Wishing to learn about Current Dumping.

Status
Not open for further replies.
Where can I learn about "current dumping" amps? Are they linear? The term dumping doesn't sound linear.

It's a feedforward technique, but it needs negative feedback for it's implementation as used by Quad. It was best explained in a publication authored by J. Vanderkooy and S.P. Lipshitz's JAES article about "Feedforward correction in power amplifiers" - funny thing is that these guys live around the corner from me !

I've read the paper several times - it is very well written. My copy (pdf) exceeds the forum file size limit or I'd post it here.
 
Thanks to the mods for creating this thread on my behalf. I just asked this question in the Apex Current Dumping amp thread.

I got the paper on Feedforward techniques from Bigun - thanks. It's heavy on math. Going to take some time to understand.

It would be good if someone could summarize it in plain English in 1 sentence. Most amps can be explained in one sentence.
 
Last edited:
The Quad 405 current dumping principle relies on a balanced bridge.

Refering to the atttached image, with the suitable values of Z1, Z2, Z3 and Z3 (in simulation, use resistors first), the voltage at the output is not influenced by the voltage or current delivered by the high power class B amp.

Clear explanations can be found here :

FEEDFORWARD ERROR CORRECTION.

and in

"Sound Engineer's Pocket Book" 2nd Edition by Michael Talbot-Smith (editor)
page 2-133
.
 

Attachments

  • Quad_amp405 'Current dumping principle'.jpg
    Quad_amp405 'Current dumping principle'.jpg
    47.5 KB · Views: 553
Last edited:
Clever isn't it. The Elektor one is probably the easiest to get to grips with and the version using HEXFET's greatly exaggerates the crossover region. Look at the current in the inductor and the 22 ohm and how they all come together to give an undistorted output signal. Reduce the signal and all the output is supplied via the class A driver stage.
 

Attachments

  • Capture2.PNG
    Capture2.PNG
    75 KB · Views: 436
I got the paper on Feedforward techniques from Bigun - thanks. It's heavy on math. Going to take some time to understand.

It looks a bit scary at first but if you follow along with a pencil and scrap piece of paper you will find much of it drops into place. Once you have been through it and can understand what they're doing you will find you have a very nice understanding of the whole thing to the extent that you can answer most of the questions you'd come up with if you had only a simple explanation.

To be honest, I remain unconvinced that this approach brings a substantial sonic benefit but I appreciate the technical elegance.
 
Last edited:
post12
Clever isn't it. The Elektor one is probably the easiest to get to grips with and the version using HEXFET's greatly exaggerates the crossover region. Look at the current in the inductor and the 22 ohm and how they all come together to give an undistorted output signal. Reduce the signal and all the output is supplied via the class A driver stage.
This simple portrayal clearly shows that the ClassA amplifier has to respond quickly (and accurately) to sudden changes in output. That is quite an onerous task for any audio amplifier.

If the ClassC amplifier were replaced with a low bias ClassAB, would the "fill in" task of the ClassA amplifier be easier to accomplish accurately?

i.e. is the bridge addition philosophy a good way to get exemplary crossover reduction for any ClassAB power amplifier?
 
Status
Not open for further replies.