Re: PP TUBES
Hi,
> Originally posted by fdegrove
> SE is ok for mid and highs but can't take control of a woofie.
It does fine in my place with a reflex loaded 15" Woofer for control.
So, SE CAN "take control" of a woofie. Blame the implementation of the system, not the principle.
Sayonara.
Hi,
> Originally posted by fdegrove
> SE is ok for mid and highs but can't take control of a woofie.
It does fine in my place with a reflex loaded 15" Woofer for control.
So, SE CAN "take control" of a woofie. Blame the implementation of the system, not the principle.
Sayonara.
Re: Re: Re: Re: PSE
Thorsten,
I have a vague idea how you would wire them in series, could you whip up a sketch?
dave
Kuei Yang Wang said:
Thorsten,
I have a vague idea how you would wire them in series, could you whip up a sketch?
dave
PSE
Hi John,
What happens when one paralells tubes?
The reported statements are correct but can be overcome.
Advantages of paralelled tubes are higher power output,lower impedance hence greater bandwidth with the same quality of OPT.
Ideally these two tubes should have independent PSU,and bias.
Nobody seems to do this,probably due to high cost.
The perceived loss of microdetail is due to the fact that paralelled tubes are often mismatched and therefore never reach stability giving a sense of restlessness.
In PP amps this is not such a problem since the PP system will cancel most of this.
This is not say the tubes do not have to be matched but the reason for matching is different.I.e. they often share a common bias.
IMO it would be wiser to build a PP amp using the same tubes than to build a PSE.
Fine,it CAN be done and if the woofer does not need the extra damping and control a PP offers then that's O.K. with me.
The fact remains however that a PP amp intrinsically offers greater damping factor then any SE amp.
Knowing how to implement technology for a task is something that is all too often overlooked.
Cheers,😉
Hi John,
What happens when one paralells tubes?
The reported statements are correct but can be overcome.
Advantages of paralelled tubes are higher power output,lower impedance hence greater bandwidth with the same quality of OPT.
Ideally these two tubes should have independent PSU,and bias.
Nobody seems to do this,probably due to high cost.
The perceived loss of microdetail is due to the fact that paralelled tubes are often mismatched and therefore never reach stability giving a sense of restlessness.
In PP amps this is not such a problem since the PP system will cancel most of this.
This is not say the tubes do not have to be matched but the reason for matching is different.I.e. they often share a common bias.
IMO it would be wiser to build a PP amp using the same tubes than to build a PSE.
Fine,it CAN be done and if the woofer does not need the extra damping and control a PP offers then that's O.K. with me.
The fact remains however that a PP amp intrinsically offers greater damping factor then any SE amp.
Knowing how to implement technology for a task is something that is all too often overlooked.
Cheers,😉
PSE v PP
Frank, thanks for your comments. I agree with much of what you say but:
Might a reason for the perceived difference between PSE and PP be the feedback normall applied in the PP circuit?
Perhaps the detail will re-appear in the PSE circuit if the same feedback is supplied to "straighten out" the transfer characteristic.
Frank, thanks for your comments. I agree with much of what you say but:
So if they're matched it's OK?
Surely it will only cancel the non-linearity (where the curve is 2nd order polynomial). So matching must be equally important?
Might a reason for the perceived difference between PSE and PP be the feedback normall applied in the PP circuit?
Perhaps the detail will re-appear in the PSE circuit if the same feedback is supplied to "straighten out" the transfer characteristic.
PSE
Hello,
Should be fine.
Correct.
Not just the feedback.
It is best to compare amps open loop without applied NFB.
PP amps suffer from more complexity due to the need for phase splitter circuits.
No amount of feedback can bring back loss of detail in a previous stage.
Cheers,😉
Hello,
Should be fine.
Correct.
Not just the feedback.
It is best to compare amps open loop without applied NFB.
PP amps suffer from more complexity due to the need for phase splitter circuits.
No amount of feedback can bring back loss of detail in a previous stage.
Cheers,😉
Frank,
From the evidence presented and surmised, I cannot see how there can be any difference between properly designed and matched PSE and PP in terms of detail.
Only the distortion signatures will differ.
But I'm happy to argue on if someone feels differently....
From the evidence presented and surmised, I cannot see how there can be any difference between properly designed and matched PSE and PP in terms of detail.
Only the distortion signatures will differ.
But I'm happy to argue on if someone feels differently....
DETAIL.
Hi John,
The less complex the design the shorter the signal path will be.
This will have inherently a better chance of reveiling micro details in the musical signal.
One of the BIG advantages of DIY is that you have much better control over the design than any commercial project.
You know what speakers this is going to be used with and you can then calculate what amount of gain and power is needed.
Most commercial products are used with a lot of gain dumped resistively which is bound to dump a fair bit of detail and dynamic range along with it.
Cheers,😉
Hi John,
The less complex the design the shorter the signal path will be.
This will have inherently a better chance of reveiling micro details in the musical signal.
One of the BIG advantages of DIY is that you have much better control over the design than any commercial project.
You know what speakers this is going to be used with and you can then calculate what amount of gain and power is needed.
Most commercial products are used with a lot of gain dumped resistively which is bound to dump a fair bit of detail and dynamic range along with it.
Cheers,😉
Re: PSE
Hi,
Originally posted by fdegrove:
> Fine,it CAN be done and if the woofer does not need the extra
> damping and control a PP offers
This statement seems to me to indicate a certain lack of understanding electrical damping in conventional Drivers. The ultimate electrical damping is limited by the DCR of the voice coil.
The difference in real "damping" between an Amplifier with a damping factor of say 1000, one with a damping factor of 10 and with a damping factor of 2.7 (typhical SE Amp), or more sensibly rendered, an output impedance of 0.008 Ohm, 0.8 Ohm and 3 Ohm is very modest.
If we have a semi-decent 10" Driver (for arguments sake the Seas CA25FEY) with Qms 1.83, Qes 0.33, Qts 0.31, Fs 31Hz, Re 5.5 Ohm and Vas 150 Liter mounted in a 80 Liter enclosure tuned to 40Hz.
Now we check the change of Qe for the different source resistances (conveniently ignoring any corssovers and cable resistances). For 0.008 Ohm the Qe basically stays at 0.33, for 0.8 Ohm source impedance the Qe becomes 0.38, for 3 ohm source resistance it becomes 0.51, qhiving a Qt respectively of 0.28, 0.315 and 0.4.
Keeping the box exactly equal for all three amplifiers the -3db point remains unaffected and the overall difference in output is greatest at around 60Hz, where we have respectively +0, +1 and +3db with the whole response flat again at 200Hz.
While such changes will likely be audible, in most cases moving the speakers a few inches will result in a much larger degree of change in frequency response.
Another way of looking at this to calculate the REAL electrical damping factor on the cone compared to the nominal 8 ohm impedance. If we include (as we must) the Voicecoil DCR our actual damping factor is 1.6 with the Amplifier having a nominal damping factor of 1000, it is 1.27 for the Amplifier with a damping factor of 10 and 0.94 for the SE amplifier.
> The fact remains however that a PP amp intrinsically offers
> greater damping factor then any SE amp.
I am not sure how you derive such a conclusion. Lets assume for calculation a 2A3 Class A1 Amplifier with a pair of 2A3's, each having an anode impedance of 800 Ohm. In one case we use a PP Circuit with a 5K A-A impedance. In the other we use a SE transformer with 1k25 primary impedance. In both cases the transformers have a voice coil nominal impedance of 8 Ohm.
Our PP Amplifier drives it's output transformer effectively from 1600 Ohm (both Anodes in series for AC) and thus attains an output Impedance of 2.56 Ohm. Our SE Amplifier drives it's output transformer effectively from 400 Ohm (both Anodes in parallel for AC) and thus attains an output Impedance of 2.56 Ohm.
As we can see, the output impedance of the same Valve complement operated in SE and PP under otherwise identical conditions is ABSOLUTELY EQUAL. It should be noted that for Class A also the attained power is Equal for an equal amount of 3rd harmonics. The PSE stage will show a higher overall THD and a much higher level of even harmonics, however, as sually the audibility of distortion is dominated by the odd order harmonics we can actually conclude that if we make sure of a level playing field the same Valves in PSE and PP offer the same amount of power for the same audibility of distortion.
So, if we rigerously apply "all else being equal" we find that for damping factor and power output for a given degree of audibility of distortion in the conventional senses there is NO difference between PP and PSE.
There are only two advantages for PP over PSE in the conventional sense, namely a lower measured THD which looks good on paper and a much easier to make and much less material consuming output transformer, which reduces production cost. In terms of power, audible distortion and damping factor neither approach has any intrinsic advantages.
In terms of listening, based on an amplifier that was switchable between PP and PSE my subjective preference was for PSE, with surprisingly the largest improvement occouring in low frequencies, which became more controlled, resolved and detailed, easier to follow.
> Knowing how to implement technology for a task is something
> that is all too often overlooked.
Exactly.
Sayonara.
Hi,
Originally posted by fdegrove:
> Fine,it CAN be done and if the woofer does not need the extra
> damping and control a PP offers
This statement seems to me to indicate a certain lack of understanding electrical damping in conventional Drivers. The ultimate electrical damping is limited by the DCR of the voice coil.
The difference in real "damping" between an Amplifier with a damping factor of say 1000, one with a damping factor of 10 and with a damping factor of 2.7 (typhical SE Amp), or more sensibly rendered, an output impedance of 0.008 Ohm, 0.8 Ohm and 3 Ohm is very modest.
If we have a semi-decent 10" Driver (for arguments sake the Seas CA25FEY) with Qms 1.83, Qes 0.33, Qts 0.31, Fs 31Hz, Re 5.5 Ohm and Vas 150 Liter mounted in a 80 Liter enclosure tuned to 40Hz.
Now we check the change of Qe for the different source resistances (conveniently ignoring any corssovers and cable resistances). For 0.008 Ohm the Qe basically stays at 0.33, for 0.8 Ohm source impedance the Qe becomes 0.38, for 3 ohm source resistance it becomes 0.51, qhiving a Qt respectively of 0.28, 0.315 and 0.4.
Keeping the box exactly equal for all three amplifiers the -3db point remains unaffected and the overall difference in output is greatest at around 60Hz, where we have respectively +0, +1 and +3db with the whole response flat again at 200Hz.
While such changes will likely be audible, in most cases moving the speakers a few inches will result in a much larger degree of change in frequency response.
Another way of looking at this to calculate the REAL electrical damping factor on the cone compared to the nominal 8 ohm impedance. If we include (as we must) the Voicecoil DCR our actual damping factor is 1.6 with the Amplifier having a nominal damping factor of 1000, it is 1.27 for the Amplifier with a damping factor of 10 and 0.94 for the SE amplifier.
> The fact remains however that a PP amp intrinsically offers
> greater damping factor then any SE amp.
I am not sure how you derive such a conclusion. Lets assume for calculation a 2A3 Class A1 Amplifier with a pair of 2A3's, each having an anode impedance of 800 Ohm. In one case we use a PP Circuit with a 5K A-A impedance. In the other we use a SE transformer with 1k25 primary impedance. In both cases the transformers have a voice coil nominal impedance of 8 Ohm.
Our PP Amplifier drives it's output transformer effectively from 1600 Ohm (both Anodes in series for AC) and thus attains an output Impedance of 2.56 Ohm. Our SE Amplifier drives it's output transformer effectively from 400 Ohm (both Anodes in parallel for AC) and thus attains an output Impedance of 2.56 Ohm.
As we can see, the output impedance of the same Valve complement operated in SE and PP under otherwise identical conditions is ABSOLUTELY EQUAL. It should be noted that for Class A also the attained power is Equal for an equal amount of 3rd harmonics. The PSE stage will show a higher overall THD and a much higher level of even harmonics, however, as sually the audibility of distortion is dominated by the odd order harmonics we can actually conclude that if we make sure of a level playing field the same Valves in PSE and PP offer the same amount of power for the same audibility of distortion.
So, if we rigerously apply "all else being equal" we find that for damping factor and power output for a given degree of audibility of distortion in the conventional senses there is NO difference between PP and PSE.
There are only two advantages for PP over PSE in the conventional sense, namely a lower measured THD which looks good on paper and a much easier to make and much less material consuming output transformer, which reduces production cost. In terms of power, audible distortion and damping factor neither approach has any intrinsic advantages.
In terms of listening, based on an amplifier that was switchable between PP and PSE my subjective preference was for PSE, with surprisingly the largest improvement occouring in low frequencies, which became more controlled, resolved and detailed, easier to follow.
> Knowing how to implement technology for a task is something
> that is all too often overlooked.
Exactly.
Sayonara.
Originally posted by Kuei Yang Wang:
Are you saying that the voice coil resistance and the amplifiers source resistance should be considered as being in series when estimating the damping factor?.... deja vu
Quoting dhaen
"Are you saying that the voice coil resistance and the amplifiers source resistance should be considered as being in series when estimating the damping factor?.... deja vu"
Yep. Don't forget to add the resistance of the crossover coils and speaker leads in series too. Oh ya, not forget to account for these when doing your speaker boxes.
Later
Bruce
"Are you saying that the voice coil resistance and the amplifiers source resistance should be considered as being in series when estimating the damping factor?.... deja vu"
Yep. Don't forget to add the resistance of the crossover coils and speaker leads in series too. Oh ya, not forget to account for these when doing your speaker boxes.
Later
Bruce
BUT OF COURSE.
Hi,
Sure enough these are in series.
What is more important for the damping factor is that the output impedance is reflected back through the transformer.
So low Zo means higher damping factor.
Most PP amplifiers use NFB loops to reduce Zo even further.
This is however NOT a microdetail.😀
Cheers,😉
Hi,
Sure enough these are in series.
What is more important for the damping factor is that the output impedance is reflected back through the transformer.
So low Zo means higher damping factor.
Most PP amplifiers use NFB loops to reduce Zo even further.
This is however NOT a microdetail.😀
Cheers,😉
So if they're in series, increasing the damping factor of the amplifer will be process of rapidly diminishing returns, since even on a poor amplifier the output R or Z is lower than 8 ohnms?
So you could use current feedback and s*d the output Z!
I'm still listening
So you could use current feedback and s*d the output Z!
I'm still listening
OUTPUT Z
John,
Hope you are still listening.
The OPT will transform the high output z into a low 8 Ohm (if you take that winding) for driving the speaker.
However the reflected (looking back) impedance is NOT 8 Ohm.
Depending on output configuration of the powerstage it will vary and if NFB from the secondary on the OPT is not applied it will be at least a number of times higher.
To simplify: suppose one would take the 16 Ohm secondary tap iso the 8 Ohm tap one would (among the power loss) notice a loss of control of the woofer due to a changed Zo.
This is especially important for OTL amps and SS designs where the feedbackloop is used to reduce the output Z and effectively includes the speaker with it.
Doing so increases damping factor and linearizes the LS (woof-woof) behaviour.
Ciao, 😉
John,
Hope you are still listening.
The OPT will transform the high output z into a low 8 Ohm (if you take that winding) for driving the speaker.
However the reflected (looking back) impedance is NOT 8 Ohm.
Depending on output configuration of the powerstage it will vary and if NFB from the secondary on the OPT is not applied it will be at least a number of times higher.
To simplify: suppose one would take the 16 Ohm secondary tap iso the 8 Ohm tap one would (among the power loss) notice a loss of control of the woofer due to a changed Zo.
This is especially important for OTL amps and SS designs where the feedbackloop is used to reduce the output Z and effectively includes the speaker with it.
Doing so increases damping factor and linearizes the LS (woof-woof) behaviour.
Ciao, 😉
Listening Sir!
Yes, yes, but:
This is drifting towards the "servo amp & piston" school.
Perhaps I should scrap my SET and buy the biggest solid state amp with uOhm o/p Z I can?
I think not... theres more to this...
Yes, yes, but:
This is drifting towards the "servo amp & piston" school.
Perhaps I should scrap my SET and buy the biggest solid state amp with uOhm o/p Z I can?
I think not... theres more to this...
Kuei Yang Wang,
Thanks for the informative post! There was a lot to chew on. Although, I disagree with the statement that THD will go up in a parallel scheme. Assuming a reasonably close pair ("... normal manufacturing tolerances have negligible effect."), the literature does not support that statement.
I guess we could simply measure it and see. Now, where'd I put my Eico...
Are you calling odd-order distortion "audible distortion"? I like that.
It does often seem that 2nd order is inaudible, even up to absurd levels. Maybe "audible distortion", could become the only THD measurement worth taking.
Frank, I'm still eagerly awaiting a description of "micro details". I want to make sure they are not falling out of my amp.
Thanks for the informative post! There was a lot to chew on. Although, I disagree with the statement that THD will go up in a parallel scheme. Assuming a reasonably close pair ("... normal manufacturing tolerances have negligible effect."), the literature does not support that statement.
I guess we could simply measure it and see. Now, where'd I put my Eico...
Are you calling odd-order distortion "audible distortion"? I like that.
It does often seem that 2nd order is inaudible, even up to absurd levels. Maybe "audible distortion", could become the only THD measurement worth taking.Frank, I'm still eagerly awaiting a description of "micro details". I want to make sure they are not falling out of my amp.
FIRST COME FIRST SERVE.
John,
You could say that.
You could also blame Harols S. Black for inventing NFB.
And where does this microOhm output Z come from?
When you roll your own you can optimize linearity and output impedance.
I'll be the last person to defend the much overused feedback loops but if you know that the linearity of triodes stemms from them having internal feedback built right in you start to look at it a bit differently.
Peter Gabriel said it all along: "Don't give up"
Caio,😉
P.S. My OTL tubeamps have an output Z of 0.14 Ohm after feedback is applied.
Without it this would be roughly 14 Ohm,you can spot the mismatch a mile away,no?
John,
You could say that.
You could also blame Harols S. Black for inventing NFB.
And where does this microOhm output Z come from?
When you roll your own you can optimize linearity and output impedance.
I'll be the last person to defend the much overused feedback loops but if you know that the linearity of triodes stemms from them having internal feedback built right in you start to look at it a bit differently.
Peter Gabriel said it all along: "Don't give up"
Caio,😉
P.S. My OTL tubeamps have an output Z of 0.14 Ohm after feedback is applied.
Without it this would be roughly 14 Ohm,you can spot the mismatch a mile away,no?
I HATE TO WAIT.
Hi Joel,
There is no definition of "micro details" IMO.
What is often described as such are small level signals present on some better recordings.
Hall or studio acoustics for instance could fall in that category.
Details that tell you about the recording venue being different from one recording to another.
Odd order distortion is very unpleasant on the human ear.
Tube amps tend to produce more even order harmonic distortion
which is still distortion but perceived as more related to the musical overtones as produced by acoustic instruments and more "pleasant" to the ear.
This one reason a lot of people prefer tubes over transistors.
The latter have a tendency to produce odd order distortion when clipping.
You could build a finely mazed net to catch them??😀
Caio,😉
Hi Joel,
There is no definition of "micro details" IMO.
What is often described as such are small level signals present on some better recordings.
Hall or studio acoustics for instance could fall in that category.
Details that tell you about the recording venue being different from one recording to another.
Odd order distortion is very unpleasant on the human ear.
Tube amps tend to produce more even order harmonic distortion
which is still distortion but perceived as more related to the musical overtones as produced by acoustic instruments and more "pleasant" to the ear.
This one reason a lot of people prefer tubes over transistors.
The latter have a tendency to produce odd order distortion when clipping.
You could build a finely mazed net to catch them??😀
Caio,😉
- Status
- Not open for further replies.