SMSL SA100 Audio Amplifier Review

[RayDunzl]

You say when you are listening at around 5W at home but not stating what an input level is fed to amplifier, I'm sure it's nominal and later attenuated by amp' volume control. If it was set to lower than sensitivity rating then you'd need to crank the volume up to reach the same spl, which would be full of distortion and noise.

Re: levels discussion above...

Maybe @amirm could loop back the AP at 0.18V to show the noise contribution of this not-2V source

 

[scooter]

All amp input sensitivities may be different

Exactly, I just thought it would be better to test them according to their designed specifications, to create an "environment" they were designed to operate within.

DAC -> 2vpk -> volume control -> 0.18V -> amplifier -> 5 watts into 4 ohms and it is likely very loud of speakers are attached

But usually you don't have volume controls on majority of DACs. So it would look like this:
DAC -> 2vpk -> amplifier -> volume control (this time not at near max but highly attenuated) ->5 watts (cleaner).

The sensitivity rating at the amplifer is for maximum output, not necessarily the same as sane listening level.

That's why I'm trying to avoid stressing an amplifier.
AFAIK sensitivity is the lowest level that will produce nominal output level (amp volume attenuation 0, max volume) and should be always better to feed more than that signal, preferably near nominal input.
This amp needs at least 0.6V to produce designed nominal output according to its sensitivity rating. I just don't understand how we'll see its real performance if signal applied is even less of that - only 0.18V.
Wouldn't we get more clear picture of an amplifier performance if it's chosen nominal input signal first and only later output level matched?

Maybe we're talking about different things

Maybe

 

[RayDunzl]

Exactly, I just thought it would be better to test them according to their designed specifications, to create an "environment" they were designed to operate within.

That's where the 5V output comes in.

If you swap amps in your system, you will likely play them all at the same output level (same volume in the room or headphones), even if some could go much much louder than another.

But usually you don't have volume controls on majority of DACs. So it would look like this:
DAC -> 2vpk -> amplifier -> volume control (this time not at near max but highly attenuated) ->5 watts (cleaner).

Nope.

DAC -> 2vpk -> volume control -> amplifier stage

Volume control occurs before the amplification stage, even if the knob is on the same box.. (there may be some rare exception)

AFAIK sensitivity is the lowest level that will produce nominal output level (amp volume attenuation 0, max volume) and should be always better to feed more than that signal, preferably near nominal input.

Nope.

Sensitivity is the highest input voltage that will produce the maximum unclipped (or whatever percentage of clipping is in vogue now) output.

The sensitivity value for an amplifier = the signal voltage at the amplifier stage to achieve wide open full output.

---

If your amp has a "volume" control, it attenuates the input voltage, so you don't overdrive the amp stage, and so you can adjust the volume.

Don't overthink it.

This amp needs at least 0.6V to produce designed nominal output according to its sensitivity rating. I just don't understand how we'll see its real performance if signal applied is even less of that - only 0.18V.

He's not looking for maximum performance.

My amps don't run 700W in here, it's insanely loud. I run them at a few watts, maybe 5 watts (average) which would still be too loud. Probably a watt or less average power.

Look at speaker sensitivity - 1 watt = 85 to 90 or more dB SPL. That's loud if continuous. The test is continuous.

He's comparing them at equivalent output levels. <- that's the key to the test.

 

[RickSanchez]

Once again it shows that nothing can be guaranteed until it is measured.

Well, maybe not "nothing" ...

" 'Tis impossible to be sure of anything but Death and Taxes."
- The Cobbler of Preston by Christopher Bullock (1716)

 

[scooter]

@RayDunzl thanks for your time and explanations. It was just hard for me to correctly imagine and understand an amplification stages. So just to confirm, in terms of performance there isn't any difference if we for example set the volume at max position (no attenuation) and apply the signal level equal to sensitivity rating, or apply a higher level signal and lower the volume with a proper ratio to achieve the same maximum unclipped power output. Meaning that an amplification circuit still "sees" a voltage equal to the sensitivity rating. Is this correct?

Sensitivity is the highest input voltage that will produce the maximum unclipped (or whatever percentage of clipping is in vogue now) output

I think this would be true if related to an amplifier with no volume control at all, in this case we can call sensitivity the highest input level, otherwise with volume control implemented a sensitivity will be the lowest level for an amp to produce unclipped full power output. So there is some headroom until an input analog channel is not clipped (maximum input level). Or am I wrong again?

 

[Ron Texas]

Perfect for boom boxes.

 

[Tks]

Well, maybe not "nothing" ...

" 'Tis impossible to be sure of anything but Death and Taxes."
- The Cobbler of Preston by Christopher Bullock (1716)

We're closing in on proving the former wrong in the future perhaps. Though the latter seems it will persist to eternity.

 

[RayDunzl]

thanks for your time and explanations.

I try. Difficult to be clear with quick texts. And I get to be wrong, now and then, too.

It was just hard for me to correctly imagine and understand an amplification stages.

I think I know a little about the subject. That makes me dangerous.

So just to confirm, in terms of performance there isn't any difference if we for example set the volume at max position (no attenuation) and apply the signal level equal to sensitivity rating, or apply a higher level signal and lower the volume with a proper ratio to achieve the same maximum unclipped power output. Meaning that an amplification circuit still "sees" a voltage equal to the sensitivity rating. Is this correct?

"No difference" is hard to achieve when looking at a circuit with the AP, easier if you're just listening.

I think you have a reasonable understanding now.

I think this would be true if related to an amplifier with no volume control at all, in this case we can call sensitivity the highest input level

Yes.

That's the "traditional" meaning.

otherwise with volume control implemented a sensitivity will be the lowest level for an amp to produce unclipped full power output.

Now you have a volume knob in there which can permit no signal to the amplifier (and no output), or maybe even amplify the input signal. The voltage level at the input in terms of "sensitivity" before the volume control becomes a bit undefined. You can apply any (reasonable) voltage, turn the knob, and end up with full output or any fractional unit thereof that you desire, or even none.

The power amplifier stage will still need Xvolts input into it from the attenuator/preamplifier stage to reach its full output, though. That's hidden inside the black box.

I presume: Sensitivity rating for an amplifier with a volume control is the voltage that, when applied to the attenuator at "full volume" will tickle the amplifier section to its maximum. I'm not sure, different vendors may have different definitions. Who knows.

So there is some headroom until an input analog channel is not clipped (maximum input level). Or am I wrong again?

In an amplifier, headroom is the difference between the maximum output it can generate and the level at which you choose to operate it.

No headroom is bad. Some headroom is good. Lots of headroom may or may not be any better. I have 699W of headroom during casual listening. I don't think about it.

---
Disclaimer: All of the above as I understand it.

 

[scooter]

The voltage level at the input in terms of "sensitivity" before the volume control becomes a bit undefined.

The power amplifier stage will still need Xvolts input into it from the attenuator stage to reach its full output, though. That's hidden inside the black box.

Thank you for these.

 

[MZKM]

So, why get the SA100 when the ~$130 Topping MX3 is a mini all-in-one (including headphone out) with better power amp measurements (similar wattage with better SINAD)?

 

[SEKLEM]

So, why get the SA100 when the ~$130 Topping MX3 is a mini all-in-one (including headphone out) with better power amp measurements (similar wattage with better SINAD)?

130 - 74 = 56. If one needs an inexpensive bluetooth amplifier for speakers, this seems like a decent option. I don't like that it doesn't have a remote, but that may not be important for many people.

 

[amirm]

Amir why do you measure SINAD without nominal input level of 2 volts? I see it is set as 0.18V.

Ray has given a lot of good explanations. Just to add, when a power amplifier does not have a volume control, then my only means for getting it below its max power to 5 watt for dashboard is to reduce the input level until it does that. You would be doing the same by inserting a preamplifier between the DAC and Amplifier to control volume.

With amplifiers like this one that have volume control, it presents a quandary. Where should I set the volume control on the amp? To have a fair comparison, I set the gain to roughly 29 dB which is sort of a standard although nothing is cast in concrete in the industry that way. I measure the amp using this setting and then try a couple of other combinations of input level and volume control. Usually the difference is very small (1 db or so). If it is not, I note it in the review usually.

Usually amplifier performance is determined by distortion and there, the volume control position doesn't make much of a difference. If the distortion is extremely low, then noise sets SINAD and there, position of volume control matters more.

So in summary, I am throwing amps with and without volume control in one bucket and that requires a bit of fiddling to get roughly equal measurements.

 

[scooter]

@amirm yes now I'm more informed, another lesson learnt from ASR today
I know that you always state if any major variables are observed during the measurements.
Thank you for doing these reviews.

 

[somebodyelse]

The similarly priced FX502S-Pro looks rather good in comparison, in the middle of the SINAD results rather than at the bottom, similar power and much lower distortion. It's got a similar dropping HF response too...

What's the load when the frequency response is measured? The datasheet shows the load affecting the Q of the output filter, with a boost or cut of a couple of dB with reasonable loads either side of the design load.

 

[restorer-john]

Or am I wrong again?

You're not wrong, but Amir's explanation above in attempting to level the playing field, is right on the money.

Amir has also confirmed when I've questioned him, that in his measurements of most gear, on the whole, there is not a lot of difference in performance between volume control position/input level variation when targeting a particular reference output voltage.

My personal experience is quite different as I test and restore a lot of vintage gear where all the rated specifications (in most cases) were either at full volume position or minimum. Large differences in residual noise, S/N, THD and frequency response variations have been observed by me and many before me, particularly when the extremes of 2.0V sources into 150mV rated front ends and various pot rotation positions.

For instance, frequency response is always measured with the pot rotation at 0dB and 1 (2.83V) watt (5W for Amir) into 8 ohms with the input level adjusted to give that. If you go the other way and turn down the volume pot, but leave the input voltage, the response plot will likely change significantly in certain designs. But you can argue, who uses the volume pot all the way around?

With digital volumes that go below and above a nominal 0dB it makes it harder again to standardize.

 

[infinitesymphony]

Compared to the ubiquity of high-SINAD gear at the $100 mark in the DAC and headphone amplifier worlds it's amazing that speaker amplifier designs have taken so long to catch up. Perhaps we'll see the trickle-down effects of Hypex and other amplifiers in a year or two.

 

[scooter]

@restorer-john thanks.

With digital volumes that go below and above a nominal 0dB it makes it harder again to standardize.

What's the root cause of going above reference 0dB on modern AVRs? Is this implemented to allow less than input sensitivity levels?
In a rough comparison, does it look like a "digital zoom" once an optical has reached its maximum on point and shoot cameras?

 

[Tks]

Compared to the ubiquity of high-SINAD gear at the $100 mark in the DAC and headphone amplifier worlds it's amazing that speaker amplifier designs have taken so long to catch up. Perhaps we'll see the trickle-down effects of Hypex and other amplifiers in a year or two.

Id rather companies just take the Benchmark route and stop wasting time with the Hypex supply issues, and make use of the THX modules.

 

[infinitesymphony]

Id rather companies just take the Benchmark route and stop wasting time with the Hypex supply issues, and make use of the THX modules.

Agreed, I forgot Benchmark used THX tech in their AHB2 power amplifier. Maybe we'll see something from Monoprice or Drop.

 

[boXem]

Id rather companies just take the Benchmark route and stop wasting time with the Hypex supply issues, and make use of the THX modules.

What supply issues are you talking about?
What is a THX module?