Topping LA90 Discrete Amplifier Review

[Vini darko]

This amp is not class D, so most likely it is class A/B. What is its idle (no signal) steady-state power consumption?
The power supply is rated at 64 V x 4 A = 256 W, yet in the bridged configuration it produced 279 W on the bench. For how long?

Robust all-aluminum enclosure with passive cooling​

I would be very concerned about heat dissipation and its negative impact on durability of this amp. The side heat sink fins are horizontal, which prevents good air convection. Also, heat sinks are small, about 8 x 15 cm each (see case dimensions at the above link).
Optimistically assuming a 65% efficiency, with the 256 W rated PS about 90 W is lost as heat. At the rated mono output of 220 W into 8 Ohms, about 118 W. The aluminum case may help a bit, but it does not appear to be vented.

The case is vented on back, sides and bottom. However yeah I doubt it will stay active at max power for long before the thermal protection kicks in.

 

[lc6]

The case is vented on back, sides and bottom. However yeah I doubt it will stay active at max power for long before the thermal protection kicks in.

Nope.

Another concern: The PS is said to be 64 V but the output stage is shown to be typical push-pull...

...which implies +/-32 V rails. But the main caps are rated at only 35 V, with the temp rating not shown (so, standard 85°C?)

While this is nominally OK, to maintain durability a design margin of substantially more than ~9% is prudent. Especially, in a hot unvented enclosure.

 

[Sokel]

I know, I made up many in my time, but there was something about the previous model which basically hobbled it when used from single ended sources, hence my wondering if this model might be better..

A really useful,somewhat close to real-world usage test would be the combined measurement of a balanced and a non-balanced dac of the blue area connected in turns.
That would be a definite answer.

(since it involves a VC wouldn't supose to have a level vs SINAD measurement? Or is there and I missed it?)

 

[capslock]

Actually, the original LA90 was the bigger feat. It likely combined an opamp input stage with an LM3886 output in a compound configuration and probably added some higher order loop compensation. The issue with this approach is that the LM3886 runs out of bandwidth quickly, making it hard to keep up the loop gain through the audio band.

A discrete VAS and power stage has a much better bandwidth. I am not sure whether the input stage is fully discrete, but designing one that rivals a modern opamp is not rocket science. The secret sauce is the loop compensation either way.

Curious about the power supply. Seems to be a single supply from the label. So do they use:
- output coupling cap (likely inside feedback loop)
- active ground
- internal bridge configuration already for each stereo channel?

 

[Tim Link]

Topping claims this amp will "...bring you the absolute amaze...". I'm intrigued!

 

[Vini darko]

Nope.

View attachment 278649

Another concern: The PS is said to be 64 V but the output stage is shown to be typical push-pull...

View attachment 278654

...which implies +/-32 V rails. But the main caps are rated at only 35 V, with the temp rating not shown (so, standard 85°C?)

View attachment 278652

While this is nominally OK, to maintain durability a design margin of substantially more than ~9% is prudent. Especially, in a hot unvented enclosure.

That render isn't that useful. Zeos gives a tour of the amp starting from 4:50 into the video. It has the vents I stated.

 

[TimW]

So we have, in real terms, a three input power amplifier with a 'massive' 41wpc@8R channel amplifier stage that hits the wall so hard it shuts down. Wow. Impressive. Not.

40wpc per channel was pretty much the minimum, the absolute bare minimum entry into high fidelity in the early 1970s! 40wpc amplifiers were less than $150(US) and they were a genuine, under-rated 40wpc. And that was with speakers boasting 90-95dBSPL@1W/M sensitivity. Most two ways these days are 80-87dB.

This is kid's basement stuff.

With inflation the price is about right:

When I listen to music using my Yamaha M-70, I can only get the meters up to the 2W mark with very loud listening. My speakers are 91dB efficient and dip down to 4 ohms in the mid bass region so that would throw off the meters which are based on 8 ohms. I'm sure they are not very accurate meters anyway and may not show instantaneous peaks, but still I doubt that I'm pushing very much power through the speakers and 40W would probably be more than enough.

 

[Matias]

@amirm I assume you measured as pictured, with the volume knob engaged, not bypassed, right? Thanks for the review as usual.

 

[VQR]

So we have, in real terms, a three input power amplifier with a 'massive' 41wpc@8R channel amplifier stage that hits the wall so hard it shuts down. Wow. Impressive. Not.

40wpc per channel was pretty much the minimum, the absolute bare minimum entry into high fidelity in the early 1970s! 40wpc amplifiers were less than $150(US) and they were a genuine, under-rated 40wpc. And that was with speakers boasting 90-95dBSPL@1W/M sensitivity. Most two ways these days are 80-87dB.

This is kid's basement stuff.

My guess would be Topping is aiming at people who want to spend under a grand to power typically 4 Ohm speakers, ie majority of modern speakers. Granted, 80 W then would not be that much power, albeit definitely clean power.

Adding on a second LA90D, we'd be looking at almost 300 W/ch and $1600 cost. It's more expensive than a lot of Class D amps with 500 W/ch, albeit not that off of many 1000 W/ch amps. I could see someone buying one LA90D for $300-$500 less than many SOTA Class D amps, then later getting a second for more power.

At 80 W, we's be looking at ~18 dB gain to your speakers, and with the 300 W, ~24 dB. Getting even 3 db more would be 600 W and for 6 dB more would be 1200 W. I'm not saying other SOTA amps like high powered class D amps have trivial benefits over two LA90D: 6 dB more is excellent for geting every bit of volume you'd want. I just wouldn't call the LA90D 'basement stuff.'

 

[TimW]

For those complaining about the features and form factor of this amplifier, keep in mind that Topping has designed it to work with their other products. The size matches the width and depth of their other 90 series components. There is no remote control because they want you to buy their other components with remote control for that. As they show in these marketing photos, the Pre90 preamp, D90 DAC, or A90 Discrete headphone amps can be used to add remote control to the LA90:

As for the included volume control and input switching, that is probably unnecessary. If Topping expects us to buy their preamp or DAC for remote volume control and input switching then there really isn't any point in having that built into the amp. However, I understand that they wanted to make the LA90 useable as a stand-alone product, and for some people 3 inputs and a hand operated volume control might be the perfect combination. In fact that is more features than the popular Aiyima A07 and similar desktop amps have. Personally I would want to use the amps as monoblocks and since the volume control can be bypassed, the unused features wouldn't bother me whatsoever. I have owned a vintage Sony power amplifier which was intended to be used with a matching preamp but also included a volume control and input switching.

 

[enricoclaudio]

For $1,600 (bridged), is it going to be a real AHB1 (edit: should be AHB2) killer this time?

Two things to consider when comparing Benchmark to Topping: Reliability and resell value. If you care about those two factors, then forget about Topping and go with Benchmark. If you add warranty and customer support to that, then it's a no brainer.... Guess who wins it all...

PS: Forgot to say THANKS @amirm for another GREAT and honest review!!

 

[Bleib]

Why no remote control? That makes it a big no-no for me and many others I reckon.

Seems they want you buy 2 and use a preamplifier with a remote.

 

[Billy Budapest]

Two things to consider when comparing Benchmark to Topping: Reliability and resell value. If you care about those two factors, then forget about Topping and go with Benchmark. If you add warranty and customer support to that, then it's a no brainer.... Guess who wins it all...

PS: Forgot to say THANKS @amirm for another GREAT and honest review!!

Another thing to compare—price. The Benchmark’s is much higher. You are in part paying for that warranty and customer support. You are also paying for U.S. labor and higher U.S. prices for electronic components.

 

[lc6]

That render isn't that useful. Zeos gives a tour of the amp starting from 4:50 into the video. It has the vents I stated.

Thanks for the clarification. The cold air intakes at the bottom of the case seem too small compared the the combined size of the outlet slits on the back pannel and the small "pass-through" sections of the heat sinks (rule of thumb is a ratio of 2+ because hot air is more "mobile" than the cold one). I'd be curious to learn the internal temp of this amp under sustained non-trivial load. Also, when two LA90s are placed side-by-side (as shown in manufacturer's pictures), the inner heat sinks would get insufficient air flow (which is anyway impeded by the horizontal orientation of heat sink fins, incl. the thicker bottom/top fins). This amp will likely be "cooking" itself and the nearby equipment, unless played at only a few Watts.

 

[andrew]

I hear what you are saying and agree to a large extent.
As a possible use case though, given their very low distortion at very low power I wonder whether they might not be an excellent pairing with super high-efficiency horns?

It’s not mainstream but that’s the use that I’m considering - specifically powering horns in an active system that run >300W. A single amp looks is a lot cheaper than the alternatives (AHB2, Purifi etc) albeit not as future proof and without proven reliability etc.

 

[gerG]

I love the looks and performance, but I am dubious about those heatsinks. The fins are oriented horizontal, but free convection moves vertically. Add to that the narrow spacing (prevents development of good turbulence) and it makes the slots just decoration. I have a feeling that owners will be clamping small fans to the sides of these. Make sure that you allow shelf space.

 

[amirm]

@amirm I assume you measured as pictured, with the volume knob engaged, not bypassed, right? Thanks for the review as usual.

I did start with the volume set to max. But then bypassed it and nothing changed so I stayed with it.

 

[amirm]

I love the looks and performance, but I am dubious about those heatsinks. The fins are oriented horizontal, but free convection moves vertically. Add to that the narrow spacing (prevents development of good turbulence) and it makes the slots just decoration. I have a feeling that owners will be clamping small fans to the sides of these. Make sure that you allow shelf space.

Cooling is a function of the impedance of the heatsink and its surface area. Here, they are going for the former with a very thick and heavy heatsink which then couples to the entire case.

 

[sarumbear]

I have a feeling that owners will be clamping small fans to the sides of these. Make sure that you allow shelf space.

That’s what I used to do for my Naim NAP250 amplifiers when partying

 

[lc6]

Cooling is a function of the impedance of the heatsink and its surface area. Here, they are going for the former with a very thick and heavy heatsink which then couples to the entire case.

The "impedance" (really, thermal resistance) of a heat sink should not be confused with its "thermal capacity" (thanks to it being a large block of metal that takes energy to heat up). What really matters is not instantaneous absorption of heat, but rather steady-state heat dissipation, which is a function of many factors, incl. shape, spacing and surface treatment of fins, not just the overall surface area. If this amp produces roughly 100 W of heat, those two small heat sinks with horizontal fins, even when (poorly) edge-coupled to an aluminum case, will hardly be sufficient to dissipate it to maintain a reasonable junction temp of the output transistors (let alone not cook the rest of the circuitry inside). And, this will be even more of a problem in the silver model.