And You can replace everything you have, paying attention to the right capacity. it will not significantly change the operation of the amplifier.
I 've already try a lot of crazy (change value - resistance , replace Trans. by Jfet , ...) all of it still working .
you should try to build it.
I 've already try a lot of crazy (change value - resistance , replace Trans. by Jfet , ...) all of it still working .
you should try to build it.
Hi Kay, all I can say is that the Hitachi ones you've mentioned are indeed lateral MOSFETs. Are the modern ones basically the same technologically and are they better performing - there are definitely better performing ones at least on paper, as for sound, few people seem to be going for those old old ones as the first choice for high fidelity. Renesas made some mosfets and as far as I understand it's basically Hitachi rebranded. They are supposedly better. I cannot really comment on the ones made in the last decade or so. I'm sure there are people on this form that can comment from a more knowledgeable pov.
Hey guys,
So today I finally had an opportunity to listen to class D amplifier monoblocks that are supposedly even better than Hypex Ncore 400 - they are from a company called Purify. It is highly resolving and the bass is straight up addictive. Haven't heard anything better from class A or AB. But it still did not meet my expectations and I ended up bursting my bubble of imagining that this could replace my Tandberg class AB. It despite being well detailed, sounded a bit dull especially at lower volumes. And the vocals especially seemed to have so little presence comparing to similarly priced Emotiva class AB monoblocks.
Tl;dr class D has amazing world class bass, but lacks the grippiness and emotion and immediacy, authoritativeness, authenticity especially in the mid-range, but dare I say also in the highs of good Class A or AB.
Since I'm looking for something with lower noise floor (which F5 turbo does not satisfy), it seems like my only options left are Modulus 86 OR either Wolverine or Honey Badger. But before I make any further choices or decisions or conclusions about that I need to spend perhaps at least another hour trying to find bad components in my Tandberg. I don't know how these two would compare, but they sound pretty similar on paper to each other and to my Tandberg. The Modulus 86 is a more unique choice which lends itself especially well to headphone use so it's also very attractive, but I'll have to weigh my choices. Any comments or insight into the differences between Wolverine and Honey Badger would be appreciated.
So today I finally had an opportunity to listen to class D amplifier monoblocks that are supposedly even better than Hypex Ncore 400 - they are from a company called Purify. It is highly resolving and the bass is straight up addictive. Haven't heard anything better from class A or AB. But it still did not meet my expectations and I ended up bursting my bubble of imagining that this could replace my Tandberg class AB. It despite being well detailed, sounded a bit dull especially at lower volumes. And the vocals especially seemed to have so little presence comparing to similarly priced Emotiva class AB monoblocks.
Tl;dr class D has amazing world class bass, but lacks the grippiness and emotion and immediacy, authoritativeness, authenticity especially in the mid-range, but dare I say also in the highs of good Class A or AB.
Since I'm looking for something with lower noise floor (which F5 turbo does not satisfy), it seems like my only options left are Modulus 86 OR either Wolverine or Honey Badger. But before I make any further choices or decisions or conclusions about that I need to spend perhaps at least another hour trying to find bad components in my Tandberg. I don't know how these two would compare, but they sound pretty similar on paper to each other and to my Tandberg. The Modulus 86 is a more unique choice which lends itself especially well to headphone use so it's also very attractive, but I'll have to weigh my choices. Any comments or insight into the differences between Wolverine and Honey Badger would be appreciated.
The 2SJ160-162/2SK1056-1058 are TOP3 versions of 2SK48-50/2SJ133-135, other devices such as the Exicon
are copies that are supposed to work decently, what is sure is that they are not the same devices, the Hitachi
ones have a negative temp coefficent at about 100mA while the Exicon are at a higher value in the 250-300mA
range, transconductance is comparable though.
Class D should be able to beat class A as it , like class A , is at its best at low power levels where the duty cycle is 50%. The TPA chips look really interesting.
A logical push-pull design would be to use independent feedback paths for the high side and low side transistors. But nobody seems even vaguely of the issue, except maybe @cumbb so we're all screwed, j/k.
In principle, a single-ended design can be thought of as an active transistor stage with a passive CCS load stacked on top or underneath. Now, to replace the CCS with a 2nd active transistor, a common feedback path messes it up for both. As one device momentarily has high current / low distortion, the other has low current / high distortion.
What everybody seems to do, is they stack the distortions together, and then try to linearise both transistors from the one NFB signal. Shakes head slowly.
Thanks a lot to both of you! I seem to recall that Hitachi used a term then that might resemble »lateral« MOSFET's, but I don't remember exactly what term it was.
Best regards!
Exicons have zero TC at 120mA according to the datasheets... But I haven't checked in practice yet.
More importantly, Exicons have quite a bit more transconductance, this is very good. See here: https://www.diyaudio.com/community/threads/uthaim-just-for-fun.306350/page-5#post-5338019
@Hotep
Elvee's "Circlophone" has a bit of a cult following on the forum. I have not heard it, but the circuit looks very interesting and MOSFET variants have been built.
Elvee's "Circlophone" has a bit of a cult following on the forum. I have not heard it, but the circuit looks very interesting and MOSFET variants have been built.
For the unique things that they are I'm going to settle with Modulus 86 and Wolverine most likely. I mean both of them, I can sell one later if I don't need it. Modulus seems to be one of the lowest distortion high fidelity DIY amps, plus it's relatively easy to make. I think I'm going to make this and the Wolverine. Make the Modulus so I finally have something to listen to, and relatively it's a lot easier and quicker to construct and finally it's cheaper than the Wolverine. Once I have this or perhaps even in parallel I think I'm going to start working on Wolverine, taking my time, not rushing, finding best parts for best prices etc. the reason I'm going with Wolverine over Honey Badger is that it is a newer project in the community after all, so likely improved by the most knowledgeable members here when comparing to earlier designs. And although it's a little bit more complex as a project than Honey Badger, if I have something to play with already (Modulus 86), I won't be in a rush to do it as quickly as possible, just enjoy the process and learn along the way (not worry about shipping times etc) and given its complexity there's probably a lot to learn. And it's a design that is more MOSFETs oriented than Honey Badger so far as I can tell. Ultimately it sounds like an amazing modern design drawing on the best knowledge of this community and hopefully it's going to.. maybe not leave my old Tandberg in the dust, but not play poorer than it either. I have high hopes for it. And while I haven't completely given up on repairing the Tandberg, I'm more inclined to make something from scratch, because without replacing pretty much most of the components in the Tandberg, it's just not gonna be what it's supposed to be at the end of the day, so no matter how good my repairs it's going to be quite flawed at the end (for starters, most if not all the resistors have completely wrong values, drifted values...). It was a bit noisy to begin with and I don't know if it's the cabling or damaged components or the design, but it had some kind of a warm veil, it's hard to describe, but it was simultaneously highly energizing and revealing and also kind of veiled... Oh and I'd have a suitable chassis and heatsinks, transformer with usable rail voltages, 15k uF ELMA high voltage power supply capacitors which would all reduce the final costs of a Wolverine build a fair bit, which is great.
Last edited:
I strongly advise first of all to gain experience in modifying the Tandberg 3026 in the field of simple matters, such as cables, connections, some components. You will find out how bad this device sounds - compared to itself. Otherwise, you would make the same sound-reducing mistakes in the design when realizing your power amplifier dream as this device has.
Having MosFets as power transistors is not necessarily the goal, I mean.
Which B&W do you have, or are you aiming for?
Having MosFets as power transistors is not necessarily the goal, I mean.
Which B&W do you have, or are you aiming for?
I'd be aiming for 702 or 703 s2 or s3 (haven't heard s1)(or 800 series if I strike the motherlode). I have something that would be closest to L82 Classic, but I'd say better. Same tweeter however, better cabinet and bass drivers IMO and hopefully in the end is gonna be better crossover (which is pretty rudimentary in the L82 anyway). Love that JBL tweeter tbh. But 700 might be better, haven't had the chance to compare side by side, but B&W is overall more detailed and in many ways more fun and emotional I feel like. I really respect those speakers since I compared them to 40-50k Focal system side by side and thought B&Ws both 700 and 800 (which are pretty close IMO) were way more fun to me.
Yesterday I also heard probably the best Amp I've ever heard to my clear memory with a super good DAC too. Everything connected to it sounded unbelievable. But it's just 30 watts per channel Class A. I dunno how B&W would feel about that, since everyone says they love lots of power, but I could potentially feed Class D to the bass and everything else with something like this Accuphase. BTW it's for sale for 5k. Given inflation it was around 20k euro new 10 years ago. Truly an astounding amp. Compared it to many other things even in 5-20k range and just wow, it's amazing. Japanese know how to do it.
Yesterday I also heard probably the best Amp I've ever heard to my clear memory with a super good DAC too. Everything connected to it sounded unbelievable. But it's just 30 watts per channel Class A. I dunno how B&W would feel about that, since everyone says they love lots of power, but I could potentially feed Class D to the bass and everything else with something like this Accuphase. BTW it's for sale for 5k. Given inflation it was around 20k euro new 10 years ago. Truly an astounding amp. Compared it to many other things even in 5-20k range and just wow, it's amazing. Japanese know how to do it.
Attachments
I have the datasheet for both Hitachi and Exicon, the latter are not well drawn but they show higher value than 120mA,
the typical transconductance is around 1S/3A for both brands, the difference is the dispersion of this parameter that range within
0.5-1.5S for the Hitachi and 0.7-2S for the Exicon, hardly a quite better gm, also max Ids is 7A for Hitachi and 8A for Exicon.
Last edited:
😉
I would like to thank you.
But my opinion - to all audio diy (!!!) interested (!!!):
I come from a different angle: once you realize that the same components sound different, that the ear is able to distinguish between the same components (same batch), that parts audibly modulate the sound, define it, you look at concepts like circuit designs in an other way: The nice ideal lines and numbers (frequency response, phases, power...) are zero statements regarding hearing and sound. And a simple consideration follows: if the ear can distinguish between identical components, what about complementary components? And what about half-wave separated amplification, amplifiers and power supplies (the majority in transistor analog audio), or channel-separated power supplies, for example? New parameters for audio amplifier (circuit) concepts follow: possible reduction of stages and possible reduction of signal separation. So what remains for the audio, hi-fi, high-end target is - just physical:
single-ended.
This is followed by classic pp (e.g.: https://www.diyaudio.com/community/threads/zero-feedback-impedance-amplifiers.42259/).
Unsuitable are symmetric complementary transistors PP concepts (e.g. Accuphase, Hiraga Classe A)
and finally non-symmetric complementary transistors PP concepts (e.g. Naim)
Nevertheless, you should not throw in the towel: the build up, choice of components and materials also have a significant influence to the sound. And for applications such as sports festivals or discos, you need power-saving crash makers. Unfortunately, these are sold at very high prices in the high-end segment, although simple PA amplifiers can easily keep up.
But it is a suggestion to think about short and half-wave separated feedback. There is certainly a lot to learn. It's been 20 years since I thought around. Unfortunately, I don't like dealing with complementary-transistor amplifiers. I reworked a NAD with only a 2-stage non-symmetric complementary-pp output, complex built up also, recently, but I always regret those wasted listening times when I connect a little non-complex SE again;-)
A real opportunity to learn a lot about audio and circuitry is John Linsley Hood's (JLH) 1969 (and subtypes) - one of the best audio (hi-fi, High-end) concepts, which offers three SE amplifiers and a SE follower. You can also quickly connect a compound transistor, or -transistors, or a CS or CCS or what else, and even feedback resistances here and there, an active power supply and so on: you're already in testing and learning heaven and will then have a different view of "amplifiers" and circuits;-)
(Some also call this concept PP, but I think it remains an SE, because the half-waves are not amplified symmetrically, and certainly never symmetrically;-)
Please forgive this somewhat more detailed text;-)
BD139 does sound not good.
My experience: Balanced circuits have the advantage that they cancel out the psu character, but the disadvantage that twice the number of parts modulate the sound. In the meantime I prefer unbalanced. But for a high power application I would connect two SE balanced.
How do the instructed Exicon MosFets sound - e.g. in comparison to simple IRF 140, 240...? My experience is that even the TO-3P housing sounds disadvantageous.
I have a friend who always wanted high-end - what he imagined it to be. After quite a while we finally realized that the whole block should hear him;-)
Trust me: this Accuphase is really steamy, and probably one of the best that Accuphase can build. The bigger and more Accuphase, the worse the sound. The complete k.o. was once a performance at some High End trade fair, where they gave every driver some super multi-way LS an own very big power amplifier, which was 12 power amplifiers, or more, whose power supplies were not even connected to each other. An acoustic disaster;-)
I would prefer even the smallest Accuphase Classe A, because Accuphase also is not able to place paralleling parts at ONE circuit, which is audible.