Graham, should you feel like "bursting into flames" as a result of reading my previous post, please wait a little... I have spent some time re-reading the whole thread, and I have found answers to some of my questions, and thus a better understanding of your descriptions of how the T-bass works. Once I have thought it through and placed it all in context, I'll try again.
One remaining sticking point, however, is the "elephant in the room" - the transformer. You appear to say that it provides current boost to the driver without the phase shift that would be caused by using a LP filter. The problem is, a transformer does provide a significant phase shift, especially at frequencies as low as 20 Hz, and requires many hundreds of henries of inductance to minimise this. In turn, this would appear to swamp the inductance of the coil (6.4mH in your original schematic) in series with the transformer. More thought needed...
Regards,
Don.
One remaining sticking point, however, is the "elephant in the room" - the transformer. You appear to say that it provides current boost to the driver without the phase shift that would be caused by using a LP filter. The problem is, a transformer does provide a significant phase shift, especially at frequencies as low as 20 Hz, and requires many hundreds of henries of inductance to minimise this. In turn, this would appear to swamp the inductance of the coil (6.4mH in your original schematic) in series with the transformer. More thought needed...
Regards,
Don.
Well, that was dumb. I misread my own notes from 40 years ago, when I was wrestling with the problem of overall NFB around a large tube amplifier. The comment above applies to the primary side of a tube amp output transformer.
A transformer for T-bass use should be adequately sized to minimise phase shift and to allow for generous wire size for low resistance.
I'll continue working to try and build an accurate model.
The components used for T-bass are large and expensive, which is a deterrent to the experimental approach to tuning. My aim is to come up with a model that will allow people wishing to try a T-bass to come up with ballpark component values to suit their intended drivers before laying out cash.
Hi Don,
E-mail notification only just come through.
No the transformer provides voltage step-up.
It is the LS back-EMF induced phase angle (reactance) which draws the current.
Core saturation is more like to be a problem than phase shift at 20Hz ?
I do so agree with you that the components are large and expensive.
I have tried several times to develop the similar response without same costs, but without success.
E-mail notification only just come through.
No the transformer provides voltage step-up.
It is the LS back-EMF induced phase angle (reactance) which draws the current.
Core saturation is more like to be a problem than phase shift at 20Hz ?
I do so agree with you that the components are large and expensive.
I have tried several times to develop the similar response without same costs, but without success.
Graham, why not trying to use an "impedance multiplier" to drastically rise LS impedance and reduce power levels?
That is, something like this: http://www.diyaudio.com/forums/chip-amps/55925-lm3886-stasis-power-buffer.html#post2114281
(there it's done with a chip, but with some mods you can use just about any NFB amplifier for the purpose).
Then you could put the T-Bass at high impedance & low power levels, in between the "front-end" amplifier and the impedance multiplier.
That is, something like this: http://www.diyaudio.com/forums/chip-amps/55925-lm3886-stasis-power-buffer.html#post2114281
(there it's done with a chip, but with some mods you can use just about any NFB amplifier for the purpose).
Then you could put the T-Bass at high impedance & low power levels, in between the "front-end" amplifier and the impedance multiplier.
Hi UnixMan,
The output impedance of the T-bass varies during music time according to the music waveform applied - exactly as back-EMF from a loudspeaker causes the loudspeaker impedance to vary during music time according to the music waveform applied - only via the T-bass the action with respect to the amplifier is oppositely phased - thus negating the effects of loudspeaker system induced back-EMF which otherwise remains trapped within and re-radiated by the loudspeaker - during loudspeaker constant time - with respect to the normally ultra-low output impedance of a voltage amplifier.
A filter in front of a NFB amplifier cannot counter dynamic loudspeaker system trapped energy from re-radiating during loudspeaker system constant controlled time. This applies to OB especially, where often LF driver Qts is allowed to boost (and colour) the lower frequencies.
(I'll drink to that comment..... Taste hits dynamically too..... Cheers!)
The output impedance of the T-bass varies during music time according to the music waveform applied - exactly as back-EMF from a loudspeaker causes the loudspeaker impedance to vary during music time according to the music waveform applied - only via the T-bass the action with respect to the amplifier is oppositely phased - thus negating the effects of loudspeaker system induced back-EMF which otherwise remains trapped within and re-radiated by the loudspeaker - during loudspeaker constant time - with respect to the normally ultra-low output impedance of a voltage amplifier.
A filter in front of a NFB amplifier cannot counter dynamic loudspeaker system trapped energy from re-radiating during loudspeaker system constant controlled time. This applies to OB especially, where often LF driver Qts is allowed to boost (and colour) the lower frequencies.
(I'll drink to that comment..... Taste hits dynamically too..... Cheers!)
Last edited:
Graham, check better my suggestion & circuit... 
that's neither an NFB voltage amplifier nor a conventional buffer or follower!
It's an "impedance multiplier". It acts sort of like a transformer (but one with power gain). Looking at its input, you do not see a (more or less) constant impedance.
On the contrary, looking at the circuit input you will see a "virtual loudspeaker" which is an exact replica of the load attached to its output, except for the module (and only that) of its impedance which is multiplied by the "gain" (current gain) of the circuit.
And its (apparent) "efficiency" (power sensitivity), which is of course multiplied by the same amount (there is fixed current gain with ~ unity voltage gain, thus a power gain ~ equal to the current gain).
That is, assuming a current gain of 1000 (as it is in my prototype), a LS which have a nominal impedance of 8ohm and a nominal sensitivity of, say, 90dB/W/m will appear at the input of the circuit as a new "virtual LS" which have a nominal impedance of 8Kohm and a sensitivity of 90dB/mW/m. Voltage and phase relationship remains exactly the same, they are not altered by the circuit (well, for the voltage there is a little attenuation introduced).
The multiplier circuit just senses input and output currents and strives to keep a fixed, costant ratio between the two. It does so "carelessly of" (ignoring, of course within limits) whatever voltage happen to be there. Thus the phase relationships between current and voltage are not altered. Whatever phase is at the output, it appears as-is also at the input (and vice-versa).
You need a "front-end", low power voltage amplifier before that, and it is only THAT one which actually "interact" with the LS complex impedance (and resonances) and "take control" of it.
If you put a properly scaled (*) version of the T-Bass circuit in between, you will get exactly the same effect as attaching the T-Bass between a conventional voltage amplifier and the LS.
(*) "properly scaled": that is, a "T-Bass" network which is designed to work properly on the new "virtual LS" with its much higher impedance.
Got it now?
that's neither an NFB voltage amplifier nor a conventional buffer or follower!
It's an "impedance multiplier". It acts sort of like a transformer (but one with power gain). Looking at its input, you do not see a (more or less) constant impedance.
On the contrary, looking at the circuit input you will see a "virtual loudspeaker" which is an exact replica of the load attached to its output, except for the module (and only that) of its impedance which is multiplied by the "gain" (current gain) of the circuit.
And its (apparent) "efficiency" (power sensitivity), which is of course multiplied by the same amount (there is fixed current gain with ~ unity voltage gain, thus a power gain ~ equal to the current gain).
That is, assuming a current gain of 1000 (as it is in my prototype), a LS which have a nominal impedance of 8ohm and a nominal sensitivity of, say, 90dB/W/m will appear at the input of the circuit as a new "virtual LS" which have a nominal impedance of 8Kohm and a sensitivity of 90dB/mW/m. Voltage and phase relationship remains exactly the same, they are not altered by the circuit (well, for the voltage there is a little attenuation introduced).
The multiplier circuit just senses input and output currents and strives to keep a fixed, costant ratio between the two. It does so "carelessly of" (ignoring, of course within limits) whatever voltage happen to be there. Thus the phase relationships between current and voltage are not altered. Whatever phase is at the output, it appears as-is also at the input (and vice-versa).
You need a "front-end", low power voltage amplifier before that, and it is only THAT one which actually "interact" with the LS complex impedance (and resonances) and "take control" of it.
If you put a properly scaled (*) version of the T-Bass circuit in between, you will get exactly the same effect as attaching the T-Bass between a conventional voltage amplifier and the LS.
(*) "properly scaled": that is, a "T-Bass" network which is designed to work properly on the new "virtual LS" with its much higher impedance.
Got it now?
Last edited:
Hi UnixMan,
Very interesting. Thanks a lot for sharing.
By < The multiplier circuit just senses input and output currents and strives to keep a fixed, costant ratio between the two. > It looks like a 'current amp'.
While a current amp shall have a very high output impedance (ideally infinte). And this will 'add' the Re of speaker and create a new Qes of the system, which might be too high.
Or, your circuit has variable output impedance?
Would you like to see my work here: http://www.diyaudio.com/forums/multi-way/155022-low-q-woofer-ob-high-zo-amp.html
Very interesting. Thanks a lot for sharing.
By < The multiplier circuit just senses input and output currents and strives to keep a fixed, costant ratio between the two. > It looks like a 'current amp'.
While a current amp shall have a very high output impedance (ideally infinte). And this will 'add' the Re of speaker and create a new Qes of the system, which might be too high.
Or, your circuit has variable output impedance?
Would you like to see my work here: http://www.diyaudio.com/forums/multi-way/155022-low-q-woofer-ob-high-zo-amp.html
it does, in a sense. Its output impedance is the output impedance of whatever is attached to its input, divided by the (current) gain of the circuit. And vice-versa.
By multiplying the current it just changes the impedance (and power) magnitudes, without interacting directly with either the source or the load. In principle (ideally), it "does not have" an input or output impedances of its own. That is, they are (should be) indeed infinite. It is in fact a current amplifier. But if you attach a load to the output, that impedance appears (scaled) to the input. Same for input: whatever impedance is there, that's what you get (scaled) at the output.
As said, the circuit looks a lot like a transformer: it let input and output "see" each other through sort of a magnifying glass, or better a ("non inverting", terrestrial) telescope.
(have you ever tried looking into one of them from either sides, seeing objects appearing bigger or smaller?
that's much the same idea...).I had the idea to use such a circuit exactly for the purpose of replacing the (bulky & expensive) OPT of a SET amplifier. With the added bonus of the possibility to get high output powers without even the need for big power tubes.
An ECC88 and an LM3886 will give you a wonderful SET amplifier capable of up to some 50W. But, differently from any other hybrid design tried that I know of, only this one will give you a true SET sound!
For you not only get the true "distorsion signature" of a real SET ('cause the triode is actually loaded by the complex impedance of the real LS exactly as it happens in a real SET with OPT), but you also have the triode actually controlling (and damping it's own way) the LS, again much the same way it happens in a real SET with OPT.
As said, basically the circuit "changes" the real LS into a virtual one with much higher impedance & sensitivity. Or, if you prefer, the other way around. It does make your little triode appear as a "super triode" which can drive directly the LS.
Either way, if the SS amplifier with its NFB loop does its job properly and does not add too much of its own (distorsion, IMD, ecc), what you end up with works exactly like a real SET. And sounds like it.
Cool, isn't it?
Last edited:
Hi UnixMan,
Have you simulated this with an equivalent loudspeaker circuit as the amplifier load,
ie. with current drawn changing during first cycle time to become out of phase with the voltage supplied ?
(I don't like SET sound - amplifiers ought not introduce any 'sound'!)
Hi CLS,
Not had time - will check your link shortly.
Have you simulated this with an equivalent loudspeaker circuit as the amplifier load,
ie. with current drawn changing during first cycle time to become out of phase with the voltage supplied ?
(I don't like SET sound - amplifiers ought not introduce any 'sound'!)
Hi CLS,
Not had time - will check your link shortly.
Last edited:
Long before I eventually built it, I did a lot of simulations. For sure I have simulated a lot of different loads to verify if the circuit was behaving as expected, but I don't think I ever had a look at "first cycle", sorry. I'll check if I can still find the files and have a look at it.
Nevertheless, I would not expect any surprise. First cycle response should be whatever is "imposed" by the front-end interacting with the LS.
One of the nice things about this little circuit is that (in itself) it is insensitive to load phase. Since the NFB loop only takes care of (in and out) currents, regardless of voltage, in principle phase lag between voltage and current should not make any difference, no matter how big and/or sudden.
well, that's a matter of taste.
As a lot of other people around, I do like the way they sound. Besides, in principle I may agree with you, but unfortunately I have never heard any amp (or even just piece of wire, for that matter...) which does not have its own "sound signature".BTW, of course I was not suggesting you to do that.
I was only suggesting the possibility to use this kind of "trick" (the impedance multiplier) to be able to implement the "T-Bass" at higher impedance / lower power level, which (perhaps) may be more convenient.Knowing your preferences, I would suggest you to try something like a modified JLH (you need the full voltage but just very small current and power... perfect for Class A operation without much dissipation) as the voltage amp front-end.
Oh, before I forget to mention... if you are going to have a very low Zin front-end, which as you said is required for T-Bass, you must use an amp which is stable at unity gain for the impedance multiplier. Thus you can not just try using an LM3875/3886 as I did in my prototype. But perhaps you can use a modified GEM...
Last edited:
Hi UnixMan.
You wrote >> One of the nice things about this little circuit is that (in itself) it is insensitive to load phase. Since the NFB loop only takes care of (in and out) currents, regardless of voltage, in principle phase lag between voltage and current should not make any difference, no matter how big and/or sudden. <<
I disagree R2 is dropping a current (phase) related voltage, thus it is wrong to study steady sines only after the dynamic changes have settled. Also, as this circuit allows output circuit reactions to modify input impedance, its voltage gain and dynamic response is additionally related to any impedance within the source. Whatever simulation tests you do, it is necessary to place an equivalent impedance for the source (your Rs) as well as the load, because a computer generated voltage source has perfect 'zero' output impedance.
If a triode is the source you will need the triode anode resistance and capacitance variation characteristics with anode voltage amplitude swing, also the internal capacitive shunt NFB characteristics between grid and anode which will be influenced by varying loudspeaker impedance transmitted via the input of the output stage to the triode anode.
So yes, I do understand the nature of this circuit, but I would not like to use it.
(The GEM is a JLH class-A type with parallel class-AB output to provide more efficient power output, but it does not require a choke bridge as in the Quad Current Dumper where their class-A is empowered with class-B drive. In both the distorting power section can be viewed as being real time corrected via parallel class-A working.)
You wrote >> One of the nice things about this little circuit is that (in itself) it is insensitive to load phase. Since the NFB loop only takes care of (in and out) currents, regardless of voltage, in principle phase lag between voltage and current should not make any difference, no matter how big and/or sudden. <<
I disagree R2 is dropping a current (phase) related voltage, thus it is wrong to study steady sines only after the dynamic changes have settled. Also, as this circuit allows output circuit reactions to modify input impedance, its voltage gain and dynamic response is additionally related to any impedance within the source. Whatever simulation tests you do, it is necessary to place an equivalent impedance for the source (your Rs) as well as the load, because a computer generated voltage source has perfect 'zero' output impedance.
If a triode is the source you will need the triode anode resistance and capacitance variation characteristics with anode voltage amplitude swing, also the internal capacitive shunt NFB characteristics between grid and anode which will be influenced by varying loudspeaker impedance transmitted via the input of the output stage to the triode anode.
So yes, I do understand the nature of this circuit, but I would not like to use it.
(The GEM is a JLH class-A type with parallel class-AB output to provide more efficient power output, but it does not require a choke bridge as in the Quad Current Dumper where their class-A is empowered with class-B drive. In both the distorting power section can be viewed as being real time corrected via parallel class-A working.)
Last edited:
of course it does, that's what provides input signal. My point (interpretation) is, both inputs of the amplifier are referred to the output terminal rather than ground. Neither one is referred to ground. Thus the amplifier should sort of "float" with respect to input and output voltages (of course it is not really floating as it is "internally" referred to ground via the power supply rails). I assume that the amplifier "see" (thus "cares" about) only the voltage drops across R1 and R2, that is input and output currents, regardless of input/output voltages. But input and output currents should be in phase with each other (except for any "internal", amplifier induced phase lag), thus the circuit itself should be "indifferent" to voltages and consequently should not be sensible to (source and/or load induced) phase lags.
Am I wrong?
of course, that's why I put it there explicitly.
right. That's exactly what I meant when I said that this circuit can be used to closely "mimic" a real SET. The loudspeaker is totally controlled by the front-end (in that case a triode) and not at all by the impedance multiplier.
you mean, you would not like to use it with a triode as a "front-end" (which is not what I was suggesting to you) or you would not like to use it at all? If this is the case, may I ask why do you think it's not worth a try in your application?
yes, I understand. I was suggesting the extra complication of turning a GEM (or some other good NFB-based amp) into an impedance multiplier and using another JLH (or whatever you like) as a front-end to be able to add the T-Bass network in the middle.
That way you would no longer need bulky iron for it, as the higher impedance and smaller power level would allow to use much smaller and cheaper parts.
Hi UnixMan.
You wrote >> may I ask why do you think it's not worth a try in your application? <<
because the voltage applied to the loudspeaker will be
1. triode dynamic/ static non linear,
2. modified by a plethora of driver/ system induced current resultants
(driver impedance charts are smoothed!)
3. amplifier performance will be modified differently by every different LS
(not what LS designers use as reference source)
4. I prefer to hear clean loudspeaker reproduction of a clean amplifier waveform
5. there would be need for much higher voltage power without the transformer.
You wrote >> may I ask why do you think it's not worth a try in your application? <<
because the voltage applied to the loudspeaker will be
1. triode dynamic/ static non linear,
2. modified by a plethora of driver/ system induced current resultants
(driver impedance charts are smoothed!)
3. amplifier performance will be modified differently by every different LS
(not what LS designers use as reference source)
4. I prefer to hear clean loudspeaker reproduction of a clean amplifier waveform
5. there would be need for much higher voltage power without the transformer.
Graham, I'm afraid perhaps you've still not understood what I was trying to tell you!
Forget about the triode. That's what I have done for a completely different goal. It has nothing to do with T-Bass. I was NOT suggesting you to use a triode anywhere!
The idea I am trying to suggest to you is to use an "impedance multiplier" to allow T-Bass controlling the loudspeaker at lower power levels.
But (of course) NOT using a tube-based "front-end"!
Using a fully SS, low distorsion, low output impedance, ecc, type of "front-end" voltage amplifier.
Whatever you like to optimally drive the T-Bass network. In principle, the voltage amplifier could even be the same power amp you are currently using. But - since you will be driving a much higher impedance and will need very little power - you'd better use a specifically designed voltage amplifier for that.
That is, setup as follows:
[SS voltage amplifier] <-> [T-Bass network] <-> [impedance multiplier] <-> loudspeaker
With this connection the T-Bass should still see a low-Z amplifier output on one side and the loudspeaker varying load (and phase) on the other. And it should interact with them in much the same way it does in its "traditional" setup.
But it does so at a greatly reduced power level. No need for bulky stuff anymore.
Oh, as said, the "Op-Amp" shown on the schematic diagram needs not to be a real integrated (power)Op-Amp chip. You'd better use some good NFB based SS amplifier (such as your GEM) for that purpose. You only have to properly alter its NFB loop to change it from a voltage amplifier to an impedance multiplier (that is a current amplifier). This should be doable with just about any NFB-based SS amplifier.
Have you got the idea now?
Conceptually, the current mirror idea makes perfect sense.
Graham may be missing the point that this circuit is a current mirror, not just a current amplifier. It may be helpful to visualise the current mirror as a step-down transformer. It's a bidirectional device, it reflects load impedance changes back to its input. That is, its input impedance reflects (tracks) the impedance at the output. For example, for a 10:1 ratio mirror, if the load impedance is 1 ohm, the input impedance will be 100 ohms. Change the load impedance to 0.5 ohms, the input impedance will change to 50 ohms.
A directly coupled T-bass has to work at loudspeaker impedances, with the requirement for large components. A current mirrored T-bass circuit can have higher impedance components. Imagine the current mirror is a stepdown transformer with a 100:1 ratio. 1 amp in = 100 amps out. This also means that a 10,000 ohm resistor at the input will look like a 1 ohm resistor at the output. Instead of having to use T-bass components with less than 0.5 ohm resistance windings (and which also have to be large enough to handle the high currents), the windings can be hundreds of ohms without causing a problem. The cores can be much smaller too because the current is only 1% of the loudspeaker-level T-bass requirements.
As for the amplifier power required, it's actually much easier to provide than in the loudspeaker-level T-bass. A current mirror driving a standard loudspeaker only has to handle the impedance presented by the speaker driver. An amplifier driving a T-bass has to handle (potentially) very low impedances at awkward phase angles. Any designer will tell you that it's a lot easier to provide a high voltage swing at the output than a high current.
Graham may be missing the point that this circuit is a current mirror, not just a current amplifier. It may be helpful to visualise the current mirror as a step-down transformer. It's a bidirectional device, it reflects load impedance changes back to its input. That is, its input impedance reflects (tracks) the impedance at the output. For example, for a 10:1 ratio mirror, if the load impedance is 1 ohm, the input impedance will be 100 ohms. Change the load impedance to 0.5 ohms, the input impedance will change to 50 ohms.
A directly coupled T-bass has to work at loudspeaker impedances, with the requirement for large components. A current mirrored T-bass circuit can have higher impedance components. Imagine the current mirror is a stepdown transformer with a 100:1 ratio. 1 amp in = 100 amps out. This also means that a 10,000 ohm resistor at the input will look like a 1 ohm resistor at the output. Instead of having to use T-bass components with less than 0.5 ohm resistance windings (and which also have to be large enough to handle the high currents), the windings can be hundreds of ohms without causing a problem. The cores can be much smaller too because the current is only 1% of the loudspeaker-level T-bass requirements.
As for the amplifier power required, it's actually much easier to provide than in the loudspeaker-level T-bass. A current mirror driving a standard loudspeaker only has to handle the impedance presented by the speaker driver. An amplifier driving a T-bass has to handle (potentially) very low impedances at awkward phase angles. Any designer will tell you that it's a lot easier to provide a high voltage swing at the output than a high current.
No Don,
It is you who is missing the point - I am not interested in this gyrator circuit.
Also this thread is for the T-bass, esp as applied to OB.
The arrangement you and UnixMan are proposing will require a separate LF amplifier of at least twice the power to equal T-bass dynamics, and whether it will actually perform the same I have no idea.
I use one good amp per channel only.
It is you who is missing the point - I am not interested in this gyrator circuit.
Also this thread is for the T-bass, esp as applied to OB.
The arrangement you and UnixMan are proposing will require a separate LF amplifier of at least twice the power to equal T-bass dynamics, and whether it will actually perform the same I have no idea.
I use one good amp per channel only.
Last edited:
No, a current mirror will *not* require an amplifier of twice the power. In fact, you have the situation almost exactly reversed. In practice, an amplifier for T-bass duty will require four times the power rating (more precisely, the ability to maintain the same output voltage at 1/4 the load impedance) as a standard (non T-bass) or current mirror circuit.
Let's take an example of an 8 ohm speaker which requires 32 watts to achieve a required SPL. That's a current of 2 amps, and the amplifier will need to provide a 16 volt output swing.
Standard circuit: Requires an amplifier capable of driving 2 amps into 8 ohms.
T-bass circuit: Requires an amplifier capable of driving 4 amps into 2 ohms, ignoring additional series L-C resonance dip.
Current mirror circuit: Requires a current mirror capable of driving 2 amps into 8 ohms.
The net result is, regardless of topology, the same amplifier *power* capability is required.
In theory, an amplifier for T-bass only needs to provide half the output voltage swing as that for a standard circuit. But at higher frequenceis, where the T-bass is effectively shorted and the normal speaker impedance is reflected at the amplifier, the higher voltage swing is again required to achieve rated SPL.
Let's take an example of an 8 ohm speaker which requires 32 watts to achieve a required SPL. That's a current of 2 amps, and the amplifier will need to provide a 16 volt output swing.
Standard circuit: Requires an amplifier capable of driving 2 amps into 8 ohms.
T-bass circuit: Requires an amplifier capable of driving 4 amps into 2 ohms, ignoring additional series L-C resonance dip.
Current mirror circuit: Requires a current mirror capable of driving 2 amps into 8 ohms.
The net result is, regardless of topology, the same amplifier *power* capability is required.
In theory, an amplifier for T-bass only needs to provide half the output voltage swing as that for a standard circuit. But at higher frequenceis, where the T-bass is effectively shorted and the normal speaker impedance is reflected at the amplifier, the higher voltage swing is again required to achieve rated SPL.
Last edited:
Sadly Don you too show you do not understand the nature of the T-bass. It is the dynamic current flows during music time which count. Loudspeaker dynamics change with frequency, and should not be driven with constant gyrator ratio. Your reasoning in the last paragraph is wrong.
Anyone can quote sine amps and volts; you are ignoring dynamics !!!!!
Go A-B compare a T-bass with the circuit you are trying to make me accept, and then let us all know your findings.
This thread is for the *already proven* T-bass - if you do not mind !
Have you tried it yet, because until you do you really will not understand what it achieves ?
ALSO I REPEAT
I DON'T WANT TO USE MORE THAN ONE GOOD AMPLIFIER, AS YOU ARE PROPOSING !
So kindly stop trying to persuade me to !
Anyone can quote sine amps and volts; you are ignoring dynamics !!!!!
Go A-B compare a T-bass with the circuit you are trying to make me accept, and then let us all know your findings.
This thread is for the *already proven* T-bass - if you do not mind !
Have you tried it yet, because until you do you really will not understand what it achieves ?
ALSO I REPEAT
I DON'T WANT TO USE MORE THAN ONE GOOD AMPLIFIER, AS YOU ARE PROPOSING !
So kindly stop trying to persuade me to !
Last edited:
Graham,
You may understand the T-bass. But you fail to understand the nature of the current mirror It *will* faithfully reflect the loudspeaker dynamics back to the "line level" T-bass circuit. It will "reflect forwards" too, and faithfully reflect the T-bass circuit characteristics to the driver.
Perhaps the analogy of power steering, or power controls in an aircraft, may help. That's a "force mirror". You can still feel the road / air "pushing back", but you don't need such big muscles to move the wheels / elevators.
How is it wrong? I used your own reasoning and explanation of T-bass operation from one of your earlier posts. I can quote you if you wish.
I specifically made no mention of sine values. I didn't have to, the bare numbers are proof enough. If I were to specifically detail the dynamic scenario of the T-bass operating conditions, I could show that the situation for amplifier power is even worse than I described before.
Where is the proof? We have only anecdotal evidence - people who say it works. But we have no recordings of reproduced tone bursts. No analytical evidence (worked examples or simulations). I'm working towards achieving some of those goals.
I believe I understand quite well what the T-bass is intended to achieve. You've done a good job of describing that thorugh this thread. But I'm not going to try it until I have a calculation method that will allow me to establish a set of component values that have a good chance of being close to correct first time. I will *NOT* "suck it and see" with random values of whatever components I have around. I've done this too many times, and seen what happens when others do it too, particularly in this thread.
Two points here:
First, please quote where I said you need more than one good amplifier.
You appeared to get that impression from something Unixman said in his first post, and he has tried vainly several times since to correct your understanding.
What you need:
- A line level preamplifier (conventional), coupled to:
- A line level T-bass circuit, coupled to:
- A current mirror amplifier, coupled to:
- A driver.
A current mirror amplifier uses the same components and mostly the same topology as a normal power amplifier. The same design rules apply, so it's no harder to build a good current mirror than it is to build a good voltage amplifier.
Second point, We are not trying to persuade you to *do* anything. You've made it abundantly clear you have no interest in working on T-bass for the foreseeable future. All your gear is packed away and you have more pressing things to worry about.
I'm working towards a design method that anyone can use to pick their driver, calculate the component values and end up with a circuit that will work "as advertised". They are then in the best position to judge for themselves the effect of a correctly working T-bass. Unixman weighed in with a good idea on how to reduce the component costs.
Now if all this is distasteful to you, just ask me to leave. I'll happily go and open a new thread to document the work.
Regards,
Don.