What do you call that unbypassed emitter resistor again ?
Remove it, rerun your sim, then come back with the results.
I assume that you quoted the wrong paragraph and your are replying to my previous post.
The "unbypassed emitter resistor" is there to give the BJT amplifier stage the same transconductance as the tube amplifier stage. You can call it feedback if you want (you could point your finger to the internal cathode impedance of the tube too). It would be rather obvious what would happen if the sim was run without it, wouldn't it?
Not sure I see a compelling reason why a tube is necessarily a better device for building linear audio circuits here (although they look cool and I use a lot of them myself).
Last edited:
You're comparing apples and oranges and cheating while at it.
Remove that resistor and adjust (= reduce the swing of) driving voltage accordingly. If BJT is indeed more linear, you should end up with the same or better result once BJT driving voltage is adjusted to accomodate for different gm figure, right ?
I quoted that paragraph because you claimed that open loop linearity of SS amplifier is better, then mentioned "no gnfb" which means nothing (who neeeds gnfb when there's plenty of local nfb to linearize the response ?) and then presented a single-stage schematic (where by definition of single stage amplifier local nfb is also global) to "prove" your point which has nothing to do with reality.
Remove that resistor and adjust (= reduce the swing of) driving voltage accordingly. If BJT is indeed more linear, you should end up with the same or better result once BJT driving voltage is adjusted to accomodate for different gm figure, right ?
I quoted that paragraph because you claimed that open loop linearity of SS amplifier is better, then mentioned "no gnfb" which means nothing (who neeeds gnfb when there's plenty of local nfb to linearize the response ?) and then presented a single-stage schematic (where by definition of single stage amplifier local nfb is also global) to "prove" your point which has nothing to do with reality.
Last edited:
2,5 KW behringer EP2500 is $284 @ B&H
i use QSC PLX myself which is more expensive but the power is similar.
Rated at 450 w/channel, that is 0.9 kW (no you can't use the 2 ohm rating)... and even at 8 ohm doesn't meet your distortion criteria (0.02%) and you can expect that to at probably quadruple into 2 ohms.
dave
You're comparing apples and oranges and cheating while at it.
Remove that resistor and adjust (= reduce the swing of) driving voltage accordingly. If BJT is indeed more linear, you should end up with the same or better result once BJT driving voltage is adjusted to accomodate for different gm figure, right ?
I quoted that paragraph because you claimed that open loop linearity of SS amplifier is better, then mentioned "no gnfb" which means nothing (who neeeds gnfb when there's plenty of local nfb to linearize the response ?) and then presented a single-stage schematic (where by definition of single stage amplifier local nfb is also global) to "prove" your point which has nothing to do with reality.
LOL!
Your entire post doesn't make a lot of sense at all an contains the usual evasions.
The whole purpose was to demonstrate the fact that the transistor amplifier operating at the same transconductance / voltage gain and DC operating points will have much better linearity. How is this perfectly fair comparison comparing "apples with oranges"? duh.
Also, the open loop linearity of the SS amp IS better (regardless of any pious conflation/confusion between gnfb and "local" feedback).
Last edited:
GK has provided a good analysis result of a BJT circuit - inasfar as that goes. But how exactly did you drive the BJT, GK - and what happens when you drive the next stage with it?
A complete amplifier usually has more than one stage. If one starts to tandem BJT stages viz-a-viz tube stages, the picture can change significantly.
With this I am not singling out a particular topology, simply saying that ... may I put it like this ... the whole 'box of apples' one way or another must be compared for what I read into Borat's aim - not just the best single example of either topology.
To bring some modest experience into the picture: I have designed perhaps a dozen of each topology over several decades, and would not like to make it a 'competition - end of story' kind of thing. Each topology has its high points, and my take would be to mention a few 'do's' vs. a few 'dont's' of either topology. (Again, guessing that that was what Borat had in mind - I did not look back.) I certainly got immaculate results audible-wise either way, but e.g. found it harder to get rid of high-order harmonic generation in semiconductors than with tubes. Then again that does not bring the economic factor into the scenario, where semiconductors must simply have the edge. But there are sadly also enough semiconductor amplifiers on the market giving 'listener fatigue' to show that not every design has achieved the mark.
A complete amplifier usually has more than one stage. If one starts to tandem BJT stages viz-a-viz tube stages, the picture can change significantly.
With this I am not singling out a particular topology, simply saying that ... may I put it like this ... the whole 'box of apples' one way or another must be compared for what I read into Borat's aim - not just the best single example of either topology.
To bring some modest experience into the picture: I have designed perhaps a dozen of each topology over several decades, and would not like to make it a 'competition - end of story' kind of thing. Each topology has its high points, and my take would be to mention a few 'do's' vs. a few 'dont's' of either topology. (Again, guessing that that was what Borat had in mind - I did not look back.) I certainly got immaculate results audible-wise either way, but e.g. found it harder to get rid of high-order harmonic generation in semiconductors than with tubes. Then again that does not bring the economic factor into the scenario, where semiconductors must simply have the edge. But there are sadly also enough semiconductor amplifiers on the market giving 'listener fatigue' to show that not every design has achieved the mark.
And just to confuse the issue further, Gary Pimm has built a SS amp, that uses transformers and runs the SS devices at very high (for typical transistors) rail voltages. The results are actually unbelievable good, considering what "everyone knows" about transformers or SS.
Look here under SS tabor for test results and circuit.
Bud
Look here under SS tabor for test results and circuit.
Bud
Not necessarily. I haven’t had any problems making complete solid-state front-ends (input stage / intermediate voltage amplification / driver) for my valve amplifiers that have an order or magnitude or greater better linearity than any comparable tube amplifier (with predominately low order distortion).
Incidentally, since anecdotes seem to carry more weight around here than anything else, I’ve owed and listened to a number of commercial valve amplifiers that have been just as “fatiguing” as the worst solid-state amplifiers.
Here is the fft comparison for that tube Vs BJT amplifier stage I posted earlier:
Attachments
I'd say that the main 'thing' Gary has done with his SS design is eliminate the voltage variable gate capacitance. All the amplifier stages are cascode-ed. That is a significant departure from the norm. The result is a heaterless pentode w/o the troubles of g2 current.
cheers,
Douglas
Evasions ? I didn't make any preposterous claims that would require evasion - you did. You made an irrelevant picture to back up your claim. If you want to prove your point, post relevant comparison.
Nah, first of all: you are comparing two models. Which is fine, but may not necessarily reflect the reality. If you wich to demonstrate that to be true in reality, you'll have to switch to RL devices.
Second: you're introducing additional factor which linearizes response of one stage. Since you're operating with models only you might as well stick perfect transconductance amplifier (source follower) on that tube, remove emitter resistor and rerun your simulation. You'll end up with same transconductance which seems to be your goal but you won't be affecting linearity of voltage amplification of either device under simulation.
And ultimately: while I find such comparisons intriguing, I fail to see the point in them as they are more along the lines of comparing F1 race car to a Hummer rather than comparing two types of race cars. To prove your claim that (I quote) "the open loop linearity of the SS amp IS better [than that of tube amp]" you really don't need to match the working parameters. Use each device at its "sweet spot" without cheating (open loop = no additional components that ensure extra feedback) and show us the results.
The only person who appears to be confused here is you
In a single stage any feedback will by definition be as "global" as they get. It spans that entire stage and that entire amplifier. You see, I don't subscribe to dogma spouted by either side in this "holy war" of amplification elements, if there ever was such a thing there - all I care about are results. Post a simulation where neither device's response is additionaly linearized and I'll be the first to admit that yes, that particular transistor's model you're using is indeed more linear than that particular tube's model, if your results should show that. Fair enough ?
Evasions ? I didn't make any preposterous claims that would require evasion - you did. You made an irrelevant picture to back up your claim. If you want to prove your point, post relevant comparison.
Nah, first of all: you are comparing two models. Which is fine, but may not necessarily reflect the reality. If you wich to demonstrate that to be true in reality, you'll have to switch to RL devices.
Second: you're introducing additional factor which linearizes response of one stage. Since you're operating with models only you might as well stick perfect transconductance amplifier (source follower) on that tube, remove emitter resistor and rerun your simulation. You'll end up with same transconductance which seems to be your goal but you won't be affecting linearity of voltage amplification of either device under simulation.
And ultimately: while I find such comparisons intriguing, I fail to see the point in them as they are more along the lines of comparing F1 race car to a Hummer rather than comparing two types of race cars. To prove your claim that (I quote) "the open loop linearity of the SS amp IS better [than that of tube amp]" you really don't need to match the working parameters. Use each device at its "sweet spot" without cheating (open loop = no additional components that ensure extra feedback) and show us the
results.
You see, I don't subscribe to dogma spouted by either side in this "holy war" of amplification elements, if there ever was such a thing there - all I care about are results. Post a simulation where neither device's response is additionaly linearized and I'll be the first to admit that yes, that particular transistor's model you're using is indeed more linear than that particular tube's model, if your results should show that. Fair enough ?
Amazing. None of the claims I have made are even remotely preposterous and this reply of yours really is beyond a joke; and you still don't get it.
I'm not saying that a transistor is a more linear device than a triode; it most definitely is not. This isn't even up for argument.
And of course I'm introducing an additional component to linearise the response of the transistor amplifier, which was the whole point! A transistor is a high transconductance device (far higher that a typical vacuum tube), and that gives scope for high degree of linearization by means of emitter degeneration (which IS a form of feedback). This is basic engineering.
To reiterate for the second time; that sim I posted was to demonstrate this fact. The basic transistor amplifier (emitter degenerated back to the same transconductance as the tube amplifier, for an apples-apples comparison) is much more linear.
The claim that a triode is a better device for "building linear audio circuits", based on the fact that it is an intrinsically more linear device is quasi-technical hogwash.
The only person who appears to be confused here is you
I was referring to that other (non-single stage amplifer) under discussion.
Last edited:
So if open loop of SS amplifier is better and transistor is not more linear device in open loop configuration I take you're acquainted with some other kind of solid-state audio amplification element the rest of us haven't heard of yet. Or is it "open loop" only if there is no visible connection from the collector to the base ?
Yup, you have indeed proven that once enough negative feedback is applied to your transistor model it behaves in a more linear fashion than your tube model when driving identical ohmic load. Nothing to argue about here.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.