Ncore design refinement

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The previous post showed a prototype which self-induced a loudness boost of the loudspeaker's bass; because of the application of positive feedback.

The attached schematic shows another prototype using the same components which self-induces a loudness cut of the loudspeaker's bass; because of the application of added negative feedback. Please note:

1. The magnitude of overall negative feedback which is inherent in the parent UcD180LP was increased; predominantly in the bass region. This has consequences; meaning [in part] a decrease in the magnitude of UcD's [Vout] at a constant magnitude of input signal [Vin]. Aka a cut in loudness.

2. The signal [V feedback] which develops across the sense resistor [0.2 Ohm] drives the non-inverting port of the Op Amp [OPA134A]. It has a line on its shaded portion to distinguish it from [Vin].

3. Assume that the output impedance of the source's input signal [Vin] is low [say ~100 Ohms]. The Op Amp amplifies [V feedback] by ~a factor of two and does not invert its phase; because it is fed to its non-inverting port.

4. Simultaneously, the input signal [Vin] is amplified by a factor of 1 and its phase inverted; because it it fed to the Op Amp's inverting port.

5. The overall consequence is the Op Amp subtracts [or differences] the non-inverted [V feedback] from the input signal [Vin]. It follows that the output signal of the Op Amp [same as input signal to UcD] was decreased in amplitude. Ditto for UcD's [Vout].

6. This prototype has a self-induced and increased overall negative feedback which is pronounced in the woofer region. Its impact on the inherent [and impressive] % THD and output impedance of UcD is unknown.

Best
Anton
 

Attachments

  • UcD180LPbasscut.jpg
    UcD180LPbasscut.jpg
    391.9 KB · Views: 850
I must apologise I won't be able to join the discussion further after this post, because things are a bit busy here. Meaning that if anyone misunderstands what follows, I shan't be here to rectify matters :)

The amplifier is not "DD". It's a straight self-oscillating fully analogue design, like all of my work of the past 15 or so years. All that time I've been saying that this is the best and most elegant way forward, and I daresay the new circuit confirms this. We've decided to call the technology Eigentakt, which is an obscure German term for self-clocking or self-oscillating. That would have been a giveaway, had Peter thought to mention the word.

I guess folks were assuming that digital was involved simply because the person to break the news is Peter Lyngdorf and he's of course historically been involved in open-loop "PWM power DACs", which is what I suppose you mean by DD. And of course, the technology historically used by Lyngdorf was developed by Lars Risbo who's my twin brother in this new venture, Purifi. He too feels no particular need to stick to digital control for its own sake. Being an engineer like me he's happy to use whatever seems the optimal solution at a given time.

So what's new in this amplifier? One thing is that I've developed a sampled domain model for self-oscillating loops that remains valid for all duty cycles, so it can predict closed loop response exactly under all conditions. This then allows finding a loop design that has very high loop gain (about 75dB at 20kHz, which is 20dB better than my previous designs) without running into stability problems near clip. A design procedure like this can't be patented because you can't prove that someone has been using it, so in a break from past style I'm not going to publish any details of this mathematical model. It's a trade secret, plain and simple.

Then there is the loop structure that allows better control of the closed loop frequency response. My previous amp* has approximately a 1st order roll-off. Now, since the output filter naturally has a second order roll-off it means that this amp could be overdriven with out of band noise from e.g. DSD recordings (in fairness, only when you cranked a quiet recording high). Also a first order response already droops noticeably by 20kHz which was sometimes remarked upon by numbers people. Eigentakt has a well-defined 2nd order response which stays dead flat in the audio band and rolls off in a fully controlled manner with a sensible -3dB point of 60kHz. The loop structure used to get this behaviour is subject of a patent application. A few other items are not really germane at the moment but were worthwhile enough to put in a separate patent application.

From a practical perspective, the enormous loop gain is partly used to allow more relaxed timing of the power stage without paying for that in distortion performance. As a result, the 400W power stage only has an idle loss of 1.7W or so. From dire experience I've learned that people don't perceive an amplifier as "cool running" if the idle losses aren't super low. It seemed a sensible move to get that sorted for once and all. In other words, the measurements you're seeing are those of a power stage actually optimized for idle loss (and hence, efficiency under real listening conditions).

Those asking "where are the patents", patience. It's a long procedure. First you file an application that gets bounced back and forth between the inventor and the patent office (with a patent lawyer in the middle), and 18 months after the first filing you see a published patent application. No sane inventor publishes details before that - those 18 months are actually there so you can take advantage of legal protection without having to tell your competitors your secrets. In other words: you will never find a patent for download when a new technology is announced. Of course, in sales communications one will easily say "patented" but actually it would be more correct to say "patent applied for" which should be enough of a heads-up for potential copycats.

Finally a note on how the various companies fit together. Purifi ApS was founded by Peter Lyngdorf, Lars Risbo and myself. No shares are owned by any of the other companies we're involved in. Purifi is run independently and will eventually supply technology to a broader base of customers. But it should be obvious why our lead customers would come from our own orbit...

_______________________________
*Forming a new company of course involved quitting my previous position, which I'm now competing with. I think it's only polite not to bandy around any of their trade names, in particular since in private we're still friendly.

Congrats Bruno on a great achievement. Can you share the make/type of the mid woofer?

thanks,

Stefan
 
He obviously did not share it but he certainly will tell us more about it on High End Munich 2019 on Purifi Audio seminaries. We will get more to know also about his NCore refinement which is indeed much of refinement, as it boasts more loop gain especially in highest frequency range (or even supersonic) which is crucial from end-side performance. Of course, together with expected performance of 6.5-inch/15 mm-Xmax woofer.. I hope that he remembers about frequency modulation distortion along with this. Long-stroke midwoofers trying to reproduce good bass and good midrange together, always fail to sound right.
 
This is reasonable but limits its practical implementations. I am curious how he will address it. Hope the depth and weight of this woofer will be reasonable to justify Sd decrease from say, 8-inch equivalent, possessing more relaxed stroke performance. I'm sorry for this off-topics.
 
Info about new amplifier is available now: - 1ET400A - Purifi

Thanks miero for the link. This amp by Purifi has an impressive performance. Congratulations. A first for me to read about an output impedance in micro Ohms!

The attached schematic is for a prototype power amp using UcD180LP. I continue to pursue potential refinements in subjective performance. This approach can be adapted to other amps like the Hypex Fusion, Ncore, Class T etc. I have shown earlier a schematic like it in the context of positive power feedback. This application is a bit different. Here is an explanation of its works.

1. The BJT power amp generates a power output signal [Vo] across the left 12 Vac winding of the toroid power transformer.

2. Simultaneously, [Vo] is attenuated by the resistor string [50 Ohm plus 25 Ohm] so as to present input signals [Vin'] and inverted [Vin'] to UcD.

3. The power output signal [Vo'] of UcD is applied across the right 12 Vac winding of the power transformer.

4. The magnitude of [Vo] and [Vo'] are made to be equal in magnitude by tweaking the 25 Ohm pot.

5. The phase of [Vo] and [Vo'] are aligned in the frequency range of 40Hz to 250 Hz. Done by differential measurement with an AC voltmeter.

6. The status of this combination amp sets up the condition of "bootstrapping". Thus, neither amp dumps power into the other's power output node. Both amps pay attention to the purity/quality of the signal across the loudspeaker load. BJT amp is the sentinel. It does minimum work in the loudspeaker compared with all the work done by UcD.

7. The toroid power transformer is like a mirror. [Vo] and [Vo'] are forced to be ~mirror images. It allows the addition to or subtraction of minimal offset power from the loudspeaker as needed. Put the AC voltmeter across [Vo and Vo'] while playing music. The voltage difference is not the theoretical zero!

8. Sounds superb. A low level high frequency oscillation from UcD appears at [Vo and Vo']. It may infiltrate the circuits of BJT and influence the overall performance.

9. BJT and a Threshold S/150 are the reference combination amp. This system does not have the switching frequency at the outputs as in point 8. This prototype sounds superb like that of BJT/UcD.

Enjoy.
 

Attachments

  • BJTUcD.jpg
    BJTUcD.jpg
    441.3 KB · Views: 539
Can UcD180LP be coerced to subjectively sound like another amp; say a BJT-based Class AB ? The attached schematic is for a prototype "compound" power amp which attempts to tackle this hypothesis.

This amp operates trouble free [no externally induced oscillation] because of the following two properties which are inherent in it:

1. The voltage gain of UcD is 13.1 dB or Vout = 4.52 times its Vin.
2. The chosen toroid transformer has a voltage step-down ratio of [115/24]=4.79

The operational trick is this ratio of 4.52 to 4.79 = 0.94.

The resultant stepped-down voltage output of UcD will always be 0.94 the output voltage of BTL. BTL is 94% voltage-shadowed [boot-strapped] by UcD. Thus BTL does little [6%] work versus UcD which does it all; up to the specified 180W/4 Ohm.

BTL is low-power and maybe Class A. Its differential power outputs drive UcD directly as required without the use of OpAmps as I showed in my previous posts.

This prototype is amenable to an A/B subjective comparison. The schematic is [say] for setup A. Setup B [not shown and suggested] requires a mechanical relay to simultaneously disconnect the primary and secondary windings of the transformer. Setup B enables BTL to drive the inputs of UcD which solely drives the loudspeaker without interference [for better or worse] from BTL in setup A.

The case of a Class T BTL instead of a BJT BTL in the compound power amp will follow in a next post.
 

Attachments

  • BTLUcD.jpg
    BTLUcD.jpg
    566.7 KB · Views: 627
The application note for UcD180LP by Hypex Electronics specifies 180 W into a 4 Ohm load with an attendant 1% THD. This level of distortion is expected to be lower in the prototype compound amp; described as follows:

1. Distortion is an electrical signal which is stepped-down from the power output of UcD to the power outputs of BTL.

2. BTL recognizes this distortion [it sez not mine!] and/but still corrects it anyway and instantly via its negative feedback loops.

3. The resultant cleansed signals at the outputs of BTL are forwarded to the inputs of UcD so as to continually effect a lowered THD content in the region of its maximum power output.

3a. The absence of distortion signals in the 24 Vac winding driven by BTL says it will also be simultaneously absent in the 115 Vac winding driven by UcD. Transformer action in action!.

The first Typical Performance Graph [7.1] in the Application Note is entitled THD vs. Power (1 kHz, 4 Ohm). The above 3 points suggest:

1. The shape and/or flatness of the graph is expected to be different; for better or worse; especially vis a vis the parent absolute THD which is superb.

2. The line in the graph at ~150 W rises sharply to indicate the onset of increased levels of THD. The slope of this line in [7.1] is expected to be diminished in the compound amp. Thus, the resultant graph/line continues with a gentle ascent.

Best
Anton
 
Hello Reactance and ds23man,

Thanks for your honesty. Healthy skepticism on your part is +1.

I hope that you consider experimenting with this compound prototype. Its assembly is easy and highly detailed. How will anybody learn its "truth"; if any?

Quite perceptive of you Reactance. I have a Philosophy Doctorate; but in Inorganic Chemistry. It spilled over.

Frankly, my DIY objective is to understand the value to performance for this suggested compound prototype; by posting or writing it out loudly... It'll help me unravel its audiophile sense or nonsense.

Best
Anton
 
Disabled Account
Joined 2005
I am a Bruno's fans since 2005.

Salute to you. Here is my DIY Amplifier, UcD fullbridge, which tested on live music, to drive 2 x 18" speaker with no problem.
If you know him, then you know that his is consistent to design class D amplifier as self oscillation post filter feedback.
Even very poor design and poor material like myself as above still amazing result I can get. Many thanks to Bruno
 

Attachments

  • UcD Fullbridge.jpg
    UcD Fullbridge.jpg
    186.5 KB · Views: 547
Last edited:
Is this new refinement going to be available in diy kits?

I asked Lars at the AES presentation about the driver and amplifier. He said both the driver and amplifier will become available for DIY market eventually.

The goal for the amplifier was by the end of summer and for the driver it is the end of the year.

Personally I would expect some delays though personally due to their significant performance improvement. I am more curious about the price range for the drivers.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.