Feedforward Error Canceling with Blameless Topology

[hbtaudio]

And there's nothing wrong with that. I'm learning a lot from reading your friendly banter back and forth.

My strange manner is due to my nature and a book whose title and subject matter still inspire me today.

Die tanzenden Wu Li Meister
[Gary Zukav]

“The most exciting intellectual adventure I've been on since reading Robert Pirsig’s Zen and the Art of Motorcycle Maintenance.”
—Christopher Lehmann-Haupt, New York Times





Gary Zukav’s timeless, humorous, New York Times bestselling masterpiece, The Dancing Wu Li Masters, is arguably the most widely acclaimed introduction to quantum physics ever written. Scientific American raves: “Zukav is such a skilled expositor, with such an amiable style, that it is hard to imagine a layman who would not find his book enjoyable and informative.” Accessible, edifying, and endlessly entertaining, The Dancing Wu Li Masters is back in a beautiful new edition—and the doors to the fascinating, dazzling, remarkable world of quantum physics are opened to all once again, no previous mathematical or technical expertise required.

In 1985, when I was a freshman in college, this book was our favorite, but so was The Hitchhiker's Guide to the Galaxy by Douglas Adams.

You may have seen the blameless-inspired design I posted here a few days ago. It came out sounding great even though I barely know what I'm doing.

👍


greetings,
HBt.

 

[jxdking]

Alternatively, you can create a separated feedforward path with a unit gain buffer. Shown as below, the class A unit gain buffer is created with CFP that has very low distortion.

 

[hbtaudio]

You can also approach the classic feedforward error correction with the now more formed cascade. Two more stages and nodes should be enough to theoretically (but never practically) reach zero % THD.

@jxdking
Please replace your ohmic 8 Ohm load with the LS dummy load with a little EMF.
Not much is left (then) of your (initial) -98dB THD, = 0.0013%. Please go to fo=3kHz and Po=5Wrms, also consider the component 9 * 3kHz.

Many thanks in advance,
HBt.

 

[jxdking]

You can also approach the classic feedforward error correction with the now more formed cascade. Two more stages and nodes should be enough to theoretically (but never practically) reach zero % THD.

The distortion from CFP in class A is very low. The only things lower than that are OPAMPs. Speaking of OPAMP, an OPAMP with bootstrapped power rails might work in that position. The limitation would be the current capability. (Perhaps, I could current dumping that OPAMP, too)

Please replace your ohmic 8 Ohm load with the LS dummy load with a little EMF.
Not much is left (then) of your (initial) -98dB THD, = 0.0013%. Please go to fo=3kHz and Po=5Wrms, also consider the component 9 * 3kHz.

I am using 10KHz.
I might go 20KHz.

I will look into LS dummy load later.

 

[hbtaudio]

I am using 10KHz.

Oh, I see - my fault; fres is 1kHz, ok.

I might go 20KHz.

Strange,
I personally am not interested in the THD20k because of the components 40kHz, 60kHz, 80kHz, 100kHz, 120kHz, 140kHz, 160kHz and 180kHz, which are outside our hearing range, or are you afraid of backfolding?

We haven't even included modulation products in the overall consideration yet, have we?

THD3k (5Wrms) is a good marker!

I will look into LS dummy load later.

Yes please, absolutely.


regards,
HBt.

 

[hbtaudio]

components 40kHz, 60kHz, 80kHz, 100kHz, 120kHz, 140kHz, 160kHz and 180kHz

To capture all of these correctly using an FFT (in realtime), you need a sampling frequency of 1.8 MHz - or is 360 kHz enough for you personally?

😉

 

[jxdking]

Yes please, absolutely.

Dummy speaker based on your #42. Not sure how to implement EMK...

 

[hbtaudio]

Let's put your SE-CLASS-A extension to the test: but in such a way that a comparison with the original expansion stage is also possible 😉

Your triangle should remain as it was.

 

[hbtaudio]

whitout EMF

EMF 1Vrms

Start and stop times of the analysis optimized!

In reality, your THD correction miracle circuit will be no better than:

-63dB <---> -81dB

 

[hbtaudio]

Since the circuit is not complicated, I suggest that you try a test setup and take real measurements. I hate to say it, but: error correction does not work as sketched here!


kindly,
HBt.

 

[hbtaudio]

Dear jxdking,
I would like to leave the discussion now and make room for other discussants.

Kind regards to Marietta.

HBt.audio

 

[jxdking]

It seems to be a common conception that you can get a blameless architecture amplifier down into the 0.005% THD range and below quite easily just by careful components selection and with very modest bias currents. What's broken that we're trying to fix with this?

That is true, just as mentioned in Douglas Self's book, the blameless with CFP has lower THD comparing to EF2. As below, with MJL3281/1302, it gets 0.005% without any mod.