I often like doing the "different" thing in life, so I came up with an interesting thought: We always like to hear about amplifiers that double up power as load halves, therefore you have an amplifier that produces for example: 100W into 8 ohms, 200W into 4 ohms, 400W into 2 ohms, 800W into 1 ohm, etc.
Now from the homework I have been doing, two very simple aspects of the power supply determine this behaviour:
1. The rail voltage determines the maximum power into higher impedence loads.
2. The available current determines the maximum power into lower impedence loads.
So, let's examine this theoretical amplifier as described above. Let's assume it has the ability to swing to + and - 28V on the output stage using 35V rails. Let's also say that the power supply is capable of delivering 30A worth of current. This gives us a theoretical transformer of 25-0-25V rated at 1500VA.
* To drive an 8 ohm load, the output stage will consume around 3.5 Amps of current
* To drive a 4 ohm load, the output stage will consume around 7 Amps of current
* To drive a 2 ohm load, the output stage will consume around 14 Amps of current
* To drive a 1 ohm load, the output stage will consume around 28 Amps of current
This is nothing new, but what is very interesting here (to me at least) is the fact that this kind of amplifier and power supply is built around the idea that it must be able to deliver the required current and voltage into the lowest required impedance. At that lowest impedance (1 ohm in this example), and at maximum output the power supply is being used to its full potential. At any higher impedance load, the power supply is sitting idling and not working nearly to its full potential.
Take for example the 4 ohm load situation. The power supply can deliver 30 Amps, but at maximum swing we are only consuming 7 Amps.
Now I hear you saying things like "headroom" to me. I understand headroom very well, but let me tell you my theory here:
I don't think there is any further headroom left when the amplifier swings to its maximum voltage capabilities into a higher impedance load despite the fact that we are only using 7A out of the possible 30A of current drive capability. This is because the only way to get more power into a 4 ohm speaker will be to increase the voltage swing capability. Only then can the current consumption by the load increase. Because of this the "headroom" argument (in my mind at least) is moot.
Our power supply is capable of delivering a massive 848W worth of power, but because we are driving a 4 ohm speaker we need to make do with 200W.
So, the reason then for all this discussion is the following:
What would happen if we were to turn this thing on its head and build an amplifier capable of delivering maximum power into the load it is presented with. Let's limit that load to anything between 1 ohm and 8 ohms for the sake of argument so we have some definite parameters to work with. We then ensure the output stage, drivers, heat sinks, etc are up to the task of delivering maximum available current at the lowest defined impedance so we have enough "balls" for the task.
Because we need both voltage swing and current in different proportions depending on the load driven, we would in theory do the following:
1. Use a power supply with a higher voltage so as to maximize power output into higher impedance loads.
2. Design the amplifier such that the internal gain can be adjusted automatically to ensure full use of the current for the given load.
In this way, the following options present itself: Let's use the same 1500VA transformer we had before, but change it to have a secondary output with 40-0-40V. This means the maximum current will now be 18.75Amps (as opposed to the 30A we had before).
The 40-0-40V rails will give us around 56-0-56V after rectification, so let's assume we can use 48-0-48V of that in the output stage.
If we recalculate the power ratings now, it looks like this:
* 288W into 8 ohms (limited by the swing voltage) using 6 Amps of current (vs the 100W we had previously)
* 576W into 4 ohms (limited by the swing voltage) using 12 Amps of current (vs the 200W we had previously)
* 703W into 2 ohms (limited by the current) using 18.75 Amps of current (vs the 400W we had previously)
* 351W into 1 ohm (limited by the current) using 18.75 Amps of current (vs the 800W we had previously)
As you can see, the power now only doubles between 8 and 4 ohms. In theory the power supply rail voltages could be pushed even further to get more out of 8 ohms, but I am quite happy with something like this - it makes a lot more sense to me than the standard way described above. How many of us need 800W into 1 ohm load anyway? Obviously there are speakers that do need this, but they are by far in the minority.
Finally, the question I am sure will be on the top of your mind: How will you know where to set the amplifier gain because speakers are very different regarding the impedance vs frequency plots, and ABC, and DEF.
You could obviously have the user select this as an option, but even then you will not know what the correct setting should be: A speaker which is advertised as a 4 ohm load could well dip down below 3 ohms at certain frequencies, so that only leaves 1 option in my mind: You make a circuit that measures the minimum impedance of the speaker connected to it, and adjusts the gain for full power at that load.
I am seriously considering building this - it sounds like it has exactly the right amount of crazy for me :banned:
Your thoughts?
Cheers,
Ian.
Now from the homework I have been doing, two very simple aspects of the power supply determine this behaviour:
1. The rail voltage determines the maximum power into higher impedence loads.
2. The available current determines the maximum power into lower impedence loads.
So, let's examine this theoretical amplifier as described above. Let's assume it has the ability to swing to + and - 28V on the output stage using 35V rails. Let's also say that the power supply is capable of delivering 30A worth of current. This gives us a theoretical transformer of 25-0-25V rated at 1500VA.
* To drive an 8 ohm load, the output stage will consume around 3.5 Amps of current
* To drive a 4 ohm load, the output stage will consume around 7 Amps of current
* To drive a 2 ohm load, the output stage will consume around 14 Amps of current
* To drive a 1 ohm load, the output stage will consume around 28 Amps of current
This is nothing new, but what is very interesting here (to me at least) is the fact that this kind of amplifier and power supply is built around the idea that it must be able to deliver the required current and voltage into the lowest required impedance. At that lowest impedance (1 ohm in this example), and at maximum output the power supply is being used to its full potential. At any higher impedance load, the power supply is sitting idling and not working nearly to its full potential.
Take for example the 4 ohm load situation. The power supply can deliver 30 Amps, but at maximum swing we are only consuming 7 Amps.
Now I hear you saying things like "headroom" to me. I understand headroom very well, but let me tell you my theory here:
I don't think there is any further headroom left when the amplifier swings to its maximum voltage capabilities into a higher impedance load despite the fact that we are only using 7A out of the possible 30A of current drive capability. This is because the only way to get more power into a 4 ohm speaker will be to increase the voltage swing capability. Only then can the current consumption by the load increase. Because of this the "headroom" argument (in my mind at least) is moot.
Our power supply is capable of delivering a massive 848W worth of power, but because we are driving a 4 ohm speaker we need to make do with 200W.
So, the reason then for all this discussion is the following:
What would happen if we were to turn this thing on its head and build an amplifier capable of delivering maximum power into the load it is presented with. Let's limit that load to anything between 1 ohm and 8 ohms for the sake of argument so we have some definite parameters to work with. We then ensure the output stage, drivers, heat sinks, etc are up to the task of delivering maximum available current at the lowest defined impedance so we have enough "balls" for the task.
Because we need both voltage swing and current in different proportions depending on the load driven, we would in theory do the following:
1. Use a power supply with a higher voltage so as to maximize power output into higher impedance loads.
2. Design the amplifier such that the internal gain can be adjusted automatically to ensure full use of the current for the given load.
In this way, the following options present itself: Let's use the same 1500VA transformer we had before, but change it to have a secondary output with 40-0-40V. This means the maximum current will now be 18.75Amps (as opposed to the 30A we had before).
The 40-0-40V rails will give us around 56-0-56V after rectification, so let's assume we can use 48-0-48V of that in the output stage.
If we recalculate the power ratings now, it looks like this:
* 288W into 8 ohms (limited by the swing voltage) using 6 Amps of current (vs the 100W we had previously)
* 576W into 4 ohms (limited by the swing voltage) using 12 Amps of current (vs the 200W we had previously)
* 703W into 2 ohms (limited by the current) using 18.75 Amps of current (vs the 400W we had previously)
* 351W into 1 ohm (limited by the current) using 18.75 Amps of current (vs the 800W we had previously)
As you can see, the power now only doubles between 8 and 4 ohms. In theory the power supply rail voltages could be pushed even further to get more out of 8 ohms, but I am quite happy with something like this - it makes a lot more sense to me than the standard way described above. How many of us need 800W into 1 ohm load anyway? Obviously there are speakers that do need this, but they are by far in the minority.
Finally, the question I am sure will be on the top of your mind: How will you know where to set the amplifier gain because speakers are very different regarding the impedance vs frequency plots, and ABC, and DEF.
You could obviously have the user select this as an option, but even then you will not know what the correct setting should be: A speaker which is advertised as a 4 ohm load could well dip down below 3 ohms at certain frequencies, so that only leaves 1 option in my mind: You make a circuit that measures the minimum impedance of the speaker connected to it, and adjusts the gain for full power at that load.
I am seriously considering building this - it sounds like it has exactly the right amount of crazy for me :banned:
Your thoughts?
Cheers,
Ian.
