SMSL PA-X Stereo Amplifier Review

[CedarX]

Well they built a circuit! It just didn't work.... ;x

Or maybe it did… After all, they don’t say what happens after the protection is triggered. What if bricking the device forever is the intended outcome?

 

[Old_School_Brad]

Or maybe it did… After all, they don’t say what happens after the protection is triggered. What if bricking the device forever is the intended outcome?

The tricky part about protection circuits is that it's not always clear what they're actually designed to protect...

 

[Roland68]

Surprising that the performance is so poor considering the price. Also a shame that the protection circuit didn't work... resulting in complete failure. I expected more considering the price. I figured it would be closer to the Topping B200...

The B200 is in a completely different league.
That's the difference between SMSL and Topping.
While SMSL mostly uses standard designs from component manufacturers, adapts them, and optimizes them based on measured values, Topping draws on its own developments in many areas and further optimizes them.

The SMSL amplifiers PA200, PA-X, and VMV A1 Pro are presumably all based on the EVAL_AUDAMP24 reference design/evaluation board from Infineon.
Based on the current measured values, I would always prefer my decades-old NAD 2200 and, of course, the A30a to the SMSL amplifiers PA200, PA-X, and VMV A1 Pro.

The question is also whether there is a significant sonic difference compared to the significantly cheaper SMSL A300, which delivers a measured 400 watts at 4 ohms and 220 watts at 8 ohms (1% THD) in BTL mode. Two SMSL A300s with 2 x 400 watts can be purchased for under €400.

 

[Jimster480]

The tricky part about protection circuits is that it's not always clear what they're actually designed to protect...

Maybe the protection of the circuit protected itself and the rest of the amp died instead.

 

[Old_School_Brad]

Maybe the protection of the circuit protected itself and the rest of the amp died instead.

 

[CedarX]

A sophisticated SMPS with normal semiconductors will always be superior to a simple GaN based power supply, and this is not clear to most users.

OT…
I assume “sophisticated SMPS” means more real estate. Is this one the reasons (the main reason?) we see more GaN stuff in compact USB-C chargers / SMPS (compact > “simple” but sub-optimal architecture/topology > mitigate with GaN HW)?

 

[amirm]

Here is an expert comparison of ordinary SMPS with standard MOSFETS compared to sophisticated one with GaN:

The current situation we have in class D amplifier is very sophisticated design with ordinary MOSFETs from likes of Hypex/Purifi, vs much simpler attempts with GaN. Given the high cost of GaN, and lack of need for further miniaturization, we probably won't see sophisticated GaN designs. A shame as I like to see what that effort would look like.

 

[classaistheway]

Some posters are missing some key points regarding the possible advantages of ganfets in Class D amps. Yes, the advantage is switching speed, but this has implications beyond simply being more power efficient, including:

1) Having the input carrier wave and comparator running at higher frequencies means the lag from input to output is reduced, which makes feedback easier to apply. The result can be either less feedback required for the same distortion performance, or more feedback can be used for even greater performance before stability problems arise. Some subjectivists might argue that less feedback is a good thing...

2) The higher switching speed reduces the distortion and lag created by the output H-bridge and again relaxes feedback requirements.

3) A higher speed means a smaller output filter can be used which reduces the distortion, damping factor reduction and phase lag created by the filter, again relaxing feedback requirements.

So in summary, having a class D amp operate at higher frequency is beneficial in every way possible and faster class D will become the norm.....if it is implemented properly.

NOTE I'm not saying the PA-X is necessarily a good implementation.

To put some numbers on this... A Fosi amp or Purify module I believe operates at a switching frequency of around 600kHz. I own the PA-X and have scoped the output. I can confirm the switching frequency is 1MHz as per the SMSL website, greater than 60% faster.

How does it sound? It's pretty good, but unfortunately I think SMSL have shot themselves in the foot by using a flat sounding opamp based volume control chip on the front end which cannot be bypassed in stereo mode. I wish this was not the case. As it is, I find the sound pretty good using the balanced input with an external preamp and the amp volume set to 100%. This is despite the Amir's testing which showed the singled-ended input to have lower distortion.

The amp does have a nice "big" sound which I enjoy, probably as much about having over 200W on tap as anything else.

 

[Crook]

I struggle see who would want this product at $699. That's £525, potentially plus import VAT at 20% into the UK.
For £490 I can get a Nord VL NC252 delivered, with a solid 5 year guarantee. If I do have to pay the import VAT, I can have the NC502 for similar money as this.
For me, it's a no brainer. No issue with Chinese production, I have an A7, but once these amps start pushing up into Hypex prices, I just can't see the allure.

 

[classaistheway]

Thanks for sharing that, I had never heard of them. They look great!

 

[LumiAudioNut]

Tharbamar just dropped a review of the SMSL PA-X.

During the review, he runs the unit through its passes in a real life stress test by operating some large tower speakers at extreme volumes for long periods of time. Eventually, the unit did trigger its internal safety and shut itself down (as it's supposed to!), but after letting it rest and cool down, the reviewer was able to power it back on continue the review. I bring this up because I know it suffered a recoverable fault during Amir's review. The point here is that the safety interlocks do seem to work under real life stress situations.

I'm operating a pair of PA-X's driving a pair of Magnepan 2.7i's which are horrendously inefficient, but present a nice stable resistive 4 ohm load (resistance changes very little with frequency) to the amplifier. So far, I haven't run into issues.

A picture of my setup is in a previous PA-X thread.

-Lumi

 

[mike7877]

Yes it is amazing how much power can come out of those ridiculously small and completely sealed wall warts. I think there is a chance for very efficient very small power amps somewhere, maybe battery operated boom boxes or something. These gan amps have been around quite a while, I wonder where have they been used.

When they switch faster, the transformer can be smaller, and since less current needs to be transferred each time, so can the capacitor(s)!

I bought a 45 watt GaN USB-C power supply when they were being hyped.
I'd never seen a new power supply technology for run of the mill stuff.
They said "Efficient GaN"

I expected efficiency - I don't know why I believed them, I must've been tired.
Anyway, when I got home and plugged it into my knock-off kill-a-watt, wall outlet power metering device...

When charging my Samsung android at the fast, not super-fast rate (the speed you'd get using one of the bundled chargers Samsung used to include when you spent a grand on a phone...) 17.2-17.3 watts were drawn. With the 10 year old "not advertised as efficient" standard Samsung? 17.6 watts!

Here's the best part... The power factor... Long story short, the power factor the power company likes represent current being drawn proportionally to the voltage, and always being drawn (for 100% f the waveform). Now, if you're drawing 20 watts, but only from the peaks half of the waveform, well, your power factor drops to 0.5, and you get billed for 40 watts!

So the Samsung charger's power factor? 0.78
The GaN charger's power factor? Varies, but at 17.2 watts, it's below 0.5...

I don't even have to do the math! The 0.46 vs 0.78 power factor DWARFS the 0.4 watt ,........

Calculation time: 17.6/17.2 = 102.3% more watts

17.6W * 1/0.78 = 22.6W (up approx. 28%)

17.2W * 1/0.46 = 37.4W (up approx. 217%)

Samsung charger uses ~2.3% watts with a 28% multiplier for how those watts are used
GaN charger uses +0.0% watts with a 217% multiplier for how those watts are used

The cost to charge the phone with the new, supposedly "efficient" GaN charger is 65.5% higher.
Why?
Because when you take only from the top of the waveform (say the top 50%), it's just as hard to generate that power as if you took that same amount for 100% of the time. What ends up happening is, because a lot of devices operate like this (less and less electronic power supplies as time goes on), the top of the, originally, sine wave, is lopped off, and the part immediately surrounding it, becomes a little wider. This is not enough to cause much of a problem for most stuff. Some transformers buzz more, some fans hum more, some things are a little bit less effic=ieh=n yeah.

Fun fax about Jan

 

[Chillazar]

SMSL may have very competent DACs and amplifiers but I don't see them advancing like Topping does. Their most interesting products (at least to me) right now are their CD players and everything else just seems to be repackaging of old internal designs.

 

[pjug]

Problems related to full-power testing below the AC-line supply frequency are often related to power supply stress.

Your testing might make SMSL rethink this, from the spec sheet:

View attachment 447119

I killed a SMSL A300 by shorting the output. The literature on that one says "Perfect protection circuit for overheating, overcurrent, etc.".
Looks like they don't use the word "perfect" anymore. Progress!

 

[Billy Budapest]

Amir, you killed that amp!

John Atkinson seems to make a habit of killing amps, so at least Amir is in good company.

 

[Palladium]

SMSL may have very competent DACs and amplifiers but I don't see them advancing like Topping does. Their most interesting products (at least to me) right now are their CD players and everything else just seems to be repackaging of old internal designs.

Topping is light years ahead in terms of DAC usability. SMSL can't even do auto standby, much less fixing their BT startup lag.

Ever since the D50 III came out there's like zero reason to buy SMSL DACs besides the SU-1. They are just ain't competitive at any other price point.

 

[amirm]

An update on the repair. A fuse was blown internally (it was a specialized, DIP style enclosed fuse). Replaced it, but as soon as I plugged it in, it blew a bunch of parts in the switching power supply. I was so confident that it would work that I didn't put a bulb inline with the AC input.

 

[kyuu]

An update on the repair. A fuse was blown internally (it was a specialized, DIP style enclosed fuse).

Dang, and I was going to ask early on if it was just a blown fuse but figured it couldn't be that simple.

 

[Roland68]

An update on the repair. A fuse was blown internally (it was a specialized, DIP style enclosed fuse). Replaced it, but as soon as I plugged it in, it blew a bunch of parts in the switching power supply. I was so confident that it would work that I didn't put a bulb inline with the AC input.

Unfortunately, that can happen, but at least now we know for sure that either the protection circuit wasn't working, or the defect originated somewhere else in the device.

For such cases, I would recommend an adjustable isolation transformer (you can sometimes get one on Ebay for an apple and an egg), plus an adjustable current limiter.
Especially useful/helpful for older devices or those with transport damage.

 

[obsessive9119]

Tortured the unit, ok, it failed, but other then that; There were already better measured results to such a degree that they are inaudible, and in every other respect it performed excellently — and your opinion on this is that it's disappointing. For this kind of money, performance like this has pretty much never existed — but if there's a 1.5 dB drop at 20 kHz, then you think it's terrible The amp sounds fucking great, it's a best buy, and you just pass it by — congrats.

Well, it measured worse than amps 1/3rd the cost and died. Excuse me if I’m not lining up for it…

(Mind you, Erin killed a Fosi ZA3 and I still bought one. But again: 1/3 the price of this).