Here another example of making up TaN resistors from SMDs to replace 0.5W leaded :
https://www.diyaudio.com/community/threads/the-holy-grail-follower-output-stage.406850/post-7697283
Patrick
.
https://www.diyaudio.com/community/threads/the-holy-grail-follower-output-stage.406850/post-7697283
Patrick
.
Attachments
My apologies for my negligence in not quoting the original sources of regulators.
To complete the picture :
https://www.linearaudio.net/superregs
The article by Walt Jung and J Walton as attachment.
And another example of "Super Regulators" for power amps that claims to be even more superior :
https://www.belleson.com/comparealw.php
(No, I haven't tried and have no intention to.)
Patrick
.
To complete the picture :
https://www.linearaudio.net/superregs
The article by Walt Jung and J Walton as attachment.
And another example of "Super Regulators" for power amps that claims to be even more superior :
https://www.belleson.com/comparealw.php
(No, I haven't tried and have no intention to.)
Patrick
.
Hey Patrick
I built mine monster (Hiraga ... c.f.r. AKIBA) in 1983, in full DIY.
2x4 units (mine octopus), PSU with caps 2F, 2SK170/2SJ74/2SC1775/2SA87/2SB716/2SD756/2SD844/2SB754/ ....
In complement with 4 way actif X-over Kaneda filter.
Sublim .... it was music in heaven (with 2x4 (full) Horn concept)
Welkom in the Monster club.
Alles, Salukis.
Karel
I built mine monster (Hiraga ... c.f.r. AKIBA) in 1983, in full DIY.
2x4 units (mine octopus), PSU with caps 2F, 2SK170/2SJ74/2SC1775/2SA87/2SB716/2SD756/2SD844/2SB754/ ....
In complement with 4 way actif X-over Kaneda filter.
Sublim .... it was music in heaven (with 2x4 (full) Horn concept)
Welkom in the Monster club.
Alles, Salukis.
Karel
By the way: I recently had my list of parts and their sound in my hand: for example, SD1047, one 40 years old: no bass, cloudy. Another 5 years old: bass behind, cloudy. Now we are building a complementary push-pull amplifier, and we don't even know what the chosen components sound like - and I don't even know any diyers who choose the complementary parts for their complex push-pull amplifier according to sound;-) And complementary transistors sound more different than equal. The advantage of Hiraga's Classe A, for example, is to use complementary parts in both half-waves. I would also build the last two stages of a Hiraga Le Monstre with the same complementary pair. And also use TO-220 types. They generally sound more focussed and the tones don't generally diverge;-)
aïe aïe aïe ....
@ patrick
* Thanks for your golden thumb.
FYI
FYI
I remember a test & home:
* With a full Horn concept (4 way) & 2x4 monsters (Octopus) I measure 105dB @ 4watt before clipping ....
Music in heaven ....
Alles, salukes.
Karel
@ patrick
* Thanks for your golden thumb.
FYI
- PCB are identique (in home made ...copy/paste ... hand made design ) ....
- Tantalium, (I présume in the actief X-over) ... I hate it .. After a long time power off it's a real danger (explosion-fire). I change it with 2,2micro/63V ...
- Trimmers .... no TOCOS .... simply, the very good BOURNS multi turn ....
FYI
I remember a test & home:
* With a full Horn concept (4 way) & 2x4 monsters (Octopus) I measure 105dB @ 4watt before clipping ....
Music in heaven ....
Alles, salukes.
Karel
In case anyone cares.. Those are from the "PFC" series, purchased at Mouser. The 1206 case code is rated 330mW. What you see there is my series feedback "resistor", 40k (4x 10k in series) with a 1.3W rating. A little B. Hofer influence there, feedback shunt resistor is another identical 10k. (The gain is set low in my FE because I have a tube preamp that is not allowed to loaf..)
SMD adapters from fleabay..
Slightly OT: I once (a century or so ago) build what I remember as a Hiraga Le Monstre. But in my memory it had just one single active device.
Anyone knows that design?
Edit: The Nemesis! https://www.linearaudio.net/nemesis-rebuild
You want to try this one, Patrick? ;-)
Jan
Anyone knows that design?
Edit: The Nemesis! https://www.linearaudio.net/nemesis-rebuild
You want to try this one, Patrick? ;-)
Jan
Attachments
Last edited:
We measured this at your place together last summer :
https://www.diyaudio.com/community/threads/the-jean-hiraga-nemesis-quadrige.401114/
Patrick
This is a real PP amplifier. Here the half-waves are amplified with (almost) identical sounding parts! I would like to repeatedly point out the difference to the complementary transistors pp, where the half-waves are often amplified by very different sounding parts. These should have no place in audio, hi-fi or high-end. Which is not to say that some concepts are at least party-fit and a lot of fun - and satisfaction;-)
Putting together TONAL pairings of complementary parts would be extremely complex and time-consuming. Nobody does that. And certainly not with the well-known parts tombs that fill glossy brochures;-)
Unfortunately: TO-3 transistors are among the worst sounding of all. And: Very few people pay attention to their current-defining connection.
Putting together TONAL pairings of complementary parts would be extremely complex and time-consuming. Nobody does that. And certainly not with the well-known parts tombs that fill glossy brochures;-)
Unfortunately: TO-3 transistors are among the worst sounding of all. And: Very few people pay attention to their current-defining connection.
How about an all-FET version, similar to the ACA-mini? I'm wondering if some small MOSFETs might shine in place of the BJT cascodes?
The cascode seems a bit tricky / sleight of hand-ish, because it sneaks in lots of voltage gain (for my lateral version, anyway), even though the JFETs are responsible for the current gain.
My simulator seems to like a J111 + J177 combo for high current gain, though I guess some premium JFETs could be paralleled.
It seems doubly interesting what happens when the gain on one side is dialed down to make the combined gain symmetrical. Does the cascaded distortion fully cancel?
The cascode seems a bit tricky / sleight of hand-ish, because it sneaks in lots of voltage gain (for my lateral version, anyway), even though the JFETs are responsible for the current gain.
My simulator seems to like a J111 + J177 combo for high current gain, though I guess some premium JFETs could be paralleled.
My understanding is the output stage has 2 current sources summed, and the shared feedback line splits the difference, so to speak. Edit: you are right, the n/p JFETs have independent gains.
It seems doubly interesting what happens when the gain on one side is dialed down to make the combined gain symmetrical. Does the cascaded distortion fully cancel?
Last edited:
It has taken us some extra time, but we finally got round to measure the Le Monstre Deluxe as well.
This has original transistors, all TaN resistors, including Shinkoh for the 220R in the feedback network, RA12P trimmer, Riedon 1R 35W thin film, etc.
As described in the article.
The surprising thing is, that H2 & H4 are ~20dB lower than the previous prototype using same active devices.
So what has changed ? 20dB less even harmonics cannot be just due to the passives.
Well, the JFETs used in this were matched at Id with 50R source resistors, instead of at Idss.
They also happened to have quite close Yfs at that current.
On top of that, we used the current source bypass to keep the 1k drain resistors and 0.7A bias.
So the trimmer was almost at its mid-point when DC offset is trimmed.
And now you get near perfect even harmonics cancellation from top and bottom halves.
And maybe the higher bias with the original output devices also contribute to some extra distortion.
Just tells you that no two Le Monstre's can be exactly the same, unless you measured every active component upfront.
Cheers,
Patrick
This has original transistors, all TaN resistors, including Shinkoh for the 220R in the feedback network, RA12P trimmer, Riedon 1R 35W thin film, etc.
As described in the article.
The surprising thing is, that H2 & H4 are ~20dB lower than the previous prototype using same active devices.
So what has changed ? 20dB less even harmonics cannot be just due to the passives.
Well, the JFETs used in this were matched at Id with 50R source resistors, instead of at Idss.
They also happened to have quite close Yfs at that current.
On top of that, we used the current source bypass to keep the 1k drain resistors and 0.7A bias.
So the trimmer was almost at its mid-point when DC offset is trimmed.
And now you get near perfect even harmonics cancellation from top and bottom halves.
And maybe the higher bias with the original output devices also contribute to some extra distortion.
Just tells you that no two Le Monstre's can be exactly the same, unless you measured every active component upfront.
Cheers,
Patrick