I understand your thermal objections, they are not to be dismissed out of hand - and probably only a test setup will shed more light here, I am thinking of the actual disaster - of the extent.
There is no question that there is a worm in the apple here. First of all, I insert the bootstrap.
THD is now -86dB
There is no question that there is a worm in the apple here. First of all, I insert the bootstrap.
THD is now -86dB
Now we can see what is all about the strange position of R1:
0.003% -> 0,001% later on it's up to 0.007% !
0.003% -> 0,001% later on it's up to 0.007% !
My /our original idea was (and is) to transform the present model in such a way that a functional and at the same time magical current dumper circuit is created. For what other reason should RF be so low impedance and be exactly 500 ohms?
However, I need the active support of all Currentdumpers for this magic - cooperation is again kindly requested. I have already wound a 4µ7H air-core coil.
🤔
However, I need the active support of all Currentdumpers for this magic - cooperation is again kindly requested. I have already wound a 4µ7H air-core coil.
🤔
In plain language:
I am in search of the Grail, and would like to achieve the "utopian THD" with the help of magic - the easiest way.
Just kidding, so sorry, sorry for this affront, no offence meant.
Greetings,
HBt.
I am in search of the Grail, and would like to achieve the "utopian THD" with the help of magic - the easiest way.
Just kidding, so sorry, sorry for this affront, no offence meant.
Greetings,
HBt.
What about a thread: solder and test a "5 minutes" amp in less than 50 minutes?😉
I think about that all the time, constantly - but the ravages of time are gnawing away at my limbs.
And above all, I own a flawless REVOX A78
(whose power amplifier module is already pretty close to the five-minute approach).
Young people are welcome to pick up a soldering iron, they have my support ..!
😎
@astx
Before I get started, I have to get the 3/4 + 3/4 amp up and running with @m0rten - heck, I have a new amp idea every few minutes, with the RIAA equalizers it's even worse, by the second. And @stv is also in the starting blocks with the Beelzebabe amp, and the construction sites are piling up. The AA50W also wants to be looked after and the 25-Wwatt Krill is also waiting impatiently.
And my wife really wants to visit Vienna with me.
So for the time being, we're staying in the first development phase of the five-minute amp, and it's all modern, purely digital and virtual with MC12.
Grüße,
HBt.
(PS erwähnte ich bereits, dass ich beim Thema Lautsprecher nicht ganz unbelastet bin?)
Before I get started, I have to get the 3/4 + 3/4 amp up and running with @m0rten - heck, I have a new amp idea every few minutes, with the RIAA equalizers it's even worse, by the second. And @stv is also in the starting blocks with the Beelzebabe amp, and the construction sites are piling up. The AA50W also wants to be looked after and the 25-Wwatt Krill is also waiting impatiently.
And my wife really wants to visit Vienna with me.
So for the time being, we're staying in the first development phase of the five-minute amp, and it's all modern, purely digital and virtual with MC12.
Grüße,
HBt.
(PS erwähnte ich bereits, dass ich beim Thema Lautsprecher nicht ganz unbelastet bin?)
It should be noted that the present design also has two inputs like the usual blameless variant. The base of Q5 is our positive input, its negative input follows the positive input - but now the negative input (Q5's emitter electrode) also represents a node, a summation node for currents. With its size, RF now essentially determines the resulting bandwidth of the now negative feedback system; RN is almost single-handedly responsible for the gain.
A special feature that should not be underestimated.
A special feature that should not be underestimated.
Normally, R2 is very rarely located at this point; on the contrary, it tends to slide upwards, i.e. between the collector of Q7 and the output node. While the emitter of Q2 is now also located on /at the output node.
And so on, we all know this - and simply don't dare to play a little more and explore other points of view. Especially if they still seem nebulous.
I have understood that and it is a really good point, which goes even further, namely to the question of whether various protective devices are still missing for safe operation. And they are still missing completely, but they are not an issue at the moment. This litter /throw should not be able to deliver much more than 16 real watts.
Oh, I see the EbersMoll number.
We also have the option of modifying the Vbe multiplier.
I have no plans to thermally couple Q1 & Q2 with Q6.
Of course, this must not happen and must be prevented, how?
So here I have a knot in my brain; let's leave out the self-heating and consider everything else as external heating (including that caused by the required dynamic load current).
What does the situation look like then, and what influence does this part really have on the instantaneous THD?
Our fears relate solely to Q6, which in itself is absolutely nothing more than an emitter follower (with all its consequences and positive characteristics) - for Q7 it represents a dynamic load (or current source).
With this view, however, R2 suddenly comes back into the picture - the old masters at Studer /Revox even set R2 to a value of 1 ohm ..!
In the given topology, there are some interesting thematic coupling variants, I keep saying that.
Q1 & Q2 are usually located together on an extra heat sink.
Either you try out the whole scenario in practice and experiment a bit (aids are: cold spray, the soldering iron, a hair dryer ...), or you set up a model and simulate the whole thermal snot.
#
As I've already written, I'm not actually that extremely worried - but the scenario we've outlined is still very frightening.
What should we do now?
There must be a solution - if the solution is the problem [loosely based on Paul Watzlawick].
Last edited:
So before I tinker with a voltage regulator IC as an audio amplifier, I'll just use the good old LM12. If everything is well prepared, I only need ten minutes with the LM12.
FYI
look into the attached pdf
FYI
look into the attached pdf
@wahab
I had actually hoped that we would use the circuit shown here as a basis for a conversion to a "current dumper". The aim of this surgical intervention must of course be to allow the already quite low THD of (min)0.001% to sink to unimaginable depths, at least to 0.00001% flat and horizontal for the entire frequency and power range.
😊
I had actually hoped that we would use the circuit shown here as a basis for a conversion to a "current dumper". The aim of this surgical intervention must of course be to allow the already quite low THD of (min)0.001% to sink to unimaginable depths, at least to 0.00001% flat and horizontal for the entire frequency and power range.
😊
Let's just add two additional resistors - and claim: by the amount that Q7 would be warmer than Q6 in the static case, this ratio reverses under dynamic consideration - oops.
What do the total harmonic distortions do now?
What do the total harmonic distortions do now?
I just did this again, over the temperature range from 20°C to 60°C -> no problem.
#
We can also experiment thermally with D1 and/or Q5 and Q3 or mentally couple them with this or that.
The Studer masters have of course placed D1 thermally on the large main heat sink.
LM12 is well known, that was an exceptional opamp, guess that the LM3875/3886 are derived fom this chip,
although nothing is as rugged as the TO3 ancestor.
My point was that with a too low xtor count 2 stages amp perfs wont be worth the effort, with this kind of design
the good balance is 7 transistors excluding the VBEM, otherwise that s a waste of ressources, i ll post some schematics
to clear this point.
Please do this, I guess I could predict what examples you will suggest, maybe the 7PN in the signal path duper.
When will you be ready? The current iteration of the "Alte Schule" is hard to top". THD consists mainly of k2, i.e. H2.
Once again in words: 0.008% fully harmonic 36Wpk into 8 ohms.
@astx
while Toni needs at least 450Wpk into 8 ohms to be really happy. However, one don't build such a fantastic amp in 50 minutes and we don't develop it in five either. This already requires n * 5.
😛
greetings,
HBt.