I've always wondered in thinking about the differences between solid state (SS) and tube amps, how the typical 50 uF in tube amp power supplies compares to the 10,000 uF in typical solid state amps. In fact, we did this calculation many years ago before some of the high end tube amps came out - time to take another look. How does the 3300 uF used in early VTL amps (stated in the "Vacuum Tube Logic Book by VTL") compare to 100,000 uF used in the heavy weight solid state amps. The two factors are energy storage to minimize droop in unregulated supplies and ripple voltage to minimize hum and intermodulation of hum with the signal.
For now let's take a look at energy storage for some typical examples:
The energy stored (Joules)in a capacitor is given by:
E = .5 * C * V^2 Joules
Lets take a typical solid state amp with 40 volt supplies and a total of 10,000 uF of capacitance in the power supply. The energy available in the caps is then:
.5 * 10,000 * 10^-6 * 40^2 = 8 Joules
High power VTL tube power amps have an amazing 3300 uF at about 500 V per mono block - the energy stored is then:
.5 * 3300 * 10^-6 * 500^2 = 412 Joules
Whereas a more typical tube amp might have 50 uF at 450 V available to the output tubes:
.5 * 50 * 10^-6 * 450^2 = 5 Joules
A high end medium power solid state amp with 100,000 uf at 40 V:
this is simply ten times the previous solid state example or 80 Joules.
A high end high power solid state amp with 100,000 uf at 80 V:
.5 * 100,000 * 10^-6 * 80^2 = 320 Joules
Typical tube and SS power supply energy compares reasonably well and is in the same ball park. It seems reasonable to upgrade medium power amps to have 20 to 100 Joules and high power amps to 100 to 400 Joules when trying to duplicate the performance of "super amps".
Pete B.
For now let's take a look at energy storage for some typical examples:
The energy stored (Joules)in a capacitor is given by:
E = .5 * C * V^2 Joules
Lets take a typical solid state amp with 40 volt supplies and a total of 10,000 uF of capacitance in the power supply. The energy available in the caps is then:
.5 * 10,000 * 10^-6 * 40^2 = 8 Joules
High power VTL tube power amps have an amazing 3300 uF at about 500 V per mono block - the energy stored is then:
.5 * 3300 * 10^-6 * 500^2 = 412 Joules
Whereas a more typical tube amp might have 50 uF at 450 V available to the output tubes:
.5 * 50 * 10^-6 * 450^2 = 5 Joules
A high end medium power solid state amp with 100,000 uf at 40 V:
this is simply ten times the previous solid state example or 80 Joules.
A high end high power solid state amp with 100,000 uf at 80 V:
.5 * 100,000 * 10^-6 * 80^2 = 320 Joules
Typical tube and SS power supply energy compares reasonably well and is in the same ball park. It seems reasonable to upgrade medium power amps to have 20 to 100 Joules and high power amps to 100 to 400 Joules when trying to duplicate the performance of "super amps".
Pete B.
Hello Pete!
You're thoughts are quite interesting, however there's something more to say on tube amps.
Tube amps have 2 more factors: some of them use pentode power tubes, and most of them have a choke.
Let's imagine the typical tube amp: a PP of pentodes, the plates connected to the first node of the psu, likely on that 47uF cap. After that, there's a 10H choke and another 47uF cap: here the screens are connected. After, a 10K resistor and another 47uF cap. Here the B+ for the driver and preamp tubes is taken.
You say that the two factors are energy storage to minimize drop in unregulated supplies and ripple voltage to minimize hum and intermodulation of hum with the signal. Well let's consider each one:
- unregulated supply in this case is not a concern: the current in a pentode is determined by the screen voltage (not the plate voltage) for a given Vg-k. In the arrangement I've discussed before, actually the screens sit on a well regulated supply (thanks to the choke). And since the screen current is 1/10th of the plate current typically, we don't need a large cap for hum reduction.
- hum reduction in a tube amp isn't a function of psu capacitance: the presence of a choke and the proper decouplement of the stages have a major role. That's because as we already said the most critical (hum intended) components of a tube amp either have an almost regulated supply (the screens) or a minimal current consumption (driver stages).
However most of those suppositions fall when the tube amp is a triode SE.
-
You're thoughts are quite interesting, however there's something more to say on tube amps.
Tube amps have 2 more factors: some of them use pentode power tubes, and most of them have a choke.
Let's imagine the typical tube amp: a PP of pentodes, the plates connected to the first node of the psu, likely on that 47uF cap. After that, there's a 10H choke and another 47uF cap: here the screens are connected. After, a 10K resistor and another 47uF cap. Here the B+ for the driver and preamp tubes is taken.
You say that the two factors are energy storage to minimize drop in unregulated supplies and ripple voltage to minimize hum and intermodulation of hum with the signal. Well let's consider each one:
- unregulated supply in this case is not a concern: the current in a pentode is determined by the screen voltage (not the plate voltage) for a given Vg-k. In the arrangement I've discussed before, actually the screens sit on a well regulated supply (thanks to the choke). And since the screen current is 1/10th of the plate current typically, we don't need a large cap for hum reduction.
- hum reduction in a tube amp isn't a function of psu capacitance: the presence of a choke and the proper decouplement of the stages have a major role. That's because as we already said the most critical (hum intended) components of a tube amp either have an almost regulated supply (the screens) or a minimal current consumption (driver stages).
However most of those suppositions fall when the tube amp is a triode SE.
-
Hi Giaime,
Thanks for your input, as I said I was not considering the ripple and hum issue.
It turns out that the VTLs do not have a choke, but I understand your point, it's hard to cover all the cases. I believe that your point is that the input and driver stages are the critical points for hum reduction and perhaps I should clarify my point. What you say is true when the amp is operating in it's linear region below clipping, but once it clips the feedback no longer corrects for ripple from the PSU and it can be seen and heard on the flat tops of the clipped waveform. Certainly, clipping should be avoided but I'm sure that many run their systems clipping part of the time. This is also when the PSU stored energy is important to hopefully keep the amp out of clipping, if possible.
Your claim of clean power for the screens being most important holds only when there is enough plate supply voltage for normal operation of the amp, here again when clipping occurs the linear model does not apply.
I agree the case of a tube amp with filter chokes is more complex but this was not the example that I had in mind.
Pete B.
Thanks for your input, as I said I was not considering the ripple and hum issue.
It turns out that the VTLs do not have a choke, but I understand your point, it's hard to cover all the cases. I believe that your point is that the input and driver stages are the critical points for hum reduction and perhaps I should clarify my point. What you say is true when the amp is operating in it's linear region below clipping, but once it clips the feedback no longer corrects for ripple from the PSU and it can be seen and heard on the flat tops of the clipped waveform. Certainly, clipping should be avoided but I'm sure that many run their systems clipping part of the time. This is also when the PSU stored energy is important to hopefully keep the amp out of clipping, if possible.
Your claim of clean power for the screens being most important holds only when there is enough plate supply voltage for normal operation of the amp, here again when clipping occurs the linear model does not apply.
I agree the case of a tube amp with filter chokes is more complex but this was not the example that I had in mind.
Pete B.
Giaime said:
- Status
- Not open for further replies.