@Dumpster, it may be useful to first understand how a current dumping amplifier would work without the "bridge" - that is, with the smaller resistor of the bridge open, and the inductor shorted. With these mods, it is a normal feedback amplifier with an under-biased output stage. Once its operation and the role of feedback in it are clear, one can contemplate the role of the "bridge". Otherwise, attempts to build an intuitive understanding of "current dumping" lack the foundation and become quite difficult.
Shall we hold a class? The only problem, the topic is a bit abstract...
Alex, I don't think it is abstract. Take the normal forward path of an amp. Nobody says: 'first the input gets out of the preamp, next it goes to the driver and a bit later it enters the output tube. Everybody understands that input signal, preamp output, driver output amd output signal are all there at the same time in a continuous situation. Nobody says it is 'chasing forward'. Only with feedback all of a sudden we say it is 'chasing around'. Maybe a lack of imagination?
Jan
Jan
Oh yes, quite a few people do. I have seen enough discussions of multiple copies of a signal chasing each other around the feedback loop in the fashion of the famous Zeno's paradox of Achilles and the tortoise.
What I mean is that not many of us would be excited to walk away with just an understanding of how something works. Something tangible - say a working current dumping amplifier - would be much more attractive I think.
You're probably right.
But I would be more interested in learning how to catch fish instead of purchasing or being given a fish. 😎
Jan
But I would be more interested in learning how to catch fish instead of purchasing or being given a fish. 😎
Jan
Just re - reading this thread - in it's entirety, as there must be something I have missed in the responses to my incessant doubt/lack of understanding.
Jan - from above you said, ''But when the dumpers start to conduct, additional feedback from Z4 (the impedance in series with the dumper output) is generated.
This is significant: in any 'normal' feedback amp, the feedback factor is constant (it's normally a two resistor divider).
But in a CD amp, the feedback factor changes depending on what the dumpers do!''
Is this not the same thing that I am banging on about, just I have it ''Arsy Versie'' - ''additional feedback from Z4 (the impedance in series with the dumper output) is generated.'' - i.e. the Trigger/Transient to re - enforce Dumper Switch - on.? - thinking out loud, the initial field generates the Pulse/trigger, but quickly disappears as the field collapses? But it can't be negative as that would drive the dumpers off - would it not. So, is it an initial '' Positive going Transient'' at the point of turn - on, which quickly collapses, leaving the class A feedback, plus additional feedback from the now conducting Dumpers, to control the amplifier, in its entirety?
At this point the ''feedback factor change depending upon what the dumpers do'' - The factor being determined by the audio signal driving the CD Amplifier.
Last edited:
I get your drift Alex - however, how the CD works is key to how all current dumping amplifiers work. The World is full of people who simply take delivery of boxes and have no interest in the technology. We British are keen on saying we did this, invented that, which simply isn't true. Some brilliant individuals develop or invent things, but we as a Nation claim ownership, when as I say the vast majority know nothing about what is inside the box.
If we are going to advance, more and more people need to understand about the way things work. Become less hedonistic, less superficial as a Nation.
What trigger/transient? It is a continuous process! As the signal level rises, at some point the base voltage on the dumpers rises and they start to conduct current and extra feedback through Z4 starts to occur. Don't take my 'starts' literally - a continuous process, changing with the signal level.
You think much too much of it.
Jan
What is the point of asking meaningless questions?
if you want to see how it works, put a scope probe on the base of Tr10. You will see a very rounded-over sine wave, the rounding being due to Tr10 turning on and taking over the heavy lifting. No switch-on transients.
Last edited:
'Current Dumping' is a marketing term.
I can define it anyway I want, and you can define it completely different.
Don't see it as a technical term - it isn't.
Jan
I can define it anyway I want, and you can define it completely different.
Don't see it as a technical term - it isn't.
Jan
I'm guessing this topic is now closed? I own a 306. I'm looking to refurbish it, as I was disappointed by an overly warm sound, lacking air and transparency in the upper frequencies. Not sure what could cause that, or whether it is the sonic nature of the amp?? Also, am I correct in assuming that the circuit topology ensures that the crossover glitch is reduced throughout the amplifier's power range? Or, is it like a class ab amplifier, where the output is pure class b after a certain point? Thanks in advance.
There is no crossover glitch, because there is no crossover. It is class A up to a certain point and then class C, with the class A filling in the gap. Please start your own thread, don’t hijack this one.
Which is what I have been driving at all along, but perhaps confusing myself and others in the way I described it. But the area which I don't understand is the region where the dumpers cut in to take over to supplement the class A section by topping up the power. The speed at which they cut in/turn on, must very quick to maintain low distortion.
This is why, rightly or wrongly, I implied "Trigger", because of the speed of operation. Decades ago, I commissioned some Broadcast Transmitters and they essentially operated in Class A up to 60% of full output - 200 Kw PEP. Above this level, a Peaking Valve came into operation. This improved overall efficiency, saving considerable money for the broadcaster.
Indeed, this was why my interest in the Quad 405 was pricked. The trouble is, with the broadcast transmitter, I could adjust the conditions to ensure the peaking valve cut in whilst maintaining low distortion. Not so with Peter Walkers design and the key is the bridge and how it achieves it.
I am afraid no one, as yet has adequately explained it and it remains just as Jan suggests - like buying a fish.
The output of the class A amplifier needs to slew faster when its voltage is in the class B amplifier's crossover notch. The required slew rate is inversely proportional to the amount of feedforward (i.e. 20% feedforward needs 5x faster slew rate). This is why current dumping never caught on.
Ed
Ed
OK, thanks for your input. Do I therefore read you correctly by saying the function of the dumpers is fixed and it is the Class A section that adjusts to minimise the error at the point where the output stages comes into service? Effectively adjusting the phase of its output to align with the dumper as they turn on?
At higher drive levels/power output, I can see; I think, that the response time needs to be faster due the the greater rate of change.
At higher drive levels/power output, I can see; I think, that the response time needs to be faster due the the greater rate of change.
Somewhere on this board is a simulation showing the output waveform of the class A amplifier. The simulation shows the class A output moving very quickly while in the crossover notch, and at normal speed once outside of the crossover notch.
Ed
Ed
So, the Class A amplifier; in an ideal world, would need a response time/bandwidth approaching that of an RF Amplifier? And if so, this limits the ability of the design due to other complications an unrestricted amplifier would introduce?
An audio amplifier needs roughly 5V/us. Five times that would be 25V/us, which most amplifiers can do, but why use up the slew rate when a normal class AB output stage does not need that.
Ed
Ed
So as brilliant; and revolutionary, Walkers design was at the time, it is inherently a redundant technique.
Thank you for your incisive contribution. Much appreciated, as is that of all contributors. Redundant or not, I remain astounded by the performance of the 405 and in particular the 909. I still cannot justify a Sugden as an alternative. However, I am happy to accept what differences there may be between them.
The Grip the 909 has on the performance of my BC1's is something I haven't experienced before, so it would take something really special to upstage it.
Now however, I have to find a way of limiting low frequency excursions by way of supplementary bass drivers to protect the BC1's from being overworked in that region.
Many thanks and kind regards,
Brian.
Thank you for your incisive contribution. Much appreciated, as is that of all contributors. Redundant or not, I remain astounded by the performance of the 405 and in particular the 909. I still cannot justify a Sugden as an alternative. However, I am happy to accept what differences there may be between them.
The Grip the 909 has on the performance of my BC1's is something I haven't experienced before, so it would take something really special to upstage it.
Now however, I have to find a way of limiting low frequency excursions by way of supplementary bass drivers to protect the BC1's from being overworked in that region.
Many thanks and kind regards,
Brian.
Sansui super feedforward claims to reduce THD, as well as crossover distortion purely based on using the feedforward technique. Sansui use a small correction amplifier. It is claimed that the low 0.004% THD @ 1khz is maintained to full power. It is said that the circuit topology is totally different to Quads CD. Is the Quad CD low THD due mainly to the feedforward part of the circuit, the class A amplifier just being a correction amplifier?