O-NOORUS D4 budget stereo amp review

[mcdn]

This is a review and detailed measurements of the O-NOORUS D4 stereo integrated amplifier. It was sent to me by the company and costs US$139 from them direct with the 48V/5A power brick. It is fundamentally a small box TPA3255 design with PFFB to give a flat response into any load, but it has some extra to offer.

The selling point for the D4 is versatility. On the input side we have RCA analog plus optical, coax, bluetooth, USB and HDMI digital inputs. There's also a single RCA out, a volume control and three tone controls - bass, mid and treble. The tone controls can be bypassed using the toggle switch on the front panel.

The amp is a small and solid unit with good thermal connections between the electronics and the case. The volume and tone knobs have a bit of play, making it hard to get repeatable results, but they are good for the price. The back panel connectors are all firm and gave me no trouble as I went through my testing. The supplied 48V/5A power brick is of good quality. I did use my more powerful 60V/20A bench supply to check for any limitations but it gave the same results.

Measurements - baseline

Amir has a fancy Audio Precision but I only have community gear. Still the results are pretty good: REW is running on a Macbook Air. A Caldigit TB3+ dock links it to an SMSL SU-1 for RCA out, a Topping D10 balanced for coax and optical toslink out (and balanced out of course, but not needed for this review), and a Cosmos ADCiso for measurements. The Cosmos is powered from a battery bank and everything else is powered from the CalDigit hub. The dummy load is my own precision one good for 400Wpc continuous power per channel. This all comes together to give this baseline:

Unless otherwise noted all measurements are taken at 5Wpc into a 4ohm load, RCA input, volume at 12o'clock, toggle switch "bypass", both channels driven and both channels loaded. This gives ~22dB of gain. Both input and output devices are set to 96KHz 24bit.​

Measurements: 1KHz FFT

Starting with the 5W 1KHz FFT we get SINAD of 78.5dB, and Noise of -108dB which would just about make it into the green zone on Amir's list of amps. The noise performance is actually superb, we're mostly limited by distortion here. Manufacturer spec is 87dB SINAD, which isn't right per my measurements but I suspect they used 1/SNR instead of 1/THD+N, which is an easy mistake to make. [edit: the manufacturer has confirmed my measurements are correct]

The specs for THD (-80dB or 0.01%) and noise (-108dB) are both met and are both good numbers, especially for a "do it all" amp such as this.

SINAD remains the same when we go to 80Wpc, although noise goes up to -102dB, which is still excellent:

Measurements - frequency response

Overall frequency response is flat, channel matching is within 0.1dB, and there is a 1dB rolloff at 20KHz.

Measurements - Multitone

Both the 1/3 octave 20-20KHz multitone and CCIF 19&20KHz are good, with ~15bits of resolution

Measurements - Power

The THD+N vs power plot shows about 120W/ch maximum output into 4ohms at 0.1% distortion, slightly above expected given the power supply. The curves are pretty much the same at 50Hz and 20KHz, which is impressive. Into 2ohms it's again power supply limited and achieves 60W/ch. I didn't test into 8ohms.

Measurements - THD vs frequency

THD is also well controlled across frequency:

Measurements - tone controls

The tone controls do what they do. I've put them all on one chart, use the legend to see what does what.

Measurements - sub out

The "sub out" is low passed with what looks like a first order 200Hz filter. I'm not sure why they bother, might as well leave it full range. The tone controls apply to it as well.

Measurements - digital inputs

All the digital inputs including HDMI and Bluetooth share the same performance. They are low passed at around 22KHz regardless of the input bandwidth, suggesting internal processing is done at 44.1 or 48KHz. Overall resolution is a bit poor at SINAD of 65dB, equivalent to 13.6 bits. There's also a suggestion of a low pass filter, seen in the slight wobble in the frequency response before 50Hz.

Continuous power

Nobody can agree exactly how to measure continuous power output, but everybody seems to want to know a number. I tested the little amp with a 20Hz-20kHz pink noise signal in an air conditioned room at an ambient temp of about 25C. It was sitting on a solid surface with plenty of room around all sides. The amp case was cooled down with a bag of ice between runs and left to return to ambient before the next run started. There was light airflow around the amp due to the aircon and the fans of the dummy load nearby.

With a 5W signal the amp case warmed up to 10C above ambient (35C) in about 10 minutes. With a 10W signal it was about the same. With a 20W signal the amp got to +15C (40C absolute) after 10 minutes. With a 40W signal it was +20C (45C absolute) after 10 minutes. It took a long time to cool it down after that run, suggesting a lot of the heat was going to internal components, not just the case. With an 80W signal the amp entered protection after 1 minute.

On/off pops

There is a 250mV power on pop and a 90mV one on power off. Both are audible but not dangerous to speakers.

Listening

The only passive speakers I have are my Sonus Faber Cremona Auditors. They sounded just as good through the D4 as through the Lyngdorf TDAi2200 I usually drive them with, minus the room correction and some peak volume. Franco Serblin would have been happy to see so many good amplifier options for his speakers.

Overall conclusions

For the size of the package and the flexibility I really like the D4. The hardware is well made and while the knobs are a bit imprecise they do the job. The digital side could be a bit higher resolution but noise is low throughout so as an "all in one" solution it works well. There is a bit of microphony (the amp making noise itself) with certain test signals, but nothing that would be audible in use. The range of inputs is great, and as a bonus the thermal design is good for the class, offering decent power under difficult test conditions.

 

[mcdn]

REW .MDAT files available at https://www.icloud.com/iclouddrive/052ORCIFlcPfJF-CYjv8l_70w#d4_mdat_files

They include some measurements not shown in the review such as THD vs power at 50Hz and 20KHz, 2ohm power testing, etc

 

[biker29]

Thanks for the review. Looks like a good low cost amp with lots of flexibility.

 

[DanTheMan]

I so wish these companies would start throwing a variable high pass on their own amplifiers. I guess one has. I mean this is still really nice, but that would just perfect the design. The input indicators are even useful. Very nice design for the money. Thanks for the review!

 

[Hayabusa]

Is measuring power with a noise signal a good idea?
An 80 watt or even a 40 watt average signal will have much large peaks and will cause clipping?

 

[mcdn]

Is measuring power with a noise signal a good idea?
An 80 watt or even a 40 watt average signal will have much large peaks and will cause clipping?

There are lengthy threads on the topic. Pink noise is more representative of music than a single tone is, but less so than m-noise. I present it here for interest, not because it’s a standard.

 

[obsessive9119]

Are the speaker outs high-passed when the sub is connected? That’s the one thing I’m really missing from my ZA3.

 

[er|κzvio1in]

The low pass filter on the subwoofer with a slope like that will eliminate audible noises from ground loops or interference without any downsides. I have digital crossovers for my subwoofers, but still I cannot disable the subs' built in analogue crossovers because there will be audible noise if I don't so I just set them to their max frequency.

 

[daniboun]

Yet another new TPA3255 with respectable performance but no revolution here... Digital inputs are not good. The Aiyima A80 seems to be its most serious competitor. Too bad there is no remote control. I would rather wait for the D3 PRO which has already been announced and which should be more interesting

 

[daniboun]

I have a question for Americans.... With Trump's new policy on import taxes from Chinese and European countries, won't it slow down purchases via Aliexpress and others? We're talking about a 25% increase

 

[Hayabusa]

There are lengthy threads on the topic. Pink noise is more representative of music than a single tone is, but less so than m-noise. I present it here for interest, not because it’s a standard.

I agree its more like music, but you should be aware of peak to average ratio to at least prevent clipping.
A quick test in audacity says -13dB. or factor 4.46. in amplitude and 20 in power.
This means your 80 watt signal has peaks of 1600 watt...
5 watt pink noise would be close to the clipping value of this amp..

 

[mcdn]

Are the speaker outs high-passed when the sub is connected? That’s the one thing I’m really missing from my ZA3.

No, the sub out and speaker outs are both always on

 

[mcdn]

I agree its more like music, but you should be aware of peak to average ratio to at least prevent clipping.
A quick test in audacity says -13dB. or factor 4.46. in amplitude and 20 in power.
This means your 80 watt signal has peaks of 1600 watt...
5 watt pink noise would be close to the clipping value of this amp..

Nothing of the sort is going on. REW is generating a pink noise signal with total power of a certain amount. Within a specific passband/fft bin there will be periodic high and low voltages but there’s no clipping and the average power remains constant

 

[HighFutility]

I have a question for Americans.... With Trump's new policy on import taxes from Chinese and European countries, won't it slow down purchases via Aliexpress and others? We're talking about a 25% increase

Pretty sure HIFI is low on the list of people's requirements right now.

Edit: The entire department of education just got wiped out.

 

[daniboun]

Edit: The entire department of education just got wiped out.

Wow! What will the consequences be? I'm curious.

 

[HighFutility]

Wow! What will the consequences be? I'm curious.

Tekton is the best speaker manufacturer of all time—dystopian times from now on.

 

[daniboun]

Tekton is the best speaker manufacturer of all time—dystopian times from now on.

Except this isn't a fictional story... I'd say it's a sad reality for many businesses. In any case, jobs are definitely going to be threatened.

 

[Hayabusa]

Nothing of the sort is going on. REW is generating a pink noise signal with total power of a certain amount. Within a specific passband/fft bin there will be periodic high and low voltages but there’s no clipping and the average power remains constant

I am not talking about clipping in signal input domain.
Its the output. If you play a pink noise signal with an average power of 80watts the peaks will be much higher.
So is that what you do? play 80watts of average power?

 

[mcdn]

I am not talking about clipping in signal input domain.
Its the output. If you play a pink noise signal with an average power of 80watts the peaks will be much higher.
So is that what you do? play 80watts of average power?

I can double check the specifics a bit later but I think the specific PN used for the test had a crest factor of 6dB

 

[mcdn]

I am not talking about clipping in signal input domain.
Its the output. If you play a pink noise signal with an average power of 80watts the peaks will be much higher.
So is that what you do? play 80watts of average power?

I checked and these are the REW settings I used:

REW handily shows the crest factor and the peak voltage, which I hadn't realised before. 80W into 4ohms is sqrt(80*4)=17.9V, so the peaks would indeed have been out of operating comfort at 35.8V or 320W. Not the 1600W in your original calculation but still well spotted, thank you @Hayabusa !

The thing is, assuming we want to avoid clipping at all, maybe a 12dB crest factor signal is more appropriate, in which case the CTA-2034 option in REW would be a better choice. But then I would need to figure out the 20ms instantaneous 1% power of the amp. It's probably pretty similar to the continuous power in this case, but REW doesn't offer an easy way to measure it.

I'll start by updating my test protocol for thermal testing to keep using the 6dB crest factor pink noise, but limiting it to -6dB ref max 1% power at 1KHz. This will allow higher RMS power levels to be reached before clipping than if using bigger crest factors. For the D4 the max thermal test level would be 25W.