SMSL A100 class D amplifier test

[pma]

This is directly at the speaker box terminals (X - Y method). The cable inductance does not help much to reduce EMI spikes.

And the differential measurement setup. Ref point is mains PE.

And this is an ASR recommended amplifier

SMSL A100 Review (Stereo Amplifier)

This is a review and detailed measurements of the SMSL A100 A100 stereo amplifier. It is on kind loan from a member. I can't find it on sale anywhere but seems to have cost US 100. I like the simply but rather elegant/symmetrical design. Same for thoughtful rear connections: Almost every...

www.audiosciencereview.com

Something is rotten in the state of Denmark.

 

[Geert]

Related topic of someone with an SMSL DA9 (using the same MA12070 chip) running into interference issues with his powerline ethernet adapter https://audiosciencereview.com/foru...d-problem-with-the-smsl-da-9-amplifier.24061/

 

[pma]

From the Infeon datasheet:

Of course the speaker wires are longer than 80cm. And I can “tune” the music from the amp on the AM transistor radio at 2nd harmonic of carrier frequency near 1300kHz in my whole sitting room. Who cares? The manufacturer definitely not.

Note: in my case, the ADC with input divider (E1DA) handles the PWM signal very well.

 

[Ropeburn]

From the Infeon datasheet:

View attachment 412123

Of course the speaker wires are longer than 80cm. And I can “tune” the music from the amp on the AM transistor radio at 2nd harmonic of carrier frequency near 1300kHz in my whole sitting room. Who cares? The manufacturer definitely not.

Note: in my case, the ADC with input divider (E1DA) handles the PWM signal very well.

That's more than just a questionable design. Expecting the customer to use a ferrite filter even under a length of 80cm? What the hell. I wonder how it passed regulations at all.

 

[jeffaegrim]

Thank you for your work @pma, your contributions to the forum have always been very valuable. I have visited your website with interest in the past - if you’re open to suggestions, a short „conclusions“ section from your perspective like the one you included in the OP post here would be most welcome in your other reviews going forward.

 

[pma]

Measurements with amplifier USB-A input

In the USB mode, the amplifier works only with the 48kHz/16bit (DVD) format. This might be acceptable, if ...

1. THD 1kHz spectrum
We can see high jitter with the USB input (skirt)

2. THD vs. frequency

Frequency response + distortion

Distortion

3. THD vs. level at 1kHz

The behaviour is quite erratic. That is no good. But it works. For a small bedroom system, it would work, if you are willing to digest enormous level of emitted HF interferences.

 

[miero]

This might be an inappropriate question, but will it blend?
- https://youtube.com/@blendtec

 

[Sokel]

Do we know the EMI limits that regulations accept?
Had a look about a European design (icepower) for safety,etc and it seems there's a bunch of them:



(chapter 14,under Safety and EMC standards)

 

[somebodyelse]

Related topic of someone with an SMSL DA9 (using the same MA12070 chip) running into interference issues with his powerline ethernet adapter https://audiosciencereview.com/foru...d-problem-with-the-smsl-da-9-amplifier.24061/

There's a level or irony here. Powerline ethernet adapters are notorious among radio amateurs for their RF emissions, and questions have been asked about how they make it through compliance testing. The amps probably use a similar loophole - test with short cables to the dummy load and it will probably get a pass.

 

[dr_mick51]

SMSL A100 amplifier review

This test review was performed to comply with FTC: 16 CFR Part 432 (July 5, 2024) - Power Output Claims for Amplifiers Utilized in Home Entertainment Products
The SMSL A100 is a small, affordable, 2-channel Class D amplifier. It embeds an Infeon MA-12070 amplifier chip and it has a built-in power supply (AC 100V-240V input), which is a nice feature. The inputs are in stereo RCA, USB-A and Bluetooth. I bought it from Amazon.de at price of 80.37 EUR with free shipping, so it is a bargain.

View attachment 411904

View attachment 411905

The amplifier works in the BTL mode and on any of the speaker binding posts (+ or -) there is +12.25Vdc against the signal common, which is at the body of RCA connectors. So be careful when measuring the amp, use well rated coupling capacitors and differential input mode of the ADC. Another note - the amplifier is a class I instrument and its signal common is connected to the PE (middle pin of the mains plug). This may worsen S/N when connected to the class I signal source like class I DAC.

1. Manufacturer's specifications
Manufacturer's specifications are listed below

View attachment 411906

Frankly speaking, the specified output power seems quite high related to the chip used, so I rather attach specifications from the MA-12070 datasheet. The SMSL A100 implementation cannot have better power rating than the chip used, itself.

MA-12070 datasheet specifications

View attachment 411907
This is something I might believe so I will consider the continuous rated power to be 2 x 30W/8ohm/1kHz.

2. Test conditions
The amplifier was tested with the built-in power supply, 230Vac, and 8ohm load in both L and R channel. Testing with 4ohm load will be done later just as an orientation measurement, as the rated power to 4ohm is unknown.

3. Pre-conditioning
First of all, the amplifier was pre-conditioned at about 1/8 of the rated power, 2 x 4.2W/8ohm, for one hour, as requested by the FCT document. At the end of the pre-conditioning period, the amplifier case was only slightly warmer than in the beginning. The distortion in the beginning and in the end of the pre-conditioning was unchanged.

4. Frequency response
Frequency response was measured at 5W output power into 8ohm. The amplifier has filterless design, without output LC filter, so the frequency response is independent of load impedance.

View attachment 411910

5. Power frequency response measured at 27W/8ohm, to 40kHz

View attachment 411909

Power frequency response was estimated as 20Hz - 20kHz. Behaviour above 20kHz is a bit erratic.

6. Distortion was measured at volume pot set to maximum, G = 40dB

6.1. Distortion at 5W/8ohm

View attachment 411911

6.2. Distortion at 21W/8ohm

View attachment 411914

6.3. THD vs. stepped output voltage into 8ohm load at 1kHz with measurement bandwidth 45kHz and measuring bursts of 0.8s length and 0.5s intervals

View attachment 411913

Maximum power is 37W/1kHz/8ohm with THD = 10.5%

View attachment 411915

The amplifier meets the specs at 30W/1kHz/8ohm with THD<1%

6.4. THD vs. frequency at 5W/8ohm

View attachment 411916

There is a steep rise of THD above 1.5kHz and THD > 1% above 13kHz with measuring BW45kHz.

6.5. 5 minutes test at rated (derated) power 2 x 27W/8ohm/1kHz

The amplifier has passed 5 minute test at 27W/8ohm/1kHz, being only mildly warm at the end of the test.

View attachment 411917

Further increasing of power resulted in distortion > 1% in the 5 minute test. But not to the shut down.

7. Conclusion
The amplifier is a mixture of good and bad points. Built quality is quite good, concerning the bargain price. It did not get hot during tests and never shut down itself. Frequency response does not have the infamous dependence on load impedance and response peaking above 10kHz, as it is a filterless design. However, this brings troubles in the field of electromagnetic interference. My wifi staying near was almost paralysed. The DAC feeding the amp with test signals was unable to work without using USB-ISO isolator. For the 5 minute rated power test I had to use analog generator, as the DAC was freezing at amplifier high power due to emitted interference.
Manufacturer's specification of power is vague and unusable. So I rather used the specs from MA-12070 chip datasheet, that declares rated continuous power 30W/8ohm/1kHz per channel. This was almost fulfilled in the 5 minute test, the amplifier was able to do it at 27W with THD << 1%.

8. Listening
I have connected the SMSL A100 with my Topping DX5 (class I) which gets the signal from S/PDIF coax, and to my surprise, there is no issue with hum, buzz or whistles. The sound is “normal”, nothing to complain. We shall see.

Pavel Macura 12/6/2024

Hello. It is my understanding you tested this with 4 and 8 ohm resistors only. Right?. This amplifier requires a 20uH inductance at the output which is normally produced by the speaker coil itself. Try adding those inductors and re-measure to make sure all that HF interference is reduced.

 

[pma]

Hello. It is my understanding you tested this with 4 and 8 ohm resistors only. Right?. This amplifier requires a 20uH inductance at the output which is normally produced by the speaker coil itself. Try adding those inductors and re-measure to make sure all that HF interference is reduced.

You never know how the crossover circuit is designed, it may start with something like R+C, few ohms + tens of uF to compensate for woofer impedance peak. You never know and the advice is useless. And, the emission is more or less same with the speaker + cable, as I have mentioned and showed the photos and measurements here. Just read this thread, above. Then you would not ask if I tested only with resistors.

I have reported the EMI issue to EU seller, Amazon.de. I do not think that attempt to explain it here makes any sense. I also do not think that Amazon will do anything. But I cannot stay silent about this EMC compliance violation.

 

[spyne]

We don't need to know if a huge, overbuilt amplifier that costs several thousand dollars can pass the testing. It probably will

You, ahhh, haven't seen too many of Amir's reviews then, have you?
Audio Research 100.2 and D300, Lyngdorf TDAI-3400, Roksan Attessa, Carver Raven 350

 

[dr_mick51]

You never know how the crossover circuit is designed, it may start with something like R+C, few ohms + tens of uF to compensate for woofer impedance peak. You never know and the advice is useless. And, the emission is more or less same with the speaker + cable, as I have mentioned and showed the photos and measurements here. Just read this thread, above. Then you would not ask if I tested only with resistors.

I have reported the EMI issue to EU seller, Amazon.de. I do not think that attempt to explain it here makes any sense. I also do not think that Amazon will do anything. But I cannot stay silent about this EMC compliance violation.

Ok, sure. Whatever you say. You already made your mind up. I wonder if have not fully read the Infineon documentation:

And saw all the tests (sections 10 to 15) are done with a filter and a 22uH inductor as part of the load:

Also, there are tests on this website where 22uH inductors are added as recommended by Infineon for testing:
https://quantasylum.com/blogs/news/first-look-infineon-merus-ma12070

So, don't say my recommendations are useless. It's all said on the documentation.

You have the right to complain if this particular amp does not work for you. I agree this model(SMSL A100) missed the output filter which should not be optional. But Infineon has tested this chip amp under many conditions, and they on their Application Note clearly show how to implement those filters.

I have this amp (not using it anymore. It has been replaced by better amps), but some people reported noise issues with some speakers(as you mention, depending on the crossover topology), others said to have issues with By-amped speakers (while connected to tweeters only, and not having enough inductance) and adding those inductors had worked for them.
I was able to internally solder those inductors(just for peace of mind, never had any noise issues).

 

[kyuu]

You, ahhh, haven't seen too many of Amir's reviews then, have you?
Audio Research 100.2 and D300, Lyngdorf TDAI-3400, Roksan Attessa, Carver Raven 350

Oh I have. I guess I should have qualified that with something to the effect of quality big, expensive amps. Not poorly engineered big, expensive amps. Which of course Amir's testing already reveals quite well without emulating the FTC...

 

[Vladimir Filevski]

It is my understanding you tested this with 4 and 8 ohm resistors only. Right?.

No. Not all tests, at least. Look at the second photo in the post #21:

 

[Chrispy]

Interference - it is a kind of issue. When my small AM radio is put near the speaker cable, it can demodulate and play the same music that goes from speakers. Tuned somewhere between 600 - 700 kHz . “FCC compliant”.

You do that often?

 

[Vladimir Filevski]

I agree this model(SMSL A100) missed the output filter which should not be optional.

User can not possibly know if that particular amp model (A100) is filter-less or not. So it is tested as-is.

But Infineon has tested this chip amp under many conditions, and they on their Application Note clearly show how to implement those filters.

Again - it is a mistake of the A100 manufacturer (SMSL), not a mistake of the reviewer (pma). Why SMSL did not include that 22 uH inductor inside their amplifier?

 

[Vladimir Filevski]

You do that often?

I do.
Not listening to an AM radio is not particularly good alibi for a bad amplifier engineering.

 

[Vladimir Filevski]

... and adding those inductors had worked for them.
I was able to internally solder those inductors(just for peace of mind,..

... and your frequency response ended up with substantial loss in the frequency response at 20 kHz.
I tried including 22 uH inductor in the loudspeaker simulation (VituixCAD) of 3-way loudspeaker I build with ScanSpeak D2905 tweeter - unacceptable loss of 3 dB at 20 kHz!
Sorry, including 22 uH inductor post festum is not good for a frequency response and for a piece of mind. Infineon is agreeing on that too - look at their measured frequency response on 4 ohm resistor without and with 22 uH inductor - huge 7 dB loss at 20 kHz and 2 dB loss at 10 kHz:

 

[pma]

SMSL A100 is sold as a completed product, thus the customer is not supposed to insert inductors or anything into speaker wires. It must work “as is”. The complex load impedance of the speaker is also not pre-defined, parameters are specified to 4 and 8 ohm. I do not understand why some try to protect manufacturers, customers are to be protected.