes this will even brake them. So we have to add a way of sorting the frequencies to go to either the tweeter or the bass. This is done by the filter normally a high pass filter and a low pass filter. The easies being a coil in series with the bass and a capacitor in parallel with the tweeter. The speaker drivers and the components being inductive and capitative will result in an advanced impedance seen from the amplifier. And the specs you read is telling us that the lowest impedance in your speakers should be at 3,7ohms at 90hz. This should be within what a 4ohm rated amplifier should be able to handle.
You just need to "size" the amp for the minimum speaker impedance (that final resistor that Paul has drawn into the board is representing the speaker). The lesser the impedance, the more current you need to pump in (I = V / R: I being the current, V being the voltage which is constant when we reach the current stage, already amplified by the voltage gain stage, R being the speaker impedance). A speaker varies its impedance as the sound frequency varies. But, if your amp is okay with 3.7 ohms, say for example at 300Hz, then it will also be okay for the other speaker impedances on all other frequencies, because they require less current.
@@christopherkise So, the speakers should not ask more from the amp than 3.7 ohms at 90 hz. Is this what you are saying? Thanks very much for your response. I've never understood all this.
@@joeb4349 do not overthink these terms. Generally- not always but generally an consumer amplifier should be stable for a range of impedances around the provided and a 4ohm speaker output is probably always +-20%. Ad the speakers tend to be less than flat. The amplifier might have earlier distortion than with a speaker above 4ohms. Generally again the amplifiers tend to have bigger problems with a high capacity in the load than a bit low impedance in total. Do not worry about 0.3ohms impedance. Yet an amplifier saying it is stable at 4, might burn up at 2ohms (so dont push to much. Always listen for distortion, and do not push your amplifier to hard
I’m not sure how the gain is created but at least I’m not mixing up transistors with the amplifier (triangle) part of the circuit any more. Does amp circuit feed the transistor controlling input (base)
Gain is achieved by applying a relatively high voltage from a power supply to a valve plate or a transistor collector that is controlled by a low voltage signal applied to a valve grid or a transistor base. There are a Very Few circuits in which the signal is applied to the cathode or the emitter, but in audio such topology is usually designed to do something other than linearly amplify a signal. Many valve RF amplifiers, such as are used in high-power radio transmitters, apply the signal to the cathode and the grid is grounded. This offers some advantages in both power supply design and the prevention of unwanted oscillations. It took me years to figure out/understand How these things work.
If you use just one channel on virtually any 2-channel power amplifier, you will have plenty of juice to drive any 2-Ohm rated speaker. You just need 2 of those then to drive 2 speakers. No need for buying expensive monoblocks or any other amps rated to be able to drive 2-Ohm loads. They are way too expensive and completely unnecessary.
At 09:15 "just tie these things together" - it looks like you're connecting the base of a transistor from one stage with the base of the transistor from the next stage. I assume it's emitter to base?
Thanks, that makes sense. "In parallel" and "load sharing" would be considered the same stage though, right? Not an additional one, like the video suggests.
