Yes, ultimately. But it's better to test everything individually in a "vacuum" first before working your way up to adding more complex variables, so you know where your results are coming from.
Tubes are not well suited to driving current into loads so they will act more non-linear when asked to do so unbuffered. But in my experience this only leads to "less good" sound rather than "euphonic" sound.
Last edited:
I think you are mixing up output impedance and optimal load impedance here; for a circuit designed to work under large-signal conditions, like an audio power amplifier, they are not necessarily the same. That is, your 4 ohm tap will be able to supply most power to a 4 ohm load, but the output impedance measured at the 4 ohm tap can range from well below 4 ohm for an amplifier with lots of shunt feedback at its output to many times higher than 4 ohm for an open-loop pentode output stage, or an output stage with series feedback.
Hopefully the designer did that. In the end what matters is what the system sounds like. As t listening, what you do on HPs tells you little about its real world performance.
dave
and i thought that pentodes are being connected as triodes becuase triodes are supposed to sound superior to pentodes...this has been hotly debated...
this i understand and has come to a realization of sorts..
witt ss amps even more so, with the sub 1 ohm output impedance, speaker impedance became less important..
no hints of distortions even a loud volumes, the amp actually became less hot.. i will try the parallel case...
i am just now discovering NP amps, will learn soon enough...
my skepticism were borne out of the fact that in college, we were taught about circuit analysis using the Thevenin's and Norton's Theorems and that there is an equivalency between the two...
Yes, in principle you can make a pretty good, but hopelessly inefficient, current drive amplifier by making an amplifier with a huge gain and huge output voltage range and inserting a series resistor with a value much greater than the loudspeaker's impedance between the amplifier and the loudspeaker.
Given that ceraations.tin aspects of the design can be optimimum for either triode or Pentode or UL (including different OPTs), the amps with switches will likely be optimum for max imum one of the configuar.
A true tridoe is typically more linear and the behaviour of a triode strapped pentode/tetrode is much closer to looking like a real triode.
Execution of the amp can easily trump that.
Some triodes sound really good in triode (ie EL84). My EL84 triode class a PP sounds stellar but has limited use with its just over 3 w output. Class AB triode would about double that, and some got more (Dyna SCA35) clain 17.5 watt, and Mojeski (RIP) did a PP EL84 with very high B+ that claimed 35W plus.
Re: Power ratings... - Roger A. Modjeski - Tubes Asylum
dave
I don't think speakers are designed for voltage sources, they are designed expecting voltage sources.. in other words, the easy way. Does this mean that if we want to use a higher Zo amp we need to conjugate the impedance? No, we could also buffer some EQ in to compensate. This is definitely outside of the purview of the non-DIY speaker designer who draws the line at the speaker cable.
ESA covers a lot of that stuff, including supporting math that may be over some heads (mine for instance, i remember just enuff to know it is solid work. He has specifc examples of current amplifiers, XOs for currant amp speakers, EQ, how to push speakers not made for the application towards suitable.
If you want to use a current amp in the bass, the resonant peak needs to be tamed or be complentary to the bass rolloff (as designed using voltage source). An apriodic enclosure really helps, one wants to start with a driver with low Qm. Or whatever Joe did with his XOs to flatten the impedance of the Elsinore Mk6. I expect the cancelation in an OB could be useful (just look at the change in Qt with the specific output impedance of your amp)
XOs are the other problem area almost always introduce bumps in the impedance, often huge. Most FRs or tweeters with copper pole pieces should have little HF impedance variation. If it rises in just the right place you might extend the top (ie Fostex FE127e).
It is very interesting to have a variable TransAmp to play with,one can listen to exactly the changes in the speakers response as the amps output impedance changes. We tried a bunch of different speakers and each had an optimum position (turned to Rout set to very low for typical speakers w XOs).
When Nelson released the F1 he wrote a short article on his appraoch to taming the impedance interaction.
http://www.firstwatt.com/pdf/art_cs_amps.pdf
and one on XOs.
http://www.firstwatt.com/pdf/art_cs_xvrs.pdf
dave
If you want to use a current amp in the bass, the resonant peak needs to be tamed or be complentary to the bass rolloff (as designed using voltage source). An apriodic enclosure really helps, one wants to start with a driver with low Qm. Or whatever Joe did with his XOs to flatten the impedance of the Elsinore Mk6. I expect the cancelation in an OB could be useful (just look at the change in Qt with the specific output impedance of your amp)
XOs are the other problem area almost always introduce bumps in the impedance, often huge. Most FRs or tweeters with copper pole pieces should have little HF impedance variation. If it rises in just the right place you might extend the top (ie Fostex FE127e).
It is very interesting to have a variable TransAmp to play with,one can listen to exactly the changes in the speakers response as the amps output impedance changes. We tried a bunch of different speakers and each had an optimum position (turned to Rout set to very low for typical speakers w XOs).
When Nelson released the F1 he wrote a short article on his appraoch to taming the impedance interaction.
http://www.firstwatt.com/pdf/art_cs_amps.pdf
and one on XOs.
http://www.firstwatt.com/pdf/art_cs_xvrs.pdf
dave
No. In a current amp, the current is fixed and the voltage is variable, in a voltag eamp the voltage is constant and the curren changes. The common ones that near the divide (coming from from the voltage side, ie typical SETs, F2).
If one wants a current amp, making one out of an LM3875 (or 3886 i guess) is pretty simple. Joe Rasmussen has a long thread, there are specific comments from Duo (aka Daniel). He turned a diyAudio/Chipamp.com/Audio Sector LM3875 board & PS, stuffin git into a fairly small Hammond box. Very enlighting to play with. Downside of the simple way is that, like many Class D amps, the feedback has to be turned for a specific load — althou i had no complaints with the LM3875 one set for 4Ω and driving 8Ω. A variable Transamp puts a potentiometer into the FB look (maybe needs some support circuitry too).
dave
In this case, however, the damping path is restored by the conjugate circuit. The speaker is no longer receiving constant current from a current amp. For these reasons it can no longer qualify as a current driven speaker when powered from a current amp.
The point is that the relations between current and voltage, current and sound pressure and voltage and sound pressure are not entirely linear in a loudspeaker. The relation between current and sound pressure is less non-linear than the relation between voltage and sound pressure, so when you force a certain signal current through the loudspeaker and let the loudspeaker determine what the voltage will be, you get less (non-linear) distortion than when you do it the other way around.
Unfortunately you also get less damping of the fundamental resonance, so you need to find another way to deal with that. Various methods to do that have been used, such as acoustic damping cloths in old valve radios, motional feedback and equalization.
- Status
- Not open for further replies.