It's really not worth discussing these things with TB; he's got a strong tendency towards ideology and NIH (Not invented here) that gets obfuscated by his use of engineery language. Or at least prepare yourself for frustration.
jan.didden. Here is the link:
Quad 405 : analyse d'un schéma
QUAD 405 analysis. In Audiophile n°1, octobre 1977
Enjoy.
Quad 405 : analyse d'un schéma
QUAD 405 analysis. In Audiophile n°1, octobre 1977
Enjoy.
Thanks, that's useful. But I was also interested in the blind testing paper as mentioned by Marcel.
Jan
My French isn't great but it seems to me that Fig 8 is discussing a voltage amplifier, G, driving the transistors. The Quad is described by Vanderkooy and Lipshitz as using a transconductance amplifier (see below). So I am not sure what the Quinquis article is trying to show.
In the V & L model it can be seen by inspection that the output impedance changes when the transistors start to conduct (because they add current gain). It can be shown by simulation too. A constant output impedance being caused by using a bridge arrangement is not correct IMO.
Attachments
Definitely prepare for frustration...I don't take prisoners and I don't put up with crap regardless of who generates it. Well, most of the time.It's really not worth discussing these things with TB; he's got a strong tendency towards ideology and NIH (Not invented here) that gets obfuscated by his use of engineery language. Or at least prepare yourself for frustration.
Please elaborate on the false premise. Is it like the diff of two digital files being null but they still sound different?
Thanks for trying this out. I think I need to see what your models are to comment usefully. The linear amp needs to be a transconductance amp. The amount of NFB needs to be the same in both.
An ideal output stage should have a constant output impedance. The challenge for the creators of the current damping topology was to achieve this with a class C ouput stage (chosen for not needing bias setting) delivering most of the power. The result may not be absolutely perfect but it is there and it works remarkably well since more than forty years, being appreciated by a huge amount of people.
I do not know an amp output circuit as ingenious as the one signed by Walker and Albinson. I think they were not obsessed by the highest performances. More recent circuits prove that current dumping with output power transistors biased in class B can give an impressive linearity.
No, I only have a second-generation paper copy made in the early 1990's. Of course I could scan that copy for you, if you like.
Believe me, I am looking for examples of genius in the Quad 405. Please suggest some more! I certainly agree that avoiding bias setting is a manufacturability advantage and the story that Quad told was excellent, if not genius, marketing. The 405 was never a sound quality leader but Quad sold a lot of them.
I like to examine whether theory of operation claims are true.
The "current dumper" idea seems to generate a lot of disagreement.
For a start there is no feed-forward involved in the proper sense. This is very much a NFB system. In the V & L simplified model, the output impedance is the result of NFB and is not linearized by the "bridge" arrangement. Regarding distortion, the perfectly balanced bridge makes it as good as conventional feedback topology and unbalanced makes it worse. There must be another reason they chose a bridge but I haven't figured that out yet (unless it was just for marketing).
The actual circuit Quad built is very complicated. They appear to have added a number of extra components that appear to be for stability reasons. It really needs to be simulated to see what it actually does.
The current dumping principle is quite clever. Walker and Albinson had to figure out that one can make a bridge circuit with an inductor, so the low-frequency output impedance and the power losses remain low, and that the path from the dumpers through the inductor to the output could be largely cancelled by the other path with a resistor, the integrating class-A transimpedance amplifier and another resistor. In fact finite DC gain of the transimpedance amplifier can be cancelled with nonzero DC resistance of the inductor.
Although it has nothing to do with current dumping as such, the QUAD-405 DC bias loop is also interesting. I haven't seen anything like it in any other audio amplifier. It results in an almost exact second-order Butterworth high-pass response.
Although it has nothing to do with current dumping as such, the QUAD-405 DC bias loop is also interesting. I haven't seen anything like it in any other audio amplifier. It results in an almost exact second-order Butterworth high-pass response.
Hey Scott, it is remarkable what people will believe. Make sure you store your audio files on the correct type of storage device and remember that the more times you copy it the more dilute it becomes and the stronger it gets.
The false premise is that if you cannot hear the extracted residual distortion on its own (played through the chosen detection system, etc.) then you cannot hear it in the presence of the source music (played through the chosen replay system, etc.).
Last edited:
Makes sense to me, you can't hear something so mix it with something else you still can't hear it. False premise proven by whom, how?
Very difficult.
Sound engineers haven't emitted much criticism towards it
Its commercial success seems to irritate you even more than its topology.
And even more of applause.
I have studied the error correction topologies. As far as I can understand them, there is always some kind of negative of feedback in their scheme.
I agree with that.