Fosi Audio ZA3 - Class D Done Right

I know many of you are aware that I'm not particularly fond of Class D amplifiers and have voiced my criticisms about them. However, this perception may be a bit misleading because it really depends on the situation. I liken Class D to CVT transmissions in cars. Would I ever want a CVT in a Dodge Charger Hellcat or a Corvette? Absolutely not. But a CVT in a sporty compact with a turbocharged four-cylinder engine? Why not?

It’s all about context. I've previously mentioned I liked Tripath Class "T" amplifiers, which are another variation of Class D. The advantages of Class D amplifiers are threefold: they are lightweight, highly efficient, and cheap. They don't require massive power transformers or a dozens of output devices to generate significant power. Class D products being sold at exorbitant prices; there simply isn't enough stuff in these amplifiers to justify such costs. If you're interested, check out this discussion for more insights: Schiit Vidar vs. NC502M

A comparable example from the automotive world would be the Acura Integra Type S, which dealers are selling for around $60,000. Seriously? A front-wheel-drive, four-cylinder Honda variant at that price? That’s outrageous.

Now, let's talk about the Fosi Audio ZA3. It features the latest and greatest TPA3255 monolithic amplifier chip from Texas Instruments. The Fosi Audio ZA3 exemplifies Class D done right. It delivers impressive power (180W at 4 ohms), generates virtually no heat, is light and compact, and is incredibly affordable at $129 from Apos.

I want to give a shout-out to Apos Audio; I’ve known them for quite some time, and they've always treated me well, even though SBAF readers typically operate in a different market. If you encounter any issues, their 45-day hassle-free return policy is a breeze to work with—good luck finding that with AliExpress*. You can upgrade to better gear within a year, pass on what you bought from Apos to a friend, and they will still honor the warranty. Plus, they’re venturing into custom products in collaboration with manufacturers, like the Apos x Audio-Technica headphones that @shotgunshane began showcasing on the loaner program. Bottom line is they are they are good people, and they've help @CEE TEE out by selling the Piety amp. Anyway, that's enough about them!
​

I have planned a few scenarios to test the Fosi Audio ZA3. The first will involve using it as a speaker amplifier with moderate to high-efficiency full-range speakers, specifically the Frugalhorns equipped with Fostex 6" Sigma drivers. I’ll be pairing it with a vintage Sonic Frontiers DAC, which has been upgraded multiple times by PCX, and the reliable pi2AES will serve as the source. While this may not be the most equitable test—especially when compared to rare, meticulously restored, and modified vintage equipment specifically optimized for these speakers—benchmarking against top-tier gear always provides valuable insights.




I plan to use an adapter to test the Fosi ZA3 with several planar magnetic headphones: a modified T50RP (original version), ZMF Caldera Closed, and a rare Grade-A Audeze LCD-2r2 (the coveted unit that previously belonged to Craig of Eddie Current). Planars love powah right? So I am particularly curious how the $129 Fosi will compare against specialty planar amps such as the Ferrum (https://www.superbestaudiofriends.o...dio-oor-hypsos-impressions-and-tidbits.11754/) and Bakoon (https://www.superbestaudiofriends.o...-review-and-impressions-under-pressure.14103/). I am hoping the gain won't be too excessive. According to ASR, supposedly the gain was increased, but Amir said the measurements were worse, so the gain was decreased (moar feedback).

Finally, if I have the time, I will try the Fosi out as a amp for my e-stat setup using the iESL energizer.

To be continued...

*Also, a kitten (or a small business owner) dies every time you buy from Amazon

 

I believe the stock opamps are NE5532 which actually are not as bad as people think. (Cue in @Psalmanazar for color commentary). The deal with opamps is that the ones most people will roll are probably to color the sound. The Fosi ZA3 as it is is very tonally similar to the restored Fisher KX-200 I have sitting to the right of it. That is, the Fosi ZA3 has a emphasis in the bass whilg being toned down somewhat in the highs. The Fosi ZA3 highs are very smooth, even moreso than the Fischer KX-200, which has sharper transients (it's still a tube amp and nothing like the more horrible examples of solid-state).

Thus I feel that rolling opamps wouldn't get us much, considering that the NE5532 are what I would consider neutral sounding. I don't see rolling in OPA2132 as improving anything. Maybe some of the latest modern OPA1612 on an adapter would be interesting. I will check to see what is in my drawer and roll opamps later.

As far as rolling in those $$$ Sparkos opamps, I dunno. This seems utterly stupid given the price of the Sparkos that would MORE than double the cost of the setup. I blame YouTube idiots shills for this. One would be better off spending more to get a better amp. Don't sidegrade. Upgrade. Although I feel cheap tweaks (with with most TI opamps) or tweaks with significant demonstrable measurements as as an @atomicbob Noise Nuke would be worthy of a learning experience.

--

The voltage gain of the amp is x32 from the balanced inputs. Signal below fed with 0.1Vrms on the input. 8-ohm brick is being used as the load.


From the unbalanced inputs using 0.1Vrms, the voltage gain is x30.


x30-x32 voltage gain is super high, therefore making it not suitable for most headphone usage with an adapter. The 10 o'clock position on the volume will output about 2Vrms with typical DAC output of 2Vrms. There won't be much usable range on the volume knob and even if we attenuated the volume, we would lose a lot of dynamic range with a raised noise floor.

 

So with the included SMPS brick, we are getting about 42W into 8-ohms. This jives with the official spec of 95W into 4-ohms as we would expect about 1/2 the power going into double the impedance. (Note specs do not list power into 8-ohms, which is more typical).

Fosi ZA3
THD (no N)

Note: XLR inputs used.

There is a large sweet spot for this amp from 1/2W to 30W. This would make it suitable for many moderate efficiency speakers. This probably isn't optimal for Grados, but I bet modified Fostex, which are a lot less sensitivity than people believe (thus truly needs power), may work decently.

 

Fosi Audio ZA3
Spectrum for 1kHz into 8-ohms
L channel only, both channels driven
2V input (to simulate DAC output)
Volume adjusted to 2.83Vrms or 1W output


Tiny second harmonic spike. Rest looks good.


Here is the same as above except at 5W or 16Vrms. I chose this because 5W is typically where most amps will around with most medium efficiency speakers in people's homes. Here we start to see some grass at most harmonics with the exception of the fourth and seven harmonic where there appears to be none.

 

Next I wanted to take a look at the switcher performance. Same 1W at 1kHz into 8-ohms. Generally I'd like to see the switching frequency higher without an extraneous garbage. Most bonafide amp designers will tell you that the higher the garbage is away from the audio range, the better. However, I don't know if it matters much with Class D.

But heck, the entire package is dirt cheap. This is a better deal than modern day Corollas (which have become kind of expensive).



One thing to mention: I'd play with one of @atomicbob's noise nukes before "upgrading" with expensive opamps.

 

The Fosi ZA3 also has a mono sub output. I can see the ZA3 as ideal with some passive desktop monitors like the ELACs or Audio Engines.

The gain is for the sub out is on the low side being at 0.2. The xover frequency is a bit high at 200Hz. However all this is probably ideal for an external sub which is going to have its own gain and xover controls.


I have a small active two way speaker project coming up. I am thinking of using the Fosi for as the bass amplifier would be ideal.

 

Ah yes! I can see why the AD827 would be great for the Fosi which a bit on the insipid side. I can see the AD827 bringing some life and inciveness to the presentation. I have some AD827 stashed somewhere. Will report back.

 

There are x2 dual opamps in the Fosi correct?

However, we can get an idea the 2.83Vrms or 1W spectrum. The noise floor isn't -140db, but there's plenty of dynamic range.


Let me come back with a THD + N graph at various output levels. The power graph above as just THD.

Oh, would mean like with a 600-ohm load instead of the 4/8-ohms.