Topping LA90

Latest reviews


100+ Head-Fier
Review and Analysis of Topping LA90 as Head-Amp
Pros: Lightweight and small footprint. Almost comparable to headphone amps.
Analog inputs take both XLR and TRS.
Great low-level details.
Clean and highly detailed. Slight hint of warmth and richness.
Very competitive technicality in all aspects.
Cons: Limited low impedance load driving.
Questionable aesthetics.
Potential product failures and/or safety issues.
Lacks the last bit of finesse and musicality found in TOTL amplifiers.


Earlier this year, I could finally acquire the Hifiman Susvara that I’ve been eager to own for half a decade since I was blown away at the first audition. And like most new owners of Susvara, it naturally led me to begin the quest to find the most optimal amplifier. After conducting quite a lot of auditions and several sighted/blinded AB testing, Benchmark AHB2 survived and ruled out everything else. It’s sleek and equipped with a bunch of fancy technologies such as class-H tracking rails. But most importantly it sounded beautiful to me with Susvara. The price was within my budget, too!

Is this review about the AHB2? Well, not really.

This story had an unexpected extra chapter. When I was waiting for the amp to pop up in the used market, one of my local audio pals invited me to his office to hear his newly purchased Chinese integrated amp. And.. I ended up placing the order for that ugly amp. Sigh.

Yes folks. That’s the amplifier today we will talk about. Topping LA90 2ch integrated amp.


Below is what the amp looks like (pictures from the product page). The LA90 largely re-used some of the design languages found in prior Topping gears. While I really like LA90’s small footprint -- one of the smallest footprint among 2ch integrated/power amplifiers except for class-D/T products. A couple of design points were hardly agreeable to me.


First, visually it’s a little boring and ugly to me. Compared to Topping’s other products like A90, the height got doubled while the width remained similar. That must be unavoidable to contain two amp modules with proper heat dissipation. But there’s no aesthetic support to compensate for this changed dimension.

Second, I dislike the idea to place controllable switches on the back. This dramatically increases user annoyance when they’re used. Considering this company has been good at keeping almost everything on the front, it feels even more disappointing. Needless to say, this makes the front panel more minimalistically looking in a bad way. Some companies manage this well but Topping was not one of them.

Volume knob feels moderate and the potentiometer seems very gradual for the first half of rotation.This, combined with low gain mode (10db), makes volume controls for headphones (i.e., more sensitive than most loudspeakers) easy and precise without the aid of digital or relay controls, which is nice.

Setup / Associated Gears​

I did most of the in-depth evaluations with Gustard X18 as dac and Hifiman Susvara as main headphones. The Susvara connected to LA90 via 4 pin XLR female to banana plugs adapter. While my Susvara is stock and has no physical mods, I selectively apply digital eq (upper midrange 1db boost and mid to upper treble 1db cut) when songs are harshly recorded. I also use the 8x upsampling software to feed X18.

In addition to Susvara, I used the KH-100 as secondary headphones to test LA90 (this is a heavily modded pair of Yamaha YH-100 transplanted to Sony ZX701; modded by my friend and modding guru @khbaur330162)

Sound Impression​

As per Topping’s own marketing material, the company claimed to put a lot of effort in eliminating noise and distortion. I honestly thought it a bit futile because it’s something good hi-fi systems aren’t troubled by in any case. That said, it was probably the LA90’s low noise+distortion level that revealed to me much more apparent low-level detail in already familiar tracks -- by “apparent” I meant that, after hearing the LA90 uncover previously unheard subtle details, I found I could then hear them when I returned to other amps. This made a much greater impression on me than any subjective awareness of a lower noise floor such as black background (which is also doubtlessly awesome with LA90).

One example of this was MISIA’s voice at the beginning phrase and the end phrase in “Orphans no Namida” (2015). Although I heard this track literally hundreds of times, I could not distinctly hear how her powerful breath as well as voice faded in and off well and mixed with ambient resonances until I played the same track with LA90 (+Susvara). It turns out the song started earlier and ended later than I thought. Once heard, this was audible through other amps as well when I returned to them. But until the LA90, I hadn’t noticed it that clearly.

These surprising detail retrievals were not restricted within low-level signals. As many Susvara owners might agree, Susvara is evidently capable of picking up any details in x-axis (small frequency deltas), y-axis (small amplitude deltas), and z-axis (transients trailing). I failed to nitpick LA90’s lack of supporting ability for that aspect. This amp could do such jobs no worse than the TOTL amplifiers I heard. For example, three differently sampled and mixed tracks in Aritomo Hamada and Takayuki Ishikawa’s collaborative rendition “∀” were unequivocally disentangled. Not many amps can disintegrate such a tricky mix as well.

I’m also impressed by the quality and quality of soundstage details. I threw in a bunch of classical recordings I often hear, the resulting impression was one of the most transparent, neutral, and lively presentations I’ve ever heard. I could just surrender to the warmly touching denouement in the fully holographic stages.

By contrast, the LA90’s ability to deliver high power has been questioned by some hobbyists/reviewers and I largely agree with this concern. Particularly when testing this amp with loudspeakers, it was more obvious. This happened to not only inefficient speakers but also moderately efficient open-baffle speakers (Spatial Audio M3) I’m using daily. It wasn’t too bad by any means but I couldn’t find dynamic impacts at the same level of other technicalities, which led me to guess LA90’s power output was somewhat limited.

Nevertheless, I would like to add that this was not the case for headphones driving. Headphones' relatively higher impedance (32-64 ohm for low-impedance cans vs 4-8 ohm typical loudspeakers) require less current for the same voltage swing. As a result, in my specific usage of the LA90 with Susvara and Yamaha orthodynamic headphones, the amp was also perceivably large in power output and showed zero hint of driving failure.

Tonal balance is also a curious audio parameter. I don’t think I'm too obsessive with even balances, but I’m still interested in how uneven or tilted frequency balance is and how I internalize. Please note that I am not necessarily referring to frequency response graphs. I do think overall perception of balance must be multifactorial and confounded/moderated by dynamic musical contexts as opposed to static test signals. To me, LA90 sounded a little less bright and less full in the bass than what I deem better speaker amps. This mid-centric character was acceptable because I had no trouble in enjoying smooth and unaccented highs as well as requisite extension in lows.

Before wrapping things up, I feel like I should briefly mention how this amp stacks against some of the best-received speaker amps in Susvara pairing. Because I wanted to first know what Susvara is truly capable of under cost no object conditions, I tried out almost all the amps that are frequently recommended in the Susvara thread. And after seriously auditioning most of them, AHB2 was the closest to my flavor. And it was eventually dethroned and removed from the shortlist by LA90.

Does this mean AHB2 sounded worse? Not at all. AHB2 indeed can do what LA90 does to me, almost equivalently. Honestly I think LA90 is minutely edged out in most of the important areas. Plus, AHB2 deliver high power in challenging conditions (e.g., current-hungry speakers) way more effortlessly. That being said, what LA90 still satisfies me better is slightly darker characters. Tad richer tone was also in my ally. I should admit the final decision was hugely made by cost and space saving aspects though.

By the way, all these evaluations are highly subjective and I am far from an unbiased reviewer. Readers should take a grain of salt in taking and translating any of my claims. This is a hobby where drastically different views are possible to exist for the same thing.


I do believe there is no strong “scientific” link between how dac or amp is measured and how we hear. There are often weak predictive hints that I used to focus on, but I did it very unscientifically.

Also please be aware that measurements may not be as objective as you may think. We always have to think about intentional and unintentional human biases that can affect how results would shape. This youtube video ("Why you can't trust audio measurements") explains how measurements can be distorted by humans. I sincerely recommend everyone to check it out.

With those in mind, I’d like to introduce several measurement graphs about the LA90, originally measured and published by L7Audiolab (2022). Please refer to the original blog post for more details.


Above is the relationship between amp output voltage level and THD+N ratio (i.e., harmonics + noise level) at relatively high load impedances (16-300 ohm). Zigzag pattern shown in the bottom right suggests the measurement hit the limitation of AP555. The amp can swing roughly up to 20 Vrms before being clipped. THD+N level was very low which may suggest LA90 employed very heavy and strong negative feedbacks anyway if not NFCA.

Unfortunately this graph could not separate noise and harmonics parts. But we can guess what the breakdown would look like by eyeballing individual FFT plots. 1W, 5W, and 40W at 8 ohm results were attached below (corresponding to 2.8, 6.3, 17.8 Vrms respectively).




Product Failure/Safety Concerns​

I have observed and/or experienced product failure cases in some Topping products before. In particular, one of their recent products, PA5 (class-D integrated amp) was notoriously defective after a certain period of time (problems begin to appear after 2-4 weeks of usage).

Also, I don’t think LA90 is equipped with as bullet-proof protection systems as Benchmark or Schiit’s amplifiers. Over-current protection seems to be included though. Admittedly, 2ch power amps are inherently subject to heat issues.

Fortunately, my friend’s unit and mine did not exhibit any issue during 8-9 weeks of heavy utilization. But I don’t think just two success cases can prove anything. Until clearer empirical evidence about safety/lack of failures is collected and validated, potential buyers have better stay aware.


Just because I think Topping’s LA90 is not absolutely perfect in all aspects doesn’t mean that I want to represent it as anything less than a marvelous sounding amplifier. The LA90 could draw more music from my Susvara and KH-100 than I’d anticipated, and, to my surprise, sounded better than a lot of amplifiers I heard. I can always confidently confirm LA90 is just right there with top performing integrated/power amplifiers in driving inefficient headphones. Combined with LA90’s superb sound quality with its desk-friendly footprint, extremely low noise, low heat, and semi-affordable price tag, it’s clearly an amplifier that I take over anything else any day. All things considered, I do think it a bargain. Worth shortlisting.
Last edited:
@gonzalo1004es Not at all. Not only I don't hear noise (hum, hiss, other types of random noises) but also I heard blacker background -- I typically hear grayer background with higher noise floor.
Great Review! I have a Red Wine Audio Signature 30.2 and I would like to compare the two amps.
I have an iFi IESL which I can connect to the RWA with speaker cables. Yesterday, I received a Focal Elear and it sounds marvelous! I also have Grado PS1000e, Kingsound KSH4 and Sennheiser HD600. My speakers are Focal Electra 946 (hardly listen to them because I need better room accoustic devices).
Did you DIY a XLR 4 pin to banana yourself? I am from Belgium and it is hard to find such a cable. This one here ( sells for 139 euro! And it is only 30 cm long (The standard length is 30cm, 20cm from the splitter to the bananas. I can build it in custom length, for example for a cable 3 meters long just add 18 Euros to the standard price)


There are no comments to display.