dvv, as always, it's all about the implementation. The Perreaux 2150B I have was a pretty straight transfer from the classic Hitachi design notes, but the mod's I did gave full rein to biting, and correct, sound - this MOSFET unit was the inspiration for all my adventures. At the same time I was checking out what the "opposition", in high end stores, was doing, and I spent most of that time shaking my head, thinking "What the ..." !! The Krells and suchlike were so thick, dead and plodding - the flaws in their sound within a system were so obvious.
What the Perreaux needed was a vastly better power supply - in raw form the "rounding off" came in far too early with rising volume; most of the work done was to cure that deficiency ...
Interesting examples, thank you. In all the amps' internals I've ever seen pics of and examined myself, I've never seen a busbar like this being employed for the supply. Its also of note the only hifi application featured in that brochure that I saw was for in-car audio.
The PS and all in series with it -- OPS --- have similar sonic affect when increased current capacity and lowered R/Z.
THx-RNMarsh
Yes, these would probably be too expensive in small quantities for many audio applications. Careful pcb layout can achieve similar results, although boards with thick copper for power circuits can be alarmingly costly.
There is also a limit on pcb copper thickness, and hence current capacity, due to problems with etching. Pcbs are much more fragile than a laminated bus, so there are problems with shipping g forces, fastener torque, and heating, when attaching large, heavy power components directly to them.
That link is an excellent reference - the example on page 3 of it gives the flavour of my gainclone: take the righthand end, the capacitors as shown, the upturned section is the chipamp, and the semiconductor fuse is the heatsink. The fun bit was doing the vias for the caps, etc - as painful as your 3D creations, 🙂.
Nothing like a bit of DIY, 😀 - I got a raw sheet of decent thickness copper from a metal supplier - lovely stuff - chopped out some rectangles, created a layered cake with uncoppered board, and squashed it all together. Of course it could have been done better with thinner insulation, but the end structure was very rugged, stiff - no problems with mounting things.
Those bus bars used to be common but I have not seen them on PCB's or in the surplus bins for years. They may not be too compatible with SMT processes. A big wave solder machine would have no issues soldering them but those machines are less common today.
Fabbing ones with sheet copper (is it OFHC or 6-9's?) with Teflon tape would probably work pretty well. The Constellation amps have big buss bars but not in laminated pairs.
Fabbing ones with sheet copper (is it OFHC or 6-9's?) with Teflon tape would probably work pretty well. The Constellation amps have big buss bars but not in laminated pairs.
For many, sure but presumably not all. For an example I think could benefit from such an approach and probably not break the budget here's a 22keuro offering from Karan which doesn't look to be optimized for low inductance - 6moons audio reviews: Karan Acoustics KA M2000
View attachment IGBT.pdf
Might be worth getting some of the SIM guru's around here to make a composite from stock parts for a IGBT.
THx-RNMarsh
The use of low Z busing, using flat, wide conductors with thin film between was first shown to audiophiles in 1982. See pg 2 -->
View attachment MC1.pdf
View attachment MC2.pdf
View attachment MC3.pdf
THx- RNMarsh
View attachment MC1.pdf
View attachment MC2.pdf
View attachment MC3.pdf
THx- RNMarsh
Last edited:
[All rights are reserved by the author -- me. So that I may publish it here or anywhere]
That's back when i was using some old flexible PCB material for speaker cables. About 6" X 24" would make it from amp to speaker. Taught me a lot about amp stability.
40 More years of learning -
Hi Demian,
That was a long time ago.... 32 YEARS. 😱 OMG ! It was new then and still new now and the principles are still good if you want the best sound/performance.
Since then i got myself some good test equipment and learned more. 30 years more.
THx-RNMarsh
Hi Demian,
That was a long time ago.... 32 YEARS. 😱 OMG ! It was new then and still new now and the principles are still good if you want the best sound/performance.
Since then i got myself some good test equipment and learned more. 30 years more.
THx-RNMarsh
Last edited:
There are a myriad of techniques and sensible things to do, for improving the quality of the system used for playback - it's just a case of applying them, each and every one of them if necessary, until finally the electronics work well enough for 'good' sound to appear ...
MC3.pdf:
Gee, written 32 years ago - gosh, we've come a long way since then, eh ... 😀, 😛
Nice one, Richard, ... 😉
Gee, written 32 years ago - gosh, we've come a long way since then, eh ... 😀, 😛
I have no iea myslef, Demian, but that's the way I receive it.
I should have added that pretty well the same thing happens when I take out a Japanese TO-3 device and stick a Motorola in. The sound becomes meatier, has more gravitas, and overall appears to be better tonally balanced. This also happens quite often when I replace a device with a BD power transistors.
This begs to be viewed as a generic difference. Even when not all, but still most, MJ and BD transistors deliver better bass lines than most Japanese power devices. They make them differently - exactly how, I don't know, this sort of reminds me of the difference in firing sequence between American and European V8 engines. Detroit V8s have a sound I'd kill for, deep down throaty, almost menacing rumble, but European V8s are constructed differently, are way more efficient, are much more easily ran at high revs, but lack that sound. Yet both types work just fine.
Same thing with MOSFETs. Nominally, they do just fine, measure them and they will do the whole gig, but in the end, they are a little too polite for my taste.
No doubt, Frank, that much of it is due to implementation, which indeed can make or break an amp.
BTW, your Perraux was one of the few MOSFET power amps I did like and would have liked to own. If that's the model which was current in 1984, when I heard it in London.
I like fairly straightforward implementations, no fancy circuits, just a straight predriver-driver-output line, with base resistors thrown in to inhibit misbehavior. Logically, with MOSFETs, I would have no need for any predrivers, therefore a simpler signal path, therefore theoretically a cleaner signal path.
I agree with that in far too many cases, MOSFETs are endowed with inadequate PSUs, for unknown reasons people seem to think they don't need large power reserves - obvious folly.
I own a €5,600 Karan integrated amp, model KA-i180. It's a model from 2003, if memory serves. Rated at 180/250W into 8/4 Ohms. Also, Milan Karan is a friend of old and a favorite coffe drinking companion.
That notwithstanding, my model is clearly optimized for higher impedances. While it will drive my AR94s with aplomb, it is not too happy to do so. With my B&M 1041 speakers, it's a match made in heaven, and I have yet to hear another integrated amp, from anyone, which can better it.
The PSU consists of parallel 10,000 uF caps, 2 per power line, 8 in total, from Fisher & Tausche, in my sample in original attire (blue coat with black writing), later on with Milan's custom packaging (black coat with gold writing, says "Karan Acoustics", but same thing). The full wave bridge rectifiers are made from discrete 1N5406 diodes. The output stage is in full balanced topology, using 4 Sanken 200W devices per channel.
Now, with all my views on soft bass sound in mind, I have to admit that his bass lines are way better and more convincing than most.
The problem seems to be all NPN is required. It isn't a problem. I have a suspicion one of us here makes his own in compound pair. Giving it more thought such a pair might be what Quad wanted in a simpler form. The compound version can be complimentary and up to 250V. It should really kick. If a tipple of Bi FET Bi it might be awesome. I supect too in my face for me. I would use the Motorola TO 220 drives if so,and Japanese ultra fast finals. Personally I would vote for F + Bi for it's multitudal simplicity. If a complimentary feedback pair I don't know. I supect a Fetlington is better if I want it to kick.
Did anyone one try Quad 405 in MOS FET ? It should work like a dream. Far better a gradual switch on than abrupt.
JLH was commissioned to do that amp I think?
- Status
- Not open for further replies.