NC252MP (class D) vs. A250W4R (classAB) burst measurements into 4ohm//2.2uF load

[AdamG]

Well if the amp isn't running hot... then all seems well in oscillation world, hence why everything is peachy I'd suggest.


JSmith

Amp runs around +/- 8-10 degrees Fahrenheit above ambient room temperature. 6 channel Amp draws average of 74-76 watts under normal load (24-25 watts per 2 channel 502 Amp module). Measurements derived from Top mounted rear venting Aircom cooler and a Fluke Multimeter validated Remote Power Monitoring switch.

 

[Matias]

NC252MP comparison of THD+N vs. power at 1kHz into 4ohm resistor load and 4ohm resistor in parallel with 2.2uF capacitor load

View attachment 273978

The "ASR chart parameter" at 1kHz/5W falls from -91.4dB to -51.0dB, 100 x higher THD+N with 4R//2.2uF load at 1kHz. Still "fine"?

This is VERY interesting indeed. Good job!

I wonder now how other amps (AHB2, LA90, Purifi, Hegel, Schiit, Accuphase, etc) would handle this load, regardless of class.

This is why Erin is about to start testing amps with complex loads too.

Erin's Audio Corner (Group) | For amplifier testing I will be using static 2/4/8 ohm resistances as is the norm | Facebook

For amplifier testing I will be using static 2/4/8 ohm resistances as is the norm. However, in addition, I will also be using (2) different reactive loads: an “easy” load and a “hard” load. The...

www.facebook.com

 

[valerianf]

"Ncore is a very good amplifier with most of "usual" speakers"

In theory is yes, but practically we do need to measure the capacitance that is present at the speaker banana plugs.
We may have some surprise depending how the crossover is designed and what is the drivers size.
I remember measuring 4.7uF with a 5.25" midrange (not connected to the crossover) using a LCR meter.
I need to make new measurements at the speaker banana plugs level.

More worrying is the Pspice model of a 12S30 BMS speaker (12" subwoofer).
https://circuitdigest.com/electronic-circuits/simulate-speaker-with-equivalent-rlc-circuit
The larger capacitance is 240uF.
When playing music, if there is no bi-amp, the class D amplifier will have to drive the large woofer capacitance and in the same time drive the tweeter without any distortion.
After the very useful measurements made by @pma I do not feel well driving a large tower speaker without a class D Bi-amp.
Distortion in the high frequencies will be an issue.

 

[valerianf]

"You can see it is 20mV in audio band, and it is a lot"

Congratulation @pma
You are showing the root cause of the Class D hiss issue.
There is noise in the high frequencies when driving a complex load at a significant power level.

 

[pma]

I have prepared an ABX test with the files recorded from these two amplifiers under 4R//2.2uF load. Here is the new thread.

 

[sernikus]

"Ncore is a very good amplifier with most of "usual" speakers"

In theory is yes, but practically we do need to measure the capacitance that is present at the speaker banana plugs.
We may have some surprise depending how the crossover is designed and what is the drivers size.
I remember measuring 4.7uF with a 5.25" midrange (not connected to the crossover) using a LCR meter.
I need to make new measurements at the speaker banana plugs level.

More worrying is the Pspice model of a 12S30 BMS speaker (12" subwoofer).
https://circuitdigest.com/electronic-circuits/simulate-speaker-with-equivalent-rlc-circuit
The larger capacitance is 240uF.
When playing music, if there is no bi-amp, the class D amplifier will have to drive the large woofer capacitance and in the same time drive the tweeter without any distortion.
After the very useful measurements made by @pma I do not feel well driving a large tower speaker without a class D Bi-amp.
Distortion in the high frequencies will be an issue.

Hmmm...are you sure ?

 

[boXem]

"You can see it is 20mV in audio band, and it is a lot"

Congratulation @pma
You are showing the root cause of the Class D hiss issue.
There is noise in the high frequencies when driving a complex load at a significant power level.

As a starter, there is no "class D hiss issue". What is shown here is that one iteration of one class D topology doesn't handle well a certain type of load.

 

[Matias]

It's funny how it is always easy to generalize "the sound/issue of class D" but noone would say the same for class A or AB.

 

[valerianf]

@boXem
"no "class D hiss issue""

Hiss in active speakers​

https://www.audiosciencereview.com/forum/index.php?threads/hiss-in-active-speakers.7358/

Post #3
"Genelec state the self-generated noise in the specifications of all of their speakers."
If the manufacturer specify the hiss of its active speakers, it is because there is a potential hiss issue with class D.
Have you ever seen an equipment using a class AB amplifier specifying a "self generated noise"????

Now there is some superior designs with a "self generated noise" that is quite not audible.
But is depends of several external factors: the total gain needed in the amplifier (if the class D stage has a low gain then the additional input buffer will increase the noise), and now we know that a capacitive load will increase the distortion for the high frequencies (thank you @pma).

Any electronic product is a compromise of what is possible at a specific price point.
I myself have a class D amp for my surrounds.

 

[HarmonicTHD]

Again. What are typical capacitance values for well engineered passive speakers eg KEF, Revel? Genuinely curious.

I have a capacitance meter. Can I just hook it up to my speakers and measure it?

 

[boXem]

@boXem
"no "class D hiss issue""

Hiss in active speakers​

https://www.audiosciencereview.com/forum/index.php?threads/hiss-in-active-speakers.7358/

Post #3
"Genelec state the self-generated noise in the specifications of all of their speakers."
If the manufacturer specify the hiss of its active speakers, it is because there is a potential hiss issue with class D.
Have you ever seen an equipment using a class AB amplifier specifying a "self generated noise"????

What is the N in THD+N? Noise is intrinsic to electronics.

Now there is some superior designs with a "self generated noise" that is quite not audible.
But is depends of several external factors: the total gain needed in the amplifier (if the class D stage has a low gain then the additional input buffer will increase the noise),

Isn't the input buffer running in class A or AB? According to what you wrote above it shouldn't generate noise.

and now we know that a capacitive load will increase the distortion for the high frequencies (thank you @pma).

For this particular design, with a too high capacitive load. Design that has a self noise far lower than many class AB designs when measured in different conditions.

 

[somebodyelse]

@boXem
"no "class D hiss issue""

Hiss in active speakers​

https://www.audiosciencereview.com/forum/index.php?threads/hiss-in-active-speakers.7358/

Post #3
"Genelec state the self-generated noise in the specifications of all of their speakers."
If the manufacturer specify the hiss of its active speakers, it is because there is a potential hiss issue with class D.
Have you ever seen an equipment using a class AB amplifier specifying a "self generated noise"????

Now there is some superior designs with a "self generated noise" that is quite not audible.
But is depends of several external factors: the total gain needed in the amplifier (if the class D stage has a low gain then the additional input buffer will increase the noise), and now we know that a capacitive load will increase the distortion for the high frequencies (thank you @pma).

Any electronic product is a compromise of what is possible at a specific price point.
I myself have a class D amp for my surrounds.

If that's your argument then there's a "class AB hiss issue" too because plenty of the hissy actives in the past have used cheap, noisy class AB chip amps. See for example EEVBlog #1118 - Why Are Studio Monitors Noisy? Some still do, but economics are pushing them towards class D now. Genelec probably used to specify "self generated noise" when they were using class AB amps too.

 

[thewas]

The load used for the test is 4ohm resistor in parallel with 2.2uF polypropylene capacitor, KEMET automotive type with max. allowed dv/dt = 200V/us. Below please see measured impedance of this load:

At 20kHz, we have 2.762ohm impedance magnitude, -40° phase and 1.07ohm EPDR (equivalent peak dissipation resistor).

Interesting results but I wonder how well a parallel capacitor, which effectively is a short circuit in high frequencies, represents a typical loudspeaker load? I only know of some "highend" electrostatic loudspeakers with similar, for example https://www.stereophile.com/content/sound-lab-1-electrostatic-loudspeaker-measurements while well engineered electrostatic loudspeakers don't necessary show such https://www.stereophile.com/content/quad-esl-989-electrostatic-loudspeaker-measurements
Personally I would like to see some similar measurements and recordings with a more typical loudspeaker load which is not essentially a short circuit.

 

[sernikus]

Again. What are typical capacitance values for well engineered passive speakers eg KEF, Revel? Genuinely curious.

I have a capacitance meter. Can I just hook it up to my speakers and measure it?

in the 2uf range, the voltage should not be higher than 3v rms peak and about 800hz ... so the measurement should be safe for the speaker ...

 

[theREALdotnet]

"Genelec state the self-generated noise in the specifications of all of their speakers."
If the manufacturer specify the hiss of its active speakers, it is because there is a potential hiss issue with class D.
Have you ever seen an equipment using a class AB amplifier specifying a "self generated noise"????

They just specify noise. My trusty old 200W Rotel class AB hissed very pronouncedly with shorted or open inputs. I think we can assume that noise to be “self-generated”.

In comparison, my current March Audio class D amp is dead silent into the same speakers under all conditions, without input signal.

In both cases I’m talking about noise heard with the ear next to the tweeter BTW, not noise heard from the listening chair.

 

[theREALdotnet]

One could learn from Bruno in the Hypex Ncore thread (diyaudio, about 10 years ago) that the limit for additional capacitance that does not affect the operating conditions of Ncore is 200nF.

NC252MP is unable to work with 4ohm//2.2uF load

Test result meets expectation, everybody is happy.

I don’t think testing the Ncore amp with ten times its maximum load capacitance is pointless. Seeing what a piece of equipment does when pushed way out of its operating range is interesting.

Another interesting test would be loading the A250W4R with a tenth of its minimum load impedance and seeing how it performs.

 

[Deleted member 48726]

Again. What are typical capacitance values for well engineered passive speakers eg KEF, Revel? Genuinely curious.

I have a capacitance meter. Can I just hook it up to my speakers and measure it?

Well. As I recall a capacitance meter sends a current at low voltage to the capacitor to charge it. A live circuit would probably not be a good idea. In reviews and tests we can see the distribution of LCR components visualized by the phase curve traced through 0-20.000 Hz in degrees + and - but not the values.

I would mean that in order to establish values one has to look at the crossover components or derive the X and C components from a known value and angle on the phase/impedance graph by vector calculation. At a given frequency we know the Z value and we know the angle and if it's positive or negative.

 

[Talisman]

Test result meets expectation, everybody is happy.

I don’t think testing the Ncore amp with ten times its maximum load capacitance is pointless. Seeing what a piece of equipment does when pushed way out of its operating range is interesting.

Another interesting test would be loading the A250W4R with a tenth of its minimum load impedance and seeing how it performs.

We could also put it in a hydraulic press and see how many kg of pressure it yields, and which element yields first,
but I don't see how this will help us if we use it as an amplifier for 99.9% of the speakers

 

[theREALdotnet]

We could also put it in a hydraulic press and see how many kg of pressure it yields, and which element yields first

I hear that’s how they test amps in Finland. Youtube is making all kinds of expensive experimentation possible…

 

[pma]

We could also put it in a hydraulic press and see how many kg of pressure it yields, and which element yields first,
but I don't see how this will help us if we use it as an amplifier for 99.9% of the speakers

@theREALdotnet

I don’t think testing the Ncore amp with ten times its maximum load capacitance is pointless. Seeing what a piece of equipment does when pushed way out of its operating range is interesting.

Another interesting test would be loading the A250W4R with a tenth of its minimum load impedance and seeing how it performs.

Would you please give a link and recommendation by Hypex regarding maximum load capacitance? IMO it does not exist. Are there manufacturers to specify rated maximum load capacitance? I do not know. Are there speakers with capacitive reactance of 2.2uF and much more? Not many, but they exist.
The mention of 200nF cable capacitance was made by Bruno unoficially at DIYaudio 12 years ago. I have not read any manufacturers warning against 2.2uF load capacitance. The testing like mine is good to warn potential users with such speakers. ASR official reviews will not inform you about such issue. You only get info how the amplifier under test operates with purely resistive load. Not many speakers have purely resistive impedance. And please avoid car analogies. They are completely pointless. I could give similar analogies, but I stick with facts and measurements, though many here stick with meaningless debates only.
You also mix the minimum load impedance with capacitive reactance. Capacitive reactance is frequency dependent and as such must be considered. At 20kHz, my load has 2.76ohm modulus, still more than 2ohm rated by Hypex. Strictly put, the 4ohm//2.2uF impedance is within rated limits of Hypex NC252MP in the audio band.