Topping LA90 Discrete Amplifier Review

[Dottie]

Hi! Great to hear that. I’ve got the same setup with some Dynaudio Craft, and honestly, I’m thrilled with it. I’ve also tried it with the Merlin Audio VSM and they sound amazing, and even with the legendary B&W 801 Matrix3. The Topping drives them all effortlessly; on the contrary, it keeps surprising in a good way!

Hi! In the past i used to have very expensive audio gear from for instance CH-Precission and Jeff Rowland. The CH-PRECISSION X1 power supply (13K euro) used for the dac smoked itself. I am totally cured now and appreciate the much more affordable gear from topping etc. very much too.

 

[vazakh]

I wonder if Topping plans to release a more powerful version of the La90.

 

[MBO]

I bought the LA90D, and to say I was blown away by the sound quality would be an understatement. I hadn't planned on buying these amps at all, but the power supply in my Purifi amp unfortunately blew, so I was looking for a replacement for my very heavy Canton Reference speakers. I decided to try the B200. These amps gave me such exceptional midrange and treble sound quality that I've never heard anywhere else in 30 years of high-end listening, no matter the price! Unfortunately, the bass was practically nonexistent. So I decided to try the LA90D. Yes, it has a slightly more moderate midrange and treble, but even very good Class D amplifiers couldn't deliver the same bass sound as the LA90D. I really hope Topping continues to develop and improve the concept of their LA90 and B200 amplifiers without losing quality!

 

[nanook]

View attachment 312493

Looking at the placement of the output inductors - shouldn't the crosstalk suffer a lot due to the (unnecessarily high) mutual inductance?

 

[capslock]

I bought the LA90D, and to say I was blown away by the sound quality would be an understatement. I hadn't planned on buying these amps at all, but the power supply in my Purifi amp unfortunately blew, so I was looking for a replacement for my very heavy Canton Reference speakers. I decided to try the B200. These amps gave me such exceptional midrange and treble sound quality that I've never heard anywhere else in 30 years of high-end listening, no matter the price! Unfortunately, the bass was practically nonexistent. So I decided to try the LA90D. Yes, it has a slightly more moderate midrange and treble, but even very good Class D amplifiers couldn't deliver the same bass sound as the LA90D. I really hope Topping continues to develop and improve the concept of their LA90 and B200 amplifiers without losing quality!

I must admit I am also blown away by the B200 (in quasi-AB comparison with Wiim Amp where it was very evident, and with PA5, 3e Audio A5se, Wiim Amp Ultra, where it was more subtle). I would have thought the Wiim Amp was perfectly sufficient based on the measurements, so I am not sure how to reconcile this with past experience and literature.

However, I don't see anything in Amir's measurements that would explain how the B200 could be low on bass:

 

[antcollinet]

(in quasi-AB comparison with Wiim Amp where it was very evident,

so I am not sure how to reconcile this with past experience and literature.

Pretty easy - your "Quasi AB was insufficiently controlled to eliminate human perceptive bias, or effects from insufficiently well matched levels.

 

[nanook]

@capslock: Don't take this the wrong way, but we've successfully deceived ourselves often enough during subjective listening sessions – at least that's true for me...

A real AB(X) is usually too exhausting for me, so I align the level to 1/10dB (voltage @ 1 kHz at the LS output terminals) and switch back and forth until I no longer know which device is playing.

If, after several rounds, I'm at about "guessing - chance" then I know that I can't hear any difference, at least not a relevant one.

 

[capslock]

A little too easy. I used a microphone to match levels, and I also deliberately allowed either option to be 1-2 dB too hot at times to exclude that effect. Perceptive bias is a bit harder since I didn't have a randomly controlled switch. However, I'd come back hours or even a day later by which time I didn't remember how I'd left the setup.

I did experience something like that before, but in that case one of the amps was borderline incompetent.

 

[nanook]

Ok., so you are aware of the risks to fool yourself.
Expectation-bias is indeed difficult to rule out - except you have a friend who randomly swaps the equipment for you.

 

[antcollinet]

I used a microphone to match levels,

Measuring in the acoustic domain is insufficiently accurate - you need volt meter at speaker terminals with test tone, and then get it to within about 0.1dB (most stepped volume controls don't allow this - my AVR for example has only 0.5dB resolution)

I also deliberately allowed either option to be 1-2 dB too hot at times to exclude that effect.

Which then provides a tell, allowing perceptive bias to get in the way.


Sorry - but there is no alternative, if you want to get reliable data on audibility - to doing blind testing correctly.

 

[capslock]

Well, I doubt 0.1 dB matching is necessary based on my experiments where I allowed either amp to be up to 2 dB too hot. But for what it's worth, I used a 1 kHz sine and a Norsonic professional sound level meter, so the bulk of the exeriments were matched to about 0.1 dB.

Yes, getting someone else to switch would have been great. In this case, though, it would have involved switching the streamer software and the speaker cables, so not doable.

 

[nanook]

Well, I doubt 0.1 dB matching is necessary based on my experiments where I allowed either amp to be up to 2 dB too hot.

When working for a Pro-Audio speaker company, a colleague of mine did an experiment with us:
Using FIR he compensated the phase of an active 3-way loudspeaker such that it became linear phase. He did the hidden switching and he told us when he swapped "something".
The rest us working in development (4 experienced listeners) were able to distinguish A and B fairly reliably. My colleague then admitted, that there's still a ca. 0.2dB level difference with the FIR he had generated. After perfectly aligning the level none of us was able to tell A/B apart. This is where my rigorous 0.1dB requirements comes from.
A few 1/10th of a dB difference doesn't sound noticeably louder - it just sounds better; voices more beaming, drums and bass-guitar more punchy etc.

So I fully agree with @antcollinet that there's no way around blind listening tests when assessing the sound quality.

Speakers are a lot more difficult - the room and the positioning will swamp many differences of the speakers and proper level alignment is not as straight forward as for DACs or Amps. Otoh. differences are usually so large that level alignment is not as critical.

 

[antcollinet]

Well, I doubt 0.1 dB matching is necessary based

You can doubt all you like, but human auditory system can start to detect differences with level mismatches in the 0.1 to 0.2dB range. Though at that level not perceived as a volume difference, but as a tone balance difference.

I used a 1 kHz sine and a Norsonic professional sound level meter, so the bulk of the exeriments were matched to about 0.1 dB.

Regardless of what the meters says, you cannot measure the sound that accurately - or you can, but the measurement is influenced by precise poisitioning of the meter, or even the position of other objects in the room (such as you) not to mention ambient noise sources. You might convince yourself with this - but you'll not convince people here that your test is valid levelled this way.

Yes, getting someone else to switch would have been great.

And if you are not testing fully blind - not knowing what is playing, not even having any chance to detect it unconsciously - then you cannot stop human perceptive biases influencing the result.

(in quasi-AB comparison with Wiim Amp where it was very evident,

And the fact that you percieved "very evident" differences, when the measurements say there shouldn't be any, is the strongest evidence that perceptive bias is at play. Even if there are actual audible differences to be heard, they will be tiny - incredibly difficult to detect, straining your hearing to do so. Perceptive bias on the other hand can (and routinely does) create "very evident" differences.

 

[Roland68]

Well, I doubt 0.1 dB matching is necessary based on my experiments where I allowed either amp to be up to 2 dB too hot. But for what it's worth, I used a 1 kHz sine and a Norsonic professional sound level meter, so the bulk of the exeriments were matched to about 0.1 dB.

Yes, getting someone else to switch would have been great. In this case, though, it would have involved switching the streamer software and the speaker cables, so not doable.

From my experience with hundreds of blind tests, I can tell you that you're wasting your time using sound level meters. Even a deviation of 1-2 mm, or a minimal change in angle, can cause problems.
Although we had someone with us who professionally handled very precise sound level measurements and the calibration of such devices, we had to repeat all our tests after a few months, and the results weren't identical or consistent in any way.

Using a multimeter and measuring directly at the speaker terminals is much more accurate and much faster.

We also tried the idea a long time ago of deliberately and alternately increasing the volume of the devices very precisely to simplify the comparison or to better hear the differences, but that also proved to be a dead end.

 

[mike7877]

LA90D - Still working great! Something around 2 years now?
I have two now - newer one's like a year old. Both have been lightly used;almost never clipped, always into 8 ohm nominal 5.something minimum. And the load, is mostly resistive (a sealed 2-way)
It shouldn't matter for longevity, but I've replaced their SMPS power supplies with one, shared, powerful, linear, regulated power supply.
Each amp has an extra 33,000uf capacitance about 2" from its DC input! How'd I do that? Simple: I bought a 10 pack of the screw-connectors Topping chose for power, and made my own DC cords which integrated the caps. (I have a couple more- might turn them into 66,000uf. But, actually, that might be more work than it's worth...
The linear supply (regulated very well), outputs 62V - perfect for my 63V caps. And it's just 2V under LA90D's expected input! Things worked out nearly perfect, and the 33,000uf caps were a happy coincidence (not bought for this).
So as they are, and connected to these speakers, it's almost impossible for me to get the amps to reach anywhere near a concerning temperature. Even artists I listen to with the most compressed levels, DNR probably about 6dB most of the time in most of the songs, and it will not cause heating past 50C (with peaks hitting 35-49W). "Normal" listening for me, with 85dB/W speakers and in the room I'm in (12x17ft), is 3-5 watts peak, with RMS 5-10% of that. Pretty standard I think. Anyway, for this, the LA90D usually stays between 30 and 40.
The only time I've actually experienced heating heating, was when I played low frequency sine waves near its least efficient level for an extended period of time (when I was setting up my new speakers; "woofer break-in" protocol lol.
I think, that this amp, should last a really long time before needing any kind of maintenance. Heat is the usual culprit behind failures over time. Of course, time itself is a problem, but it's mostly time at elevated temperatures doing the damage. Enough time at 25C and you're still cooked, though lol)
Using the linear regulated supply+ upgrading the ~20-22 gauge DC power line (yes it's that thin, no clue why when the rest of the dang thing is so thick)+adding extreme capacitance as close as possible to the unit, resulted in perceived better bass performance. To me. And I'd say that about 3/4 of the improvement is done by the supply itself, and the next 1/4 is the cord+cap with the supply. Also worth noting, the first 3/4 is qualitatively different than the last 1/4. Yes, they do share some characteristics (we are talking bass, here), but they are different.
If someone is inspired and wants to try upgrading their power supply, first just get the linear regulated. That way, if you don't hear a difference, or the if the improvement is too small to justify the cost, then things are simple: you just have one thing to box up and ship back. Oh, and one wire to reattach to the original power supply PCB (if you borrowed it - you should if you do, it's easy).
A couple days ago I took a picture of my system's leftmost LA90D when I was testing out my 2025 Christmas present. A couple things are off, but I thought the lighting reflecting from the shelf onto the printed "LA90 Discrete" looked pretty cool.

Long live the LA90 Discrete! And the rest of my system. And your system! And Happy New Year! It's January 6th - THE day to disassemble and put away the Christmas tree.
Hopefully 2026 goes better than it looks lol. And keep your speakers dry!

 

[J84]

I bought the LA90D, and to say I was blown away by the sound quality would be an understatement. I hadn't planned on buying these amps at all, but the power supply in my Purifi amp unfortunately blew, so I was looking for a replacement for my very heavy Canton Reference speakers. I decided to try the B200. These amps gave me such exceptional midrange and treble sound quality that I've never heard anywhere else in 30 years of high-end listening, no matter the price! Unfortunately, the bass was practically nonexistent. So I decided to try the LA90D. Yes, it has a slightly more moderate midrange and treble, but even very good Class D amplifiers couldn't deliver the same bass sound as the LA90D. I really hope Topping continues to develop and improve the concept of their LA90 and B200 amplifiers without losing quality!

Hi with the power of the LA90D with 2×SVS SB2000 pro subs fill a whole house when listening to music?

 

[terryforsythe]

Hi with the power of the LA90D with 2×SVS SB2000 pro subs fill a whole house when listening to music?

Much has to do with listening volume, speaker efficiency, speaker placement, room characteristics, whether you are listening to compressed or uncompressed music, etc. There are too many unknowns to give a definitive answer. For some people a single LA90D may be sufficient, and for some people it will not be. Keep in mind that each successive doubling of power only adds 3dB of SPL.

 

[J84]

Much has to do with listening volume, speaker efficiency, speaker placement, room characteristics, whether you are listening to compressed or uncompressed music, etc. There are too many unknowns to give a definitive answer. For some people a single LA90D may be sufficient, and for some people it will not be. Keep in mind that each successive doubling of power only adds 3dB of SPL.

Hi thanks. I plan to use the Wharfedale Linton.

 

[terryforsythe]

Hi thanks. I plan to use the Wharfedale Linton.

Those are pretty efficient speakers. Wharfedale specifies the sensitivity to be 90dB @ 2.0V/1m. Assuming that is accurate, they would be around 93dB @ 2.83V/1m. They are specified as 6 ohm nominal impedance. Interpolating Amir's output power measurements, the LA90D should output around 56W/ch with low distortion.

With two speakers pulled away from the wall at least 1m or so into the room, the maximum SPL would be around 109dB at 2m and 106dB at 3m. With the speakers against the wall, the maximum SPL would be around 112dB at 2m and 109dB at 3m.

For uncompressed music, you probably want at least 20dB of headroom for dynamics. For compressed music (typical of modern recordings) 6dB of headroom may be sufficient.

Example 1: Speakers pulled into the room, 3m from listening position, and listening to compressed music (-20dB for headroom): You could listen at 86dB nominal SPL at 3m listening position. That is louder than I ever listen to music.

Example 2: Speakers against the wall (+3dB due to boundary reinforcement), 2m from listening position, and listening to uncompressed muisic (-6dB for headroom): You could listen at 106dB nominal SPL at 2m listening position. That level of SPL for prolonged periods will cause hearing damage.

Here is a spreadsheet you can use to play with different scenarios: https://docs.google.com/spreadsheets/d/1PBFl_gYj_g6XYeHZHYhnBxNm-JjJD0FJFguxEhE_2d0/edit?usp=sharing

EDIT: Also, since you have two subwoofers, that relieves some of the load off of the LA90D.

 

[J84]

Those are pretty efficient speakers. Wharfedale specifies the sensitivity to be 90dB @ 2.0V/1m. Assuming that is accurate, they would be around 93dB @ 2.83V/1m. They are specified as 6 ohm nominal impedance. Interpolating Amir's output power measurements, the LA90D should output around 56W/ch with low distortion.

With two speakers pulled away from the wall at least 1m or so into the room, the maximum SPL would be around 109dB at 2m and 106dB at 3m. With the speakers against the wall, the maximum SPL would be around 112dB at 2m and 109dB at 3m.

For uncompressed music, you probably want at least 20dB of headroom for dynamics. For compressed music (typical of modern recordings) 6dB of headroom may be sufficient.

Example 1: Speakers pulled into the room, 3m from listening position, and listening to compressed music (-20dB for headroom): You could listen at 86dB nominal SPL at 3m listening position. That is louder than I ever listen to music.

Example 2: Speakers against the wall (+3dB due to boundary reinforcement), 2m from listening position, and listening to uncompressed muisic (-6dB for headroom): You could listen at 106dB nominal SPL at 2m listening position. That level of SPL for prolonged periods will cause hearing damage.

Here is a spreadsheet you can use to play with different scenarios: https://docs.google.com/spreadsheets/d/1PBFl_gYj_g6XYeHZHYhnBxNm-JjJD0FJFguxEhE_2d0/edit?usp=sharing

EDIT: Also, since you have two subwoofers, that relieves some of the load off of the LA90D.

Thanks. I would prefer to go with the topping B200 but it's seem you can't connect the wimm ultra streamer with it or something