Starkrimson® Ultra Amplifier Modules -- 250/500W into 8/4ohm

[orchardaudio]

Here it is with the crest factor adjustment.

 

[orchardaudio]

Thanks .... so why are the 2 spectral plots (one referenced above and one in the test report shown below) so different? I see much more distortion products at multiples of 1khz (before 20kHz) plus a much different (noisier) profile for the switching frequency by-products for the former spectral plot. I'm assuming these plots pertain to the exact same amplifier design.

Lastly, what are the 1-Watt and full-power bandwidths of this amp in 8 and 4 ohms. Is it also rated for 2 ohms operation?

One spectral plot is at full power, yes there is more distortion at full power than at -40dB output which is only 0.25W.

I cannot tell you the exact bandwidth because the audio analyzer tops out at 80kHz, and at that point, the amp is still not -3dB down. Hence I list the bandwidth as 80kHz+. The bandwidth does not change with power.

The amplifier is 2-ohm stable and is capable of 20A peak current.

Here is the 8ohm frequency response:

Here is the 4ohm frequency response:

 

[David_M]

Here it is with the crest factor adjustment.
View attachment 138328

Nice ... Noise < -120dB ==> ENOB > 20bits which is in line with the best DACs and headphone amplifiers/Preamps Amir has ever measured, such as the Element X and Matrix PRO by Matrix Audio that uses the ultrahigh-resolution ESS9830PRO DAC. Congratulations!

 

[David_M]

One spectral plot is at full power, yes there is more distortion at full power than at -40dB output which is only 0.25W.

I cannot tell you the exact bandwidth because the audio analyzer tops out at 80kHz, and at that point, the amp is still not -3dB down. Hence I list the bandwidth as 80kHz+. The bandwidth does not change with power.

The amplifier is 2-ohm stable and is capable of 20A peak current.

Here is the 8ohm frequency response:
View attachment 138329

Here is the 4ohm frequency response:
View attachment 138330

With such wide bandwidth, it'd be nice to see what a full-power 10kHz squarewave looks like, time permitting ofcourse . The shape of the output waveform not only highlights the amplifier bandwidth (and slew rate) but also its stability.

 

[Veri]

Very good measurements, indeed.

 

[orchardaudio]

Here it is with the crest factor adjustment.
View attachment 138328

Also the updated test report.

 

[orchardaudio]

Here it is with the crest factor adjustment.
View attachment 138461

I updated the quoted post as the previous one did not show 2 of the 32 tones as the x-axis was clipped on both sides.

 

[orchardaudio]

With such wide bandwidth, it'd be nice to see what a full-power 10kHz squarewave looks like, time permitting ofcourse . The shape of the output waveform not only highlights the amplifier bandwidth (and slew rate) but also its stability.

10kHz full power square wave into 8ohm both channels. 1MHz bandwidth, this is the highest bandwidth the audio analyzer can measure.

 

[David_M]

This is an excellent high-frequency response at full power. I estimate a 3us rise within a 120V voltage swing ==> Slew rate = 40V/us. Really high and superb! Most amplifier 10kHz output waveforms are seen at 1V output, which I consider meaningless.

Hopefully, the amp should be able to drive electrostatic speakers (by Martin Logan) with 0.5 ohms at 20kHz with no problem as their panels present as a capacitive load to an amplifier with decreasing impedance with frequency?

Using the formula of fmax = Slewrate / (2*pi*Vp), where fmax is the highest frequency the amp is able to produce before slew-rate induced distortion (SID) kicks in, I estimate fmax = 40e6/(2*pi*60) ==> 106kHz full power bandwidth in 8 ohms... a figure you mostly see at 1W for many amps.

Congratulations on these excellent specs!

 

[David_M]

Uh-Oh ... here I am again with another (tiny) request that doesn't require any measurement

Is it possible to overlay the input 32-tone signal from the AP analyzer output with the output of the amp shown above? I was curious to see the noise +distortion floor of the AP signal when compared to that of your amplifier. Thanks!

 

[orchardaudio]

Uh-Oh ... here I am again with another (tiny) request that doesn't require any measurement

Is it possible to overlay the input 32-tone signal from the AP analyzer output with the output of the amp shown above? I was curious to see the noise +distortion floor of the AP signal when compared to that of your amplifier. Thanks!

I cannot overlay the two because they have different reference levels. The difference is the gain of the amp which is 19.05dB.

However here is what the multi-tone looks like when you loop the audio analyzer back on itself.

 

[David_M]

Thank you! The low-level intermodulation distortion products are almost identical to those from the AP unit, with the exception of some minuscule 60Hz hum and its harmonics, plus those at higher tone frequencies, all still at low levels under -110dB. Excellent performance!!!

 

[David_M]

Since your power amplifiers have a differential output based on your proprietary DC-coupled, fully balanced dual feedback modulator AND since (good quality) differential circuits eliminate all even-order harmonics in a signal (once converted to single-ended afterward:

1. Will the amps work for loudspeakers that ground their signal return terminal, essentially forcing the amp to work in a single-ended (SE) operation?

2. Will the amps eliminate even-order distortion in loudspeakers, essentially reducing loudspeaker distortion? Essentially they'd be driving loudspeakers transducers differentially, reducing their distortion profile. I think the emitted differential sound wavefront sum as it enters reaching the left/right ear converting to a SE wavefront into the ear canal with no even-ordered distortion?

Thanks!

 

[orchardaudio]

1. Will the amps work for loudspeakers that ground their signal return terminal, essentially forcing the amp to work in a single-ended (SE) operation?

No. This is shorting out an amplifier output.
In the instructions, I specifically state not to connect either of the amplifier's outputs to ground/earth ground.

 

[kipman725]

Congratulations on the great performance! particularly impresed with the THD+N Ratio W.R.T frequency result. Perhaps a technical article on the modulator?

 

[David_M]

No. This is shorting out an amplifier output.
In the instructions, I specifically state not to connect either of the amplifier's outputs to ground/earth ground.

The point is most users have no clue whether their speakers have grounded signal returns or not as they have no problems using amplifiers that are single-ended, which is the majority of all analog amplifiers out there. Is there a protective circuit to prevent this possible damage?

 

[orchardaudio]

The point is most users have no clue whether their speakers have grounded signal returns or not as they have no problems using amplifiers that are single-ended, which is the majority of all analog amplifiers out there. Is there a protective circuit to prevent this possible damage?

Absolutely.

 

[David_M]

... and on the topic of distortion elimination, I mentioned in my original post?

 

[orchardaudio]

... and on the topic of distortion elimination, I mentioned in my original post?

Yes, the fully balanced nature of the amp helps to eliminate, or better to say lower the effect of even-order harmonics. No circuit can be 100% balanced due to component tolerances so even order harmonics are not going to be completely nulled. I use 0.05% and 0.1% resistors.

 

[jgillingham]

Will the amps eliminate even-order distortion in loudspeakers, essentially reducing loudspeaker distortion? Essentially they'd be driving loudspeakers transducers differentially, reducing their distortion profile.

Thanks!

The balanced nature of the amp can dramatically reduce its own even-order distortion, but this does not affect the even-order of the loudspeaker itself. Typical dynamic loudspeakers may generate distortion as a result of non-linearities in their suspensions and magnetic circuits. An amp would have to generate an inverse distortion in order to cancel the loudspeaker's distortion.

I think the emitted differential sound wavefront sum as it enters reaching the left/right ear converting to a SE wavefront into the ear canal with no even-ordered distortion?

Thanks!

No, this does not happen.