Topping LA90 Discrete Amplifier Review

[stefanovesa]

Your sources all have XLR (or TRS) out?

DAC and Pre Phono yes, for the SACD player I go with a RCA to XLR cable. Only DAC has variable output, so I have to use the LA90 potentiometer. The amp is in high gain mode.

 

[mike7877]

DAC and Pre Phono yes, for the SACD player I go with a RCA to XLR cable. Only DAC has variable output, so I have to use the LA90 potentiometer. The amp is in high gain mode.

What speakers and DAC do you have?
In most cases, the only problem possible from using the built in volume on the LA90D is the lack of gain - if a source is quiet, or a recording on a source is quiet, there is no extra gain available beyond the input

The woofer itself, no, but the circuit as measured from the binding posts, yes. I also plugged the ports, which also reduces the bass extension. It started to roll off bellow 70Hz when I measured it outside. It also has a baffle step response.

I have a shelving filter to address the baffle step response, but minimal EQ other than that. I cross it over to my subwoofer at around 100Hz and use Dirac Live for the final tuning. Mostly I have used 8th order filters, but currently I am evaluating the use of 4th order filters.

Of course, not just the woofer xD. Always good to clarify!

Nice with the filter - 8th order is pretty steep, are you just talking about the woofer here? If so, I think 4th is still viable with your 70Hz rolloff if you want to keep the same 100Hz crossover point.

All of my favourite speakers up til now (of which I know their crossover configurations, at least), have all had their tweeter>mid or tweeter>midwoofer crossover a 3rd order Butterworth. I don't know how a concentric driver configuration might play into things, but it might be worth a try for the top end. It's especially good for a full sound off axis (which I think is something I think you're striving for).

With a 4 inch/100mm driver, the maximum ideal crossover point would be 2.2kHz, though you could probably get more than acceptable results at up to 2.5kHz. I'd recommend going as low as you can where average power doesn't start heating things up, and there's still xMax for transients (obviously this changes with your preferred genre, with average power not being a consideration for most genres except metal/rock, transients always applicable -- electric guitar can have some heating potential lol)

The Elac's were my first speakers with a concentric driver for the midrange and tweeter. So impressed with the performance of the concentric driver as a point source, I also got a pair of KEF LS60s, which I use in my office. (The boss said no to moving the LS60s into the family room. )

Nice. Well that's her loss - more for you! haha
I like KEF - I've got two of their Kube subs for my 2.2 fronts. They're sealed, 10", 300W RMS, flat down to 22Hz (with DSP, obviously lol, or they'd be soooo inefficient!), and an xMax of somewhere around 20-25mm (one way).
My only gripe with them is not being able to switch from 24dB/oct to 12 :/
Oh, and that the 12" version wasn't available.. Though it hasn't hampered me at all - sure, they've got another 3Hz extension, but I've found those dual 10s have been able to move more than enough air

Yes. At this point the plan is to mount the measurement swivel with the speaker onto my 19ft A-frame ladder, well away from the house. Also, I am contemplating trying to further mitigate the ground reflections by building an A-frame out of two sheets of plywood, covering it with insulation, and placing that between the ladder and the microphone stand. We'll see. I am still thinking about it. The last time I used an 8ft ladder, but ground reflections were an issue. Going to 19ft should reduce the impact of the ground reflections, but I don't know if it will reduce them enough to get good data without using some sort of contraption to deflect/absorb the energy directed toward the ground.

Yes, there's the whole height thing... I'm thinking of doing it near the base of a large 80-100ft tall tree that's climb-friendly in the spring before the leaves come in. I don't know how brave I'll be, I haven't looked too closely at it lol

 

[terryforsythe]

Nice with the filter - 8th order is pretty steep, are you just talking about the woofer here? If so, I think 4th is still viable with your 70Hz rolloff if you want to keep the same 100Hz crossover point.

For a long time I had 8th LR on everything. On the subwoofer in particular, 8th order does a good job at keeping the subwoofer location inaudible, even at 100Hz. On the other hand, 4th order should exhibit a little bit better step response. Recently I started experimenting with 4th order, but I am still evaluating and not sure which works best for my setup.

I have a room mode resulting in a significant dip in the frequency response around 100Hz. To combat that, I set the crossovers on my woofers to 90Hz and on my subwoofer 110Hz, and let Dirac Live run its algorithms to generate the IIR and FIR filters. I read about using that strategy somewhere in the Dirac Live documentation, or perhaps in one of their notes, but I don't remember. Nonetheless, it works fairly well.

With a 4 inch/100mm driver, the maximum ideal crossover point would be 2.2kHz, though you could probably get more than acceptable results at up to 2.5kHz. I'd recommend going as low as you can where average power doesn't start heating things up, and there's still xMax for transients (obviously this changes with your preferred genre, with average power not being a consideration for most genres except metal/rock, transients always applicable -- electric guitar can have some heating potential lol)

Based on my initial measurement data, my model in VituixCAD showed that 2.5kHz would provide the best directivity index when using 8th order LR filters. But, I didn't measure at enough angles, which is why I am going to re-run the outdoor measurements to get a much more thorough data set. I have experimented with crossover frequencies between 1.8kHz and 2.5kHz, equalizing for each setup. Currently I am evaluating 4th order at 2kHz. Subjectively, the differences are very subtle.

 

[mike7877]

For a long time I had 8th LR on everything. On the subwoofer in particular, 8th order does a good job at keeping the subwoofer location inaudible, even at 100Hz. On the other hand, 4th order should exhibit a little bit better step response. Recently I started experimenting with 4th order, but I am still evaluating and not sure which works best for my setup.

I have a room mode resulting in a significant dip in the frequency response around 100Hz. To combat that, I set the crossovers on my woofers to 90Hz and on my subwoofer 110Hz, and let Dirac Live run its algorithms to generate the IIR and FIR filters. I read about using that strategy somewhere in the Dirac Live documentation, or perhaps in one of their notes, but I don't remember. Nonetheless, it works fairly well.

Very nice - I'll remember that trick for bass in the future.

Ah, Dirac... I only use DSP correction in my HT system. I've got Denon's X3700H though, came out the model before Dirac support. It's got Audyssey MultEQ XT32 of which I run room correction up to 300Hz. It sucks that Dirac support came the next generation with the X3800H 4800H etc. because I probably would have played with it more (I'm sure there's enough processing power in the DSP chips. They're probably the same, actually - a firmware update could have probably added Dirac. But of course, Denon won't add a purchase option of Dirac in a special firmware downlod for $100 or whatever, because then they couldn't sell as many of the new models with the same DACs and same amplifiers and same (probably) IC volume control, just updated HDMI (I think full 2.1 support instead of 1 or 2 2.1 and 6x 2.0 (whatever lol).

Based on my initial measurement data, my model in VituixCAD showed that 2.5kHz would provide the best directivity index when using 8th order LR filters. But, I didn't measure at enough angles, which is why I am going to re-run the outdoor measurements to get a much more thorough data set. I have experimented with crossover frequencies between 1.8kHz and 2.5kHz, equalizing for each setup. Currently I am evaluating 4th order at 2kHz. Subjectively, the differences are very subtle.

Hmm, 2.5kHz with 8th order - I presume with a 4th order the model would push that up another 100-200Hz?
Well at the end of the day it's how it sounds that's most important. Though this will probably be something near what measures best as well
I really think there's something to 3rd order Butterworth though, in keeping things sounding natural through more of the room, if you have time or the extra measurements aren't too numerous to be too annoying to add. I'm a bit unusual probably with how much I'm up and around my system vs. right in the optimal spot (probably a 30/70 split up/sitting) so I might place more importance on this aspect than most.
What do you listen to mostly?

 

[stefanovesa]

What speakers and DAC do you have?
In most cases, the only problem possible from using the built in volume on the LA90D is the lack of gain - if a source is quiet, or a recording on a source is quiet, there is no extra gain available beyond the input

Elac DBR62 and RME ADI-2 DAC FS.

 

[staticV3]

Elac DBR62 and RME ADI-2 DAC FS.

Easy choice then:

Disable the LA90D's volume control to reduce hiss and eliminate channel imbalance.

Set the LA90D to Low gain.

Disable any voltage limiters in the ADI-2, then use the ADI-2 to control volume, making sure that Auto Ref Level is turned on.

Use the ADI-2's 5-band PEQ + Bass/Treble control to apply speaker correction using Amir's and Erin's anechoic data, as well as room correction using your own in-room frequency response measurements.

Enable the ADI-2's Loudness compensation to get automatic tonal correction when listening especially loud/quiet.

 

[stefanovesa]

Easy choice then:

Disable the LA90D's volume control to reduce hiss and eliminate channel imbalance.

Set the LA90D to Low gain.

Disable any voltage limiters in the ADI-2, then use the ADI-2 to control volume, making sure that Auto Ref Level is turned on.

Use the ADI-2's 5-band PEQ + Bass/Treble control to apply speaker correction using Amir's and Erin's anechoic data, as well as room correction using your own in-room frequency response measurements.

Enable the ADI-2's Loudness compensation to get automatic tonal correction when listening especially loud/quiet.

Thanks for the hints. Actually I can not do that for day by day use, because a pre phono and a SACD player are connected too to LA90, and these have not a variable output. My original post was mainly to know if numbers, taken from the Amir test, take into account the potentiometer effects or not, having read somewhere here that its quality is not so good. Said that, I can witness that the quality of the amp is so high, with the pot in, that probably bypassing it would not have any audible effect.

 

[staticV3]

Thanks for the hints. Actually I can not do that for day by day use, because a pre phono and a SACD player are connected too to LA90, and these have not a variable output. My original post was mainly to know if numbers, taken from the Amir test, take into account the potentiometer effects or not, having read somewhere here that its quality is not so good. Said that, I can witness that the quality of the amp is so high, with the pot in, that probably bypassing it would not have any audible effect.

Here's what enabling the volume knob does to the output of the original LA90:

Source

The impact on the LA90 Discrete is likely identical.

If I were you, I'd feed the pre phono and SACD player into the ADI-2 DAC as well, using its Coax and Toslink inputs.

This is easy to do, using devices like the miniDSP PocketADC, miniDSP ADept, or Parks Audio Waxwing.

That way, the same frequency response corrected sound can be enjoyed regardless of the audio source.

 

[terryforsythe]

Hmm, 2.5kHz with 8th order - I presume with a 4th order the model would push that up another 100-200Hz?

4th order just shallows the slopes. The issue is trying to keep the frequency response consistantly linear at various angles off axis, which goes back to why I need more FR measurement data to put in my VituixCAD model. Tuning it only using REW only gets me so far. Don't get me wrong, it sounds fantastic, but there may be some meat left on the bone, perhaps only from a technical perspective, though. But, I enjoy the process. Tweaking and optimizing is satisying.

I really think there's something to 3rd order Butterworth though, in keeping things sounding natural through more of the room,

The issue with 3rd order is a 15 degree tilt of the polar response caused by a 90 degree phase difference in the high pass and low pass filter sections. With DSP I can adjust the time delay on the tweater or midrange to null this out, but given that the midrange and tweater themselves are almost perfectly time aligned in the concentric configuration, I prefer to use even order cross over slopes, e.g., 2nd order with reverse polarity on the tweeter, 4th order or 8th order. Keep in mind that I am all active, and thus not having to address driver impedances in the filter network.

Decades ago I designed a number of speakers using a MTM configuration. On those designs 3rd order butterworth filters, a.k.a. D'Appolito network, worked well for the midrange/tweeter crossover.

What do you listen to mostly?

I'm all over the place with my listening preferences, depending on my mood. Very eclectic. Mostly I prefer music that is new to me. I listen to "Daily Discovery" on Tidal quite a bit. I love finding gems I had never heard before. Recently I have been listening to early 1960's surf instrumentals, e.g., Apache, Pipeline and Rumble at Waikiki. Fun stuff. I tend to avoid rap, classic rock, country and hardcore, but everything else is fair game. There are some exceptions, though. My oldest son plays bass for Brett Statska (country/surf), and my middle son plays bass for Gauge Away (hardcore). I do listen to those.

 

[stefanovesa]

Here's what enabling the volume knob does to the output of the original LA90:
View attachment 506905 View attachment 506904
Source

The impact on the LA90 Discrete is likely identical.

If I were you, I'd feed the pre phono and SACD player into the ADI-2 DAC as well, using its Coax and Toslink inputs.

This is easy to do, using devices like the miniDSP PocketADC, miniDSP ADept, or Parks Audio Waxwing.

That way, the same frequency response corrected sound can be enjoyed regardless of the audio source.

Ordered miniDSP ADept

 

[fwwwk]

Sorry I am not familiar with entire topic. Are any significant flaws on long term usage of this amplifier? Any potential dangers/ safety concerns? Are there revisions of this amplifier or it is only 1 the same batch/ version? Thank You All in advance

 

[stefanovesa]

Sorry I am not familiar with entire topic. Are any significant flaws on long term usage of this amplifier? Any potential dangers/ safety concerns? Are there revisions of this amplifier or it is only 1 the same batch/ version? Thank You All in advance

Mine is working OK since February 2025. It has only, let me say, a cosmetic issue: one of the loudspeakers pots has the hole slighty oversized, so that one of the banana connector is not completely firm. If you plan to use banana terminals, I suggest the locking ones. as a did, to overcome the issue. I've chosen not to inform that dealer of the issue because the amp is working fine. My only advise is purchasing it from an authorized dealer. Probably at about 900 euros is Europe, today it is a bit overpriced for the power it delivers.

 

[terryforsythe]

Sorry I am not familiar with entire topic. Are any significant flaws on long term usage of this amplifier? Any potential dangers/ safety concerns? Are there revisions of this amplifier or it is only 1 the same batch/ version? Thank You All in advance

I have had a pair of them for over 2 years now. So far, so good.