A few components, some electricity and a new adventure begins - a journey. How can we describe the fascination?
Ever since I was a child, I've been fascinated by amplifiers - you can turn them or push them, even the "on" pop is somehow part of it, or the switches, the rocker switches of the 70s.
Long story short:
I have never heard a JLH, not even an original Hiraga. Other concepts have. I don't need much power and it's already warm enough anyway, why not make do with 10 or 20 honest watts?
Maybe the attached ideas are interesting - the components are slumbering and gathering dust in my boxes. I originally wanted to use the IRFP044 in the ZEN4, but then it became an Aleph5 and the HEXFETs have been lying unused on the shelf wrapped in aluminum foil for decades.
Now it's about time...
Kind Regards,
HBt.
I would be pleased to receive comments, suggestions and cooperation
.
Ever since I was a child, I've been fascinated by amplifiers - you can turn them or push them, even the "on" pop is somehow part of it, or the switches, the rocker switches of the 70s.
Long story short:
I have never heard a JLH, not even an original Hiraga. Other concepts have. I don't need much power and it's already warm enough anyway, why not make do with 10 or 20 honest watts?
Maybe the attached ideas are interesting - the components are slumbering and gathering dust in my boxes. I originally wanted to use the IRFP044 in the ZEN4, but then it became an Aleph5 and the HEXFETs have been lying unused on the shelf wrapped in aluminum foil for decades.
Now it's about time...
Kind Regards,
HBt.
I would be pleased to receive comments, suggestions and cooperation
.
Idle-Current
Perhaps a few more comments or thoughts on the three ideas would be helpful or of interest? I can't resist the charm of the two concepts, Jean Hiraga's symmetrical structure and the unusual JLH69 approach. I don't even understand why, but it's probably because of the simple elegance of Hiraga and the naturalness of a rock massif in the JLH concept. I simply don't know, but both of these representatives of their guild just won't let go of me. With the SYMM by Helge Peters (Klang & Ton magazine 1991) and the Class A LFA by Elektor, also from 1991, the question of class began to haunt me. Both amplifiers sounded good and were technically perfectly fine. But as an infected person, I wasn't satisfied and other DIY attempts followed. Among friends, we compared and listened to ... and we always came back to the MF-A1 or the NAD3020 or a good old Technics. Both the 3020 and the A1 have a special magic - I know them inside out.
This long preamble is actually only intended to address the quiescent current of the A1, statically 0.8A and dynamically 1.6A -> Tim de Paravincini himself provided the explanation for this and it immediately made sense to me - the Elektor LFA also follows this principle.
The JLH69, also from 1996, is completely different in this respect - more of a single-ended principle, but opinions differ widely on this.
This 0.8A always in the back of my mind is the motor for the first and also the following designs, a constant.
Perhaps a few more comments or thoughts on the three ideas would be helpful or of interest? I can't resist the charm of the two concepts, Jean Hiraga's symmetrical structure and the unusual JLH69 approach. I don't even understand why, but it's probably because of the simple elegance of Hiraga and the naturalness of a rock massif in the JLH concept. I simply don't know, but both of these representatives of their guild just won't let go of me. With the SYMM by Helge Peters (Klang & Ton magazine 1991) and the Class A LFA by Elektor, also from 1991, the question of class began to haunt me. Both amplifiers sounded good and were technically perfectly fine. But as an infected person, I wasn't satisfied and other DIY attempts followed. Among friends, we compared and listened to ... and we always came back to the MF-A1 or the NAD3020 or a good old Technics. Both the 3020 and the A1 have a special magic - I know them inside out.
This long preamble is actually only intended to address the quiescent current of the A1, statically 0.8A and dynamically 1.6A -> Tim de Paravincini himself provided the explanation for this and it immediately made sense to me - the Elektor LFA also follows this principle.
The JLH69, also from 1996, is completely different in this respect - more of a single-ended principle, but opinions differ widely on this.
This 0.8A always in the back of my mind is the motor for the first and also the following designs, a constant.
Eisenport: The first idea
Nothing speaks against a MOSFET version and somehow I also like the idea of Nelson Pass, his PLH version. So I thought I'd give it a try over the last few days. And since the HEXFETs are really quite unique - especially regarding their tendency to undesirable behavior, we sadly can see it for example in Elektor's HEXFET MPA from 1993 ... but work great in the pass amplifiers, I want to take up the challenge. A quiescent current stabilization or regulation is obviously essential, International Rectifier confirms this anyway, so the JLH96 modification is the obvious choice. That is the background to this idea. Two-stage amplification like the Aleph series, but also with a differential amplifier as the first stage - so that the negative feedback can really get to work, let's see what it can iron out.
Nothing speaks against a MOSFET version and somehow I also like the idea of Nelson Pass, his PLH version. So I thought I'd give it a try over the last few days. And since the HEXFETs are really quite unique - especially regarding their tendency to undesirable behavior, we sadly can see it for example in Elektor's HEXFET MPA from 1993 ... but work great in the pass amplifiers, I want to take up the challenge. A quiescent current stabilization or regulation is obviously essential, International Rectifier confirms this anyway, so the JLH96 modification is the obvious choice. That is the background to this idea. Two-stage amplification like the Aleph series, but also with a differential amplifier as the first stage - so that the negative feedback can really get to work, let's see what it can iron out.
Attachments
Less than 24 hours later, I was tormented by the question of whether a good design and concept would not require a field-effect transistor on the input and output side. In any case, Eisenport "der Alber" can be modified immediately in this direction (if it works at all in reality) and a cascode does no harm, quite the opposite in fact. We can use almost any type of jFet, not only Toshibas famos dual one. But I still have two pairs hidden somewhere.
Somehow the second idea was very obvious
Somehow the second idea was very obvious
Attachments
Eisenport: The third idea
My gut feeling is that there could be problems with the behavior of the connected speaker, the complex impedance and the emf. With a purely ohmic resistance, the simulations are promising.
If the rear fidgets, pulls and pushes, who knows how the power amplifier will behave?
A quasi-complementary de-amplifier, a push-pull operation, could perhaps combat the evil - and I like my father-in-law's ancient DUAL amplifier (CV120/121), it sounds amazingly good and works perfectly after almost fifty years (is in daily use). The quiescent current stabilization works here via an NTC resistor. So just swap the BJTs for N-MOS-FET.
The driving question is whether it (NTC) works with the HEXFET, as active quiescent regulation can still take place (see for example D. Self's Novel controller).
My gut feeling is that there could be problems with the behavior of the connected speaker, the complex impedance and the emf. With a purely ohmic resistance, the simulations are promising.
If the rear fidgets, pulls and pushes, who knows how the power amplifier will behave?
A quasi-complementary de-amplifier, a push-pull operation, could perhaps combat the evil - and I like my father-in-law's ancient DUAL amplifier (CV120/121), it sounds amazingly good and works perfectly after almost fifty years (is in daily use). The quiescent current stabilization works here via an NTC resistor. So just swap the BJTs for N-MOS-FET.
The driving question is whether it (NTC) works with the HEXFET, as active quiescent regulation can still take place (see for example D. Self's Novel controller).
Attachments
There are an abundant amout of issues addressed in this single topic. A large dish which needs a lot of chewing through.
Nevertheless is this 'Eisenport-concept' ("Iron-gate" ? - vergib mir meine seit 150 Jahren vergessene Muttersprache) quite different from the boring repititions of those well known and copied concepts mounting and suffocating the digital 'smart' portals.
You refer to Hiraga's design, but there's not that much left in this Eisenport concepts.
Jean Hiraga deliberately choose for an absolute symmetrical topology to have any uneven harmonics suppressed as much as possible - a pita for all unsym designs relying on sufficient amounts of overall negative feedback to 'keep the race car into track', an approach I sincerely disapprove.
There are more issues to address here, yet, appreciated!
Nevertheless is this 'Eisenport-concept' ("Iron-gate" ? - vergib mir meine seit 150 Jahren vergessene Muttersprache) quite different from the boring repititions of those well known and copied concepts mounting and suffocating the digital 'smart' portals.
You refer to Hiraga's design, but there's not that much left in this Eisenport concepts.
Jean Hiraga deliberately choose for an absolute symmetrical topology to have any uneven harmonics suppressed as much as possible - a pita for all unsym designs relying on sufficient amounts of overall negative feedback to 'keep the race car into track', an approach I sincerely disapprove.
There are more issues to address here, yet, appreciated!
Running one of the output transistors at constant current is doomed to very low efficiency.
The third idea can approach 50% efficiency.
IMO, the way to design class A is to start with a modern class AB topology. The supply voltages and biasing can be adapted to class A operation. This will produce a superior amplifier compared to old class A designs.
Ed
The third idea can approach 50% efficiency.
IMO, the way to design class A is to start with a modern class AB topology. The supply voltages and biasing can be adapted to class A operation. This will produce a superior amplifier compared to old class A designs.
Ed
Hmm... was the AB operation not a more power efficient solution to the 'warming' class A operation, available by substantial different topologies?
And what differs 'new' from 'old' designs? What kind of truely different approaches are made since then?
The OP's proposition is leading in this discussion.
Yes, but class AB topologies can also run in class A. That gives up efficiency and power output but retains all other desirable traits.Citizen124032 said:
Modern designs have DC coupling, higher loop gain, stable bias, good PSRR, and take advantage of complementary pairs.Citizen124032 said:
Ed
I think you got that wrong. Even harmonics are asymmetrical and uneven harmonics are symmetrical !
The way Hiraga got the desired spectrum was by utilising the fact that k170 and j47 are not real complementary thereby creating 2.order harmonics.
Maybe check some writings of Nelson Pass about this.
😉
Many thanks to all of you .
The dish served hides something that is true and is exactly as intended, namely the essential points of a concept that are exciting and profitable in a discussion.
Dear @Citizen124032,
I would be delighted if you could give free rein to your thoughts on this. By the way, Eisenport (iron-gate) etymologically correctly expresses what you probably suspect: an inner and outer attitude, a so-called "Hut" (i.e. in historical fencing language something like a guard, a starting point also). It is noble and seemingly harmless. Just like the conceptual ideas of the three designs.
Jean Hiraga only fits into the introduction in so far as it (the concept) has been in my head for decades, the SYMM amplifier mentioned is very similar in design and I still have two monos of it. I believe Accuphase also pursued this concept commercially. Hiraga doesn't necessarily fit into this thread, but you can put these two different concepts on a scale and weigh the pros and cons against each other - so keep them in mind.
At the moment I am very interested in your opinions on the above menu, idea 1 (and consequently idea 2) in comparison to idea 3 - and also completely independently.
Completely open on every point.
Kind Regards,
HBt.
.The dish served hides something that is true and is exactly as intended, namely the essential points of a concept that are exciting and profitable in a discussion.
Dear @Citizen124032,
I would be delighted if you could give free rein to your thoughts on this. By the way, Eisenport (iron-gate) etymologically correctly expresses what you probably suspect: an inner and outer attitude, a so-called "Hut" (i.e. in historical fencing language something like a guard, a starting point also). It is noble and seemingly harmless. Just like the conceptual ideas of the three designs.
Jean Hiraga only fits into the introduction in so far as it (the concept) has been in my head for decades, the SYMM amplifier mentioned is very similar in design and I still have two monos of it. I believe Accuphase also pursued this concept commercially. Hiraga doesn't necessarily fit into this thread, but you can put these two different concepts on a scale and weigh the pros and cons against each other - so keep them in mind.
At the moment I am very interested in your opinions on the above menu, idea 1 (and consequently idea 2) in comparison to idea 3 - and also completely independently.
Completely open on every point.
Kind Regards,
HBt.
Dear @EdGr,
I agree with you on all points - and that is precisely the crux of the matter.
Douglas Self and other authors speak of B operation, which we Germans classically refer to as AB operation, i.e. a bias voltage of the base-emitter paths, which now inevitably leads to a quiescent current flow ... I think there is a consensus here. Historically, we refer to B operation as completely currentless, i.e. basically without the VBE bias voltage. This often leads to confusion. In the meantime, however, I completely follow Self's diction.
The design specification is 100% plausible and with the Trimodal Amplifier approach Douglas Self puts his finger directly in the wound.
But then we could immediately set aside any motivation and the magic is gone.
At this point, I would like to take a closer look at the JLH or PLH-like device with the aforementioned 0.8A quiescent current.
Who knows exactly what surprises can be found?
The distortion spectrum has already been mentioned; compared to the Blameless B concept, all three ideas perhaps leave a watermark, a fingerprint. They may not even work in interaction with our loudspeakers(?).
Greetings,
HBt.
I agree with you on all points - and that is precisely the crux of the matter.
Douglas Self and other authors speak of B operation, which we Germans classically refer to as AB operation, i.e. a bias voltage of the base-emitter paths, which now inevitably leads to a quiescent current flow ... I think there is a consensus here. Historically, we refer to B operation as completely currentless, i.e. basically without the VBE bias voltage. This often leads to confusion. In the meantime, however, I completely follow Self's diction.
The design specification is 100% plausible and with the Trimodal Amplifier approach Douglas Self puts his finger directly in the wound.
But then we could immediately set aside any motivation and the magic is gone.
At this point, I would like to take a closer look at the JLH or PLH-like device with the aforementioned 0.8A quiescent current.
Who knows exactly what surprises can be found?
The distortion spectrum has already been mentioned; compared to the Blameless B concept, all three ideas perhaps leave a watermark, a fingerprint. They may not even work in interaction with our loudspeakers(?).
Greetings,
HBt.
All aspects are welcome, I am happy about it.
In addition to post #2, as information only:
Hiraga
JLH
MF-A1
Ultimately, I'm interested in the Eisenport project as a hopefully interesting and unusual alternative to the well known ones. With the potential to discover one thing or another.
Last edited:
Do these first two links work for you?
Not so here, Rod's site seems down now for some time.
Yes, of course.
No problem here, it works fine, maybe choose https://sound-au.com/tcaas/
It is hosted by ESP.
No problem here, it works fine, maybe choose https://sound-au.com/tcaas/
It is hosted by ESP.
Strange Hugo,
coming from Germany, https://www.sound-au.com/ also works perfectly.
coming from Germany, https://www.sound-au.com/ also works perfectly.
Then perhaps an overview of the circuits (that are permanently haunting the back of my mind) will be helpful, including ZEN4 and PLH in the context of the Project. I think most people are familiar with the sources, so I won't mention them here.
It may also be advantageous to decide in favor of one of the three ideas; as the author, this is (strangely enough) completely impossible for me.
It may also be advantageous to decide in favor of one of the three ideas; as the author, this is (strangely enough) completely impossible for me.
I will probably use the jFet-BJT cascode as the first stage after all, as it somehow feels a little better suited to the concept of the Eisenport project.
#
The simulations of the distortion behavior are basically very promising, but I can see an anomaly on the rising edge of the signal at higher frequencies.
Like a dead time in the positive zero crossing, and strangely enough this also seems to be level-dependent, at required power levels of less than 5Wrms everything seems to be fine.
Clearly visible with THD10k, but it is still < 0.1%, k3 dominates. Strange. Is it due to the different switching times (on and off) of the high-side switch (behavior!), i.e. the Q6?
Or is it caused by the gate stopper resistance of Q5?
I actually wanted to avoid a test setup for a short while, but apparently that's not possible - I don't really trust the simulation.
Greetings,
HBt.
#
The simulations of the distortion behavior are basically very promising, but I can see an anomaly on the rising edge of the signal at higher frequencies.
Like a dead time in the positive zero crossing, and strangely enough this also seems to be level-dependent, at required power levels of less than 5Wrms everything seems to be fine.
Clearly visible with THD10k, but it is still < 0.1%, k3 dominates. Strange. Is it due to the different switching times (on and off) of the high-side switch (behavior!), i.e. the Q6?
Or is it caused by the gate stopper resistance of Q5?
I actually wanted to avoid a test setup for a short while, but apparently that's not possible - I don't really trust the simulation.
Greetings,
HBt.
In general there are only a few options (not to elabourate on A/AB/B/others):
- symmetrical input to output
- semi-symmetrical input to symmetrical output
- symmetrical input to semi-symmetrical output
- semi-sym in to semi-sym out
- unsymmetrical
- buffers (Au=1) with sym/asym in- and/or output
- tube (-like) topology with output transformer
- OTL (output transformer less) with tube and/or solid state
- cyclotrons
- V-fets with the drains to the output (there exists no 'p-ch tubes'; this is an exotic option with only two commercial products so far)
Other approaches:
- 'instrumentation' designs with high accuracy, depending on high nfb levels, which sounds 'confronting' (device better then ears)
- 'natural' designs, lower or none (intrinsic/local) nfb, which sounds 'nice'
- combinations of these to 'balance' the final result
More sides:
- objective vs subjective
- measuring results vs listening experience (personal / audiences / common knowledge)
- preferences (education, social, ethnic, &c)
- perception (biological, talent, experience)
Also:
- the nature of the human species ('lazy opportunists')
- the indestructible belief in one's own right and almost divine immortality ('the final me')
These are the four cardinals ('where the winds blow from') that determine the camps to which one is stationed to.
Agreements, disagreements, discussions, opinions, and so on.
DIYaudio is a well maintained & peacefull landscape, chapeau!
Everything above in italics prooves very controversial.
It's more like asking: "Does one add the pasta to the sauce, add the sauce to the pasta, do not add these components together, doesn't matter, what is pasta and what is sauce, I do not cook, Just hoovering with my saucer here but I'm not from Earth."
Maybe something to set up an 'Ultimate Pasta Poll" in the Lounge forum? (Aliens excluded for obvious reasons.)