That's what I question. In the other post, I used the best of my knowledge ( which does not mean much!!!) to compare stage by stage on distortion. I don't see why tubes by default have more distortion.
Two things I notice that stack against tubes.
1) Lower loop gain even from 3 stages of tubes compare with SS IF designed correctly. The first stage is only about 100 even using CCS or mu-follower. Second LTP is about 50. The power tube with OT is unity or less. So it is lower than SS power amp. This does not reduce the harmonics as well.
2) Output impedance is higher with tube amps so it is more sensitive with speakers.
But the very strong point is triodes by nature has very low odd and higher harmonics. Major harmonics is 2nd. It is a lot easier to deal with. SS is like pentode that has high harmonics way to 7th order. Seems like from my limited knowledge, each has strong or weak points, nobody whip the floor with the other. Remember even in SS amp, when there is higher harmonics, the harmonics of 5KHz can extend to 50KHz. At 50KHz, the loop gain is going to be like 6dB to 10dB for an amp with 100KHz BW. It's not effective in lowering 50KHz harmonics.
Last edited:
As Johan was saying about OT distortion, I would agree. Magnetizing current is the main distortion culprit there, and that appears on the primary side where it affects tube linearity. (and is a significant issue mainly for SE amplifiers anyway) This can be fixed readily by local feedback to a driver stage with sufficient gain. There is still winding resistance and frequency roll-off to consider, but these can be fixed acceptably with a modicum of global feedback.
I think the usual complaints about OT distortion derive from traditional amplifiers where only global feedback was used, so limited loop gain would be available for correction in that case to get super low distortion figures now. Global feedback would have been fine for the typical 0.1 % distortion goal for Hi-Fi back then, but would be an issue trying to reach PPM distortion levels of extreme SS amps today. (but local feedback can still be called to fix that, it just wasn't much needed traditionally.)
The H-K Citation II used internal local feedbacks (beside global FDBK) , as did the Mac, ElectroVoice, Quad, A-R and certainly a few others.
Now that many try to get away with some cheapo OT (power toroids) for cost reasons (DIY), we have been much more attentive about local feedbacks.
Per Alan's concern about competing with SS on dist., this could possibly be done in a tube amp using a lot of nested feedbacks and more gain stages. The output tubes are the main distortion issue, and there are super techniques like Error Correction (Hawksford) and differential current FDBK that could be applied there. (as in SS amps) (these don't actually require much additional circuitry, and could be done using SS circuits anyway, if reliability is a concern.) One could also go Hog-Wild using a ton of parallel tubes and even class A, with the consequent heat output be da_ned. (OTL amp owners have already done that..., must be in the air conditioning business.)
The issue most tube-aholics would have with that, is the amp just starts sounding very clinical, and becomes indistinguishable from a Best Buy amp. The view from the tube world is that SS designers have run out of meaningful problems to solve, so are just pushing the decimal place on dist. specs further and further beyond what is audible. Nothing wrong with pushing the state of the art for design, but a big part of the attraction of tube amps is the very limited complexity needed to reach acceptable sound. (And the improved reliability when limited numbers of tubes are used. Plus tubes show aging effects on their gm, so maintaining a high performance tube amp would be difficult. Although there are some SS/CPU monitoring schemes around.)
I think the usual complaints about OT distortion derive from traditional amplifiers where only global feedback was used, so limited loop gain would be available for correction in that case to get super low distortion figures now. Global feedback would have been fine for the typical 0.1 % distortion goal for Hi-Fi back then, but would be an issue trying to reach PPM distortion levels of extreme SS amps today. (but local feedback can still be called to fix that, it just wasn't much needed traditionally.)
The H-K Citation II used internal local feedbacks (beside global FDBK) , as did the Mac, ElectroVoice, Quad, A-R and certainly a few others.
Now that many try to get away with some cheapo OT (power toroids) for cost reasons (DIY), we have been much more attentive about local feedbacks.
Per Alan's concern about competing with SS on dist., this could possibly be done in a tube amp using a lot of nested feedbacks and more gain stages. The output tubes are the main distortion issue, and there are super techniques like Error Correction (Hawksford) and differential current FDBK that could be applied there. (as in SS amps) (these don't actually require much additional circuitry, and could be done using SS circuits anyway, if reliability is a concern.) One could also go Hog-Wild using a ton of parallel tubes and even class A, with the consequent heat output be da_ned. (OTL amp owners have already done that..., must be in the air conditioning business.)
The issue most tube-aholics would have with that, is the amp just starts sounding very clinical, and becomes indistinguishable from a Best Buy amp. The view from the tube world is that SS designers have run out of meaningful problems to solve, so are just pushing the decimal place on dist. specs further and further beyond what is audible. Nothing wrong with pushing the state of the art for design, but a big part of the attraction of tube amps is the very limited complexity needed to reach acceptable sound. (And the improved reliability when limited numbers of tubes are used. Plus tubes show aging effects on their gm, so maintaining a high performance tube amp would be difficult. Although there are some SS/CPU monitoring schemes around.)
Last edited:
OH, and power supply issues are another big distortion/noise issue for tube amps. But these are quite fixable with regulated supplies, attention to amp PSRR, and DC power for the filaments/heaters.
What happened to:
All amplifiers sound the same<<< (ALL that's SS or tube in whatever form or all topologies) even a mix of both.
At the same THD and distortion levels every amplifier is indistinguishable from another in blind testing?
Like I said before are you listening to the music or the equipment? IF the above statement is true then your listening to the equipment!
And this seems to be what the OP is trying to discover..the problem is if you can hear a difference its the distortion..if the level of distortion is the same you can't tell.
But then we are back to the same problem whats the point of "High end" equipment..😀 (is the distortion not low enough in midfi so you can hear a difference with high end (read expensive)?)
And if distortion is the only measure why do people go back to the older designs and will go to the trouble to build one?
In theory it seems the OP is asking can you hear a topology? (And tell one from another in a blind test) the above statement says no you can't. So next question whats the point of different topology? (other than power saving etc)
Regards
M. Gregg
All amplifiers sound the same<<< (ALL that's SS or tube in whatever form or all topologies) even a mix of both.
At the same THD and distortion levels every amplifier is indistinguishable from another in blind testing?
Like I said before are you listening to the music or the equipment? IF the above statement is true then your listening to the equipment!
And this seems to be what the OP is trying to discover..the problem is if you can hear a difference its the distortion..if the level of distortion is the same you can't tell.
But then we are back to the same problem whats the point of "High end" equipment..😀 (is the distortion not low enough in midfi so you can hear a difference with high end (read expensive)?)
And if distortion is the only measure why do people go back to the older designs and will go to the trouble to build one?
In theory it seems the OP is asking can you hear a topology? (And tell one from another in a blind test) the above statement says no you can't. So next question whats the point of different topology? (other than power saving etc)
Regards
M. Gregg
Last edited:
The distribution of harmonics (low order versus high order) would make a difference in sound for the same total dist., but beyond maybe 0.1% THD for benign low order or possibly 0.01% THD for obnoxious high order, one probably wouldn't notice much. (There are those who claim they can hear nano and un-measurable effects, I'll leave that out for now.)
One thing that should be mentioned is the soft clipping behavior using low NFBK, which is possible using tubes. There are many adherents of low or no FDBK systems for various reasons. Too many complex issues there to easily sort out all high NFDBK effects, but I think if properly done in all respects, there is really no issue with high NFDBK other than the claimed "clinical", "all sound the same" effects. Not too surprising that NO distortion amps would all sound the same, some subtle coloration missing I think.
-----------------------------
Here is a link to a long ago thread on the Wolcott amplifier, which is a high NFDBK amplifier currently made commercially. Unfortunately, the "white papers" on its design are gone these days, but patent info is still around.
http://www.diyaudio.com/forums/tubes-valves/58352-wollcott-cross-coupled-circuit-3.html#post655355
It uses positive feedback around a linearized, bootstrapped triode to obtain high loop gain for a set of nested local and global NFDBK loops. The front end complexity seems over done.... One might consider an approach similar to this for a "Blameless" tube amp. Lots of other factors to consider besides, of course. One could maybe compress the high gain stage into a cathode degenerated, frame grid pentode LTP with Gyrator loads. Some follower/buffers (could be Mosfet, "Powerdrive") then for driving the output tube grids still. Any type of high gain amp will need some attention to overload/recovery characteristics.
One thing that should be mentioned is the soft clipping behavior using low NFBK, which is possible using tubes. There are many adherents of low or no FDBK systems for various reasons. Too many complex issues there to easily sort out all high NFDBK effects, but I think if properly done in all respects, there is really no issue with high NFDBK other than the claimed "clinical", "all sound the same" effects. Not too surprising that NO distortion amps would all sound the same, some subtle coloration missing I think.
-----------------------------
Here is a link to a long ago thread on the Wolcott amplifier, which is a high NFDBK amplifier currently made commercially. Unfortunately, the "white papers" on its design are gone these days, but patent info is still around.
http://www.diyaudio.com/forums/tubes-valves/58352-wollcott-cross-coupled-circuit-3.html#post655355
It uses positive feedback around a linearized, bootstrapped triode to obtain high loop gain for a set of nested local and global NFDBK loops. The front end complexity seems over done.... One might consider an approach similar to this for a "Blameless" tube amp. Lots of other factors to consider besides, of course. One could maybe compress the high gain stage into a cathode degenerated, frame grid pentode LTP with Gyrator loads. Some follower/buffers (could be Mosfet, "Powerdrive") then for driving the output tube grids still. Any type of high gain amp will need some attention to overload/recovery characteristics.
Attachments
Last edited:
Ok here is a question,
If you were going to build a tube amp to equal or beat a Krell SS amp in performance what topology would you use..(performance meaning sound not power levels). ie to sound the same as or close to..
I'm just interested in your thoughts. Would this fit into the Blameless category?
Regards
M. Gregg
If you were going to build a tube amp to equal or beat a Krell SS amp in performance what topology would you use..(performance meaning sound not power levels). ie to sound the same as or close to..
I'm just interested in your thoughts. Would this fit into the Blameless category?
Regards
M. Gregg
I'm not real familiar with the Krell amps, but I assume it uses massive parallel outputs (low dynamic stress on each output device for lower distortion and high peak power) and some form of error correction or high NFDBK.
For a tube amp, we would want to avoid massive numbers of tubes for reliability reasons.
The OT is the first issue to consider. Only P-P need apply for the job. Low primary impedance makes for an easier to get "perfect" OT design, so I would go with a Circlotron, Mac, or possibly a MOSFET switched ferrite OT (since that might be considered cheating, we'll cross that one off for now). A toroidal, grain oriented core for the OT, with no gaps, would give high primary inductance and low leakage L. (DC servo biasing) Optimised interwinding insulation for best distributed C versus leakage L tradeoff. All windings equal resistance for the Mac version (Circlotron equal R on halves of the primary). (The internal winding R is used later for differential current sensing.)
High current capable output tubes would allow for a lower primary OT Z, so some big TV sweep tubes or DC regulator tubes would be preferred. These might not be the most linear tubes, but the non-linearities of these are low order and easily fixed by high NFDBK.
We want to avoid any sharp discontinuities in gm that would cause high order distortion, so we would go with either class A operation or class AB operation with differential current feedbacks (generally only takes 2 or 4 resistors back to the driver stage, but see below). Hawksford EC would be another option, but this requires an additional LTP stage to compute the error correction signal.
The driver stage would be a VERY high gain frame grid pentode LTP with gyrator loads and some cathode R degeneration (for linearity) and using cathode follower (or Mosfet follower) buffers after it for driving the output tube grids. (CCS loading of the followers) Since the Circlotron or Mac configuration has large CFB signals, we cannot just run the differential local CFB feedbacks to the LTP cathodes via 2 wires, so these will require some R attenuation with the prescribed cathode degeneration resistors. (CFB derived feedbacks are NOT taken to the driver grids, since that would load down the phase inverter stage, and require crossed feedbacks which would not sit well with class B through the OT.)
(an alternative would be to run the diff. CFB feedbacks to the LTP driver screens. These feedbacks would need to be amplitude scaled so as to respect the internal Mu of the pentodes during operation. --keeping the screen currents linear--) Since these require crossed feedbacks, only class A operation of the OT/outputs would be acceptable in this case.
Since the CFB feedbacks also contain the proportional output voltage signal, these will also serve as the local driver V feedbacks. (ie, local CFB in the outputs automatically AND the driver stage via the R connections) Most of the huge driver gain is consumed in the local feedback operation (reduced to internal Mu territory).
The phase splitter could be Concertina/triode (or HV Mosfet) with a leading input tube stage of sufficient gain (Gyrator loaded for high gain and linearity) to provide allowable global Neg. FDBK. (which could be considerable with the bandwidth optimised OT) Balance adjustment for P-P gains.
Global feedback from the secondary to the input tube cathode per usual design.
The power supply would be oversized, LCLC and regulated. Individual regulation of all voltages used in the amp. (could be quietized/filtered switch mode power supplys) For Circlotron B+ power supplies (two) would require dual split bobbin power xfmrs and floating regulators for low common mode capacitance. DC heater supplies. Servo bias control of the output tubes for balanced OT primary current.
Class AB output biasing would be set up so that the gm tails of the output tubes are positioned at the edges of the class A region, for smoothest gm variation out of the flat region of the gm wingspread diagram. (ie, adjust bias level with a PC FFT analyzer)
For a tube amp, we would want to avoid massive numbers of tubes for reliability reasons.
The OT is the first issue to consider. Only P-P need apply for the job. Low primary impedance makes for an easier to get "perfect" OT design, so I would go with a Circlotron, Mac, or possibly a MOSFET switched ferrite OT (since that might be considered cheating, we'll cross that one off for now). A toroidal, grain oriented core for the OT, with no gaps, would give high primary inductance and low leakage L. (DC servo biasing) Optimised interwinding insulation for best distributed C versus leakage L tradeoff. All windings equal resistance for the Mac version (Circlotron equal R on halves of the primary). (The internal winding R is used later for differential current sensing.)
High current capable output tubes would allow for a lower primary OT Z, so some big TV sweep tubes or DC regulator tubes would be preferred. These might not be the most linear tubes, but the non-linearities of these are low order and easily fixed by high NFDBK.
We want to avoid any sharp discontinuities in gm that would cause high order distortion, so we would go with either class A operation or class AB operation with differential current feedbacks (generally only takes 2 or 4 resistors back to the driver stage, but see below). Hawksford EC would be another option, but this requires an additional LTP stage to compute the error correction signal.
The driver stage would be a VERY high gain frame grid pentode LTP with gyrator loads and some cathode R degeneration (for linearity) and using cathode follower (or Mosfet follower) buffers after it for driving the output tube grids. (CCS loading of the followers) Since the Circlotron or Mac configuration has large CFB signals, we cannot just run the differential local CFB feedbacks to the LTP cathodes via 2 wires, so these will require some R attenuation with the prescribed cathode degeneration resistors. (CFB derived feedbacks are NOT taken to the driver grids, since that would load down the phase inverter stage, and require crossed feedbacks which would not sit well with class B through the OT.)
(an alternative would be to run the diff. CFB feedbacks to the LTP driver screens. These feedbacks would need to be amplitude scaled so as to respect the internal Mu of the pentodes during operation. --keeping the screen currents linear--) Since these require crossed feedbacks, only class A operation of the OT/outputs would be acceptable in this case.
Since the CFB feedbacks also contain the proportional output voltage signal, these will also serve as the local driver V feedbacks. (ie, local CFB in the outputs automatically AND the driver stage via the R connections) Most of the huge driver gain is consumed in the local feedback operation (reduced to internal Mu territory).
The phase splitter could be Concertina/triode (or HV Mosfet) with a leading input tube stage of sufficient gain (Gyrator loaded for high gain and linearity) to provide allowable global Neg. FDBK. (which could be considerable with the bandwidth optimised OT) Balance adjustment for P-P gains.
Global feedback from the secondary to the input tube cathode per usual design.
The power supply would be oversized, LCLC and regulated. Individual regulation of all voltages used in the amp. (could be quietized/filtered switch mode power supplys) For Circlotron B+ power supplies (two) would require dual split bobbin power xfmrs and floating regulators for low common mode capacitance. DC heater supplies. Servo bias control of the output tubes for balanced OT primary current.
Class AB output biasing would be set up so that the gm tails of the output tubes are positioned at the edges of the class A region, for smoothest gm variation out of the flat region of the gm wingspread diagram. (ie, adjust bias level with a PC FFT analyzer)
Last edited:
Of course the lower the power demanded from an amplifier versus the design maximum, the lower the distortion as well. So just put some super efficient speaker on the "super" amp, and you will get phenomenally low distortion figures.
So, if we can cheat a "little", why not cheat a lot.... I would add a SS current output amplifier onto the output terminals of the "super" tube V amp, to carve off some of the current load from the V amp. The V amp still sets the speaker voltage. Of course we could also design a large current output tube amp for this, but since it has no direct effect on the sound, why not just cheat and put it (SS Iout) in a little black box under the tube amp with some meters on it for the tube amp power supply.
Other more exotic tube amp topologies might include using a large bank of Beam Deflection Tubes in parallel for the outputs. These are remarkably linear when selected and de-magnetized. Like 0.1% distortion open loop, try to get that out of a power tube!! I've got a big box of 6JH8 BDTs here from a tube sale. 6ME8's are cheap too. And these BDT's can be used to make a great phase splitter too.
Or use a big bank of frame grid tubes in parallel for the tube outputs. Should work great for a CFB design. Not sure how well those fragile grid wires would hold up to a massive bass drum overload though. (8417 outputs seem to have had some reliability issues. ) Add some current limiting.
So, if we can cheat a "little", why not cheat a lot.... I would add a SS current output amplifier onto the output terminals of the "super" tube V amp, to carve off some of the current load from the V amp. The V amp still sets the speaker voltage. Of course we could also design a large current output tube amp for this, but since it has no direct effect on the sound, why not just cheat and put it (SS Iout) in a little black box under the tube amp with some meters on it for the tube amp power supply.
Other more exotic tube amp topologies might include using a large bank of Beam Deflection Tubes in parallel for the outputs. These are remarkably linear when selected and de-magnetized. Like 0.1% distortion open loop, try to get that out of a power tube!! I've got a big box of 6JH8 BDTs here from a tube sale. 6ME8's are cheap too. And these BDT's can be used to make a great phase splitter too.
Or use a big bank of frame grid tubes in parallel for the tube outputs. Should work great for a CFB design. Not sure how well those fragile grid wires would hold up to a massive bass drum overload though. (8417 outputs seem to have had some reliability issues. ) Add some current limiting.
Last edited:
If that's the criterion rather than number of decimal places in distortion, then many PP topologies could be made to work. The hard part is output impedance (and thus frequency response into a speaker load)- inaudible distortion is easy.
"The hard part is output impedance (and thus frequency response into a speaker load)- inaudible distortion is easy."
Global feedback and a large oversize OT (low winding R) should be able to do a fairly good job on the output Z at the lower frequencies for the standard Voltage out tube amp where there is plenty of NFDBK to enforce that.
The current output "cheater" or Iout amplifier can effectively lower the output Z of the V amp if it is configured to minimize current through a sense resistor in the return ground of the V amp. (essentially carving out all the load current it can, leaving a high Z load for the V amp)
A powerful SS Iout amp could do this trick nicely, but if we are to stay on the up and up, we could design a tube Iout amp using a ferrite OT with just high frequency capability and very low winding resistance. This would complement the Vamp's low freq. capability for Zout.
It would have to be coupled to the Vamp's output using a small C to avoid shunting the Vamp at low frequencies. The Iout amp's bandwidth would have to extend somewhat below the crossover frequency of the coupling cap and the Vamp Zout, so that both amps would be in harmony through the full range where they interact. So the Iamp might be something like 5 KHz and above (say 20 KHz).
Since audio typically doesn't have much power in this upper range (if it does, the listener soon goes deaf anyway), the Iout amp can be much smaller than the Vout amp (powerwise).
--------------------------------------------------------------
If this is all too complicated, then another approach is possible. A tube Iout amp does not see leakage L in the OT as long as it has enough voltage compliance to overcome the voltage drop. It does however see distributed capacitance in the primary. So the Iout amp would use a typically low leakage L toroidal OT with progressive wound windings for low distributed C. And good insulation between the primary and secondary and core, which would increase the leakage L, but minimizes the distributed C. The Iout amp covers the full bandwidth this way and uses the same current sense in the Vamp ground return for it's inverting signal input. (ie, it still tries to minimize the the current draw through the Vamp. ) This scheme however requires a large Iout amplifier to cover the full frequency range, but it lowers Zout of the Vamp through the entire range. And so the Vout amplifier can be shrunk down to a small amplifier instead.
Or one could size both amps at around 1/2 the rated power and let them both do 1/2 the work by putting a little input signal into the non-inverting input of the Iout amp.
Global feedback and a large oversize OT (low winding R) should be able to do a fairly good job on the output Z at the lower frequencies for the standard Voltage out tube amp where there is plenty of NFDBK to enforce that.
The current output "cheater" or Iout amplifier can effectively lower the output Z of the V amp if it is configured to minimize current through a sense resistor in the return ground of the V amp. (essentially carving out all the load current it can, leaving a high Z load for the V amp)
A powerful SS Iout amp could do this trick nicely, but if we are to stay on the up and up, we could design a tube Iout amp using a ferrite OT with just high frequency capability and very low winding resistance. This would complement the Vamp's low freq. capability for Zout.
It would have to be coupled to the Vamp's output using a small C to avoid shunting the Vamp at low frequencies. The Iout amp's bandwidth would have to extend somewhat below the crossover frequency of the coupling cap and the Vamp Zout, so that both amps would be in harmony through the full range where they interact. So the Iamp might be something like 5 KHz and above (say 20 KHz).
Since audio typically doesn't have much power in this upper range (if it does, the listener soon goes deaf anyway), the Iout amp can be much smaller than the Vout amp (powerwise).
--------------------------------------------------------------
If this is all too complicated, then another approach is possible. A tube Iout amp does not see leakage L in the OT as long as it has enough voltage compliance to overcome the voltage drop. It does however see distributed capacitance in the primary. So the Iout amp would use a typically low leakage L toroidal OT with progressive wound windings for low distributed C. And good insulation between the primary and secondary and core, which would increase the leakage L, but minimizes the distributed C. The Iout amp covers the full bandwidth this way and uses the same current sense in the Vamp ground return for it's inverting signal input. (ie, it still tries to minimize the the current draw through the Vamp. ) This scheme however requires a large Iout amplifier to cover the full frequency range, but it lowers Zout of the Vamp through the entire range. And so the Vout amplifier can be shrunk down to a small amplifier instead.
Or one could size both amps at around 1/2 the rated power and let them both do 1/2 the work by putting a little input signal into the non-inverting input of the Iout amp.
Last edited:
There is also a composite OT scheme in RDH-4 with low and high frequency OTs combining outputs. The low freq. OT would be the toroid OT and the high freq. OT would be the ferrite OT. Their outputs are combined with an L and C network. This enables the same (tube) amplifier to operate both OTs over the combined frequency range.
Some clever feedback scheme might enable the low Freq. OT to be operated in V mode and the high Freq. OT to operate in I mode. Or the amp could just operate conventionally (V mode) over an extended frequency range. That would allow NFDBK to operate with full force well up into 100's of KHz, if not even MHz, for low Zout over the full audio range.
Some clever feedback scheme might enable the low Freq. OT to be operated in V mode and the high Freq. OT to operate in I mode. Or the amp could just operate conventionally (V mode) over an extended frequency range. That would allow NFDBK to operate with full force well up into 100's of KHz, if not even MHz, for low Zout over the full audio range.
You need to look at the whole picture which is not just distortion at rated power but distortion as function of power (from very low to rated power).
Then you will see that the operative conditions come first and class A or high current class AB win hands down. You can see this already in data-sheets where you get a distortion number at rated power and in many cases graphs of THD as function of Pout.
Triode has the advantage of lower source impedance which is good with conventional speakers. But this is not a universal rule! For example, if you use full range drivers the larger the size of the driver the more desirable can be higher Zout. High frequency response can improve considerably. At low frequency you just need need to have a flexible combination of parameters to adjust the alignment for the actual source impedance. Although it is not hi-fi, guitar amps with large 12" cones achieve a good high frequency response and if you look at the OPS it's invariably pentode connection without fb. Switch to triode and there will be no trebles....easy to prove.
UL is more efficient. With triode-strapped EL34 in PP you can get 15W in pure class A or 20-25W in AB. "Designed for 40W" is not a good point for comparison. Supply voltage and PSU capability are more important as you can change the output transformer for pure class A or Class AB operation.
Pure Class A in triode can be achieved with approx. 400V (and no less than this)/50mA and 10K plate-to-plate. To get more power in AB you just reduce the primary load to 4-6K plate-to-plate, for example.
Crossover distortion is negligible for class A and high current AB tube amps. If not something is wrong....
As said the main advantage is distortion at low level and better (lower) Zout. But Pout will be somewhat lower. In practical applications you are not going to notice a difference in SPL between 15W and 20W. I can bet on this...
None for me.
- Status
- Not open for further replies.