Review and Measurements of Lyngdorf TDAI-3400 Amp & EQ (Part 1)

[MZKM]

0.1% translates to a SINAD of 60 dB so yes, that is definitely not a useful target to me.

For absolute audibility thresholds, yes. However, music masks distortion, which allows there to be distortion above the room’s noise floor that isn’t audible.

https://www.bksv.com/media/doc/BO0385.pdf#page4

As we can see in Fig. 6, for 100dB [EDIT: at 1kHz], the 2nd harmonic needs to be above -30dBFS to be audible, -40dB for the 3rd harmonic, -60 dBFS for the 8th harmonic, etc.

Thus, simply stating THD and not showing the spectrum of harmonics across the frequency range is not the best to do either, as the same % of THD can become more audible if the distortion is mostly higher order. And of course higher THD at 40Hz can be tolerated in relation to 4kHz, which is why seeing THD vs frequency graphs help as well.

So yeah...it’s complicated. However, I still feel 0.1% THD is a good target. Stereophile uses 1% as clipping usually as that’s what is normally stated for audibility thresholds for music, and the specs for most all the solid state amps they measure list their output at 1% THD.

 

[restorer-john]

...and the specs for most all the solid state amps they measure list their output at 1% THD...

Where are you getting this figure from? 1% is a joke. It may be being used by some unscrupulous manufacturers to inflate their power figures, but any company with a modicum of pride and aspirations to true high fidelity wouldn't entertain specifying such a poor number.

The simplest, most repeatable and level playing field way of determining output power is simple: The onset of visible waveform clipping. That way, amplifiers with deliberately high THD across their power band can be compared with ultra low THD designs.

All amplifiers run out of steam when they hit the rails. It's not necessary to pick an arbitrary THD number, especially if that number changes (stereophile with tube amplifiers for instance), just measure actual clipping and call it for what it is.

 

[soundwave76]

Wow. These measurements are really bursting many audiophile brand bubbles and showing, that price does not correlate with quality. Kudos to you @amirm (and nwavguy back in the days )!

 

[Jimster480]

Wow. These measurements are really bursting many audiophile brand bubbles and showing, that price does not correlate with quality. Kudos to you @amirm (and nwavguy back in the days )!

Yes it basically shows that you might as well buy cheap gear, because likely it is going to perform the same as expensive gear.

 

[graz_lag]

Wow... Yet another "high end top quality" brand dethroned!

It really does seem like companies are designing by marketing rather than engineering these days.

We need to consider that the markets for luxurious items do not know recessions, never ever ...
That's true for the arts, jeweleries, cosmetics, cars, hifi stuffs, among dozens of others ...

They (the folks @ Lyngdorf in this particular case) simply do not care if their finished gear does not perform at the top level in this or that area, being this particular machine full of features and so sold as a super-luxurious Hi-Fi Swiss knife, they look at its average performance.
Mr. Lyngdorf, who is an extremely well respected businessman in the Danish audio industries, for sure knows and would have the resources to do a much nicer job instead, his goal is simply different.

Ohh, last but not least, within the Hi-Fi shop auditioning rooms, this type of gear is plugged up within chains composed of as much expensive sources and speakers, which render an overall quality that makes the audiophiles - looking for luxurious brands, getting that very satisfaction they were looking for.

 

[Sander_web]

Welcome to the forum.


The owner asked me the same question. My experience with anything broken is that the effect is global and quite extreme. The issues I see of noise, ultrasonic content, etc. does not get caused by broken unit. It could however be sample to sample variation. This said, I plan to perform listening tests and if the issues don't manifest themselves, I will be sure nothing is broken.

As to contacting Lyngdorf, it is my policy to not go and chase manufacturers. I expect them to monitor the web for reviews of their products and if interested, contact me.


My testing reflects the interest of the forum. Here, people like to know if the DAC subsystem of any audio product is up to par independent of any other functionality. It is a good litmus test of engineering excellence. Also, people like to eliminate their music-centric DAC if they can with these integrated systems.


I tested S/PDIF and its performance was the same as USB from what I recall. The DAC implementation is the limit of performance here, not the input.


Oh, I did not realize it had enabled them. I will check. For my testing though, I lowered the levels more than what I measured and it made no difference. So I don't think there is an issue with clipping.

The owner questions if you had a faulty unit, I seriously do and also people over at another forum with a lively Lyngdorf section do. My stereo setup with the TDAI-3400 in the chain is as transparent and resolving as the direct headphone output of my Chord DAVE.

In having a science approach, you know as no other that N=1, especially when you encounter these abnormal results, is very poor. I don't know how many followers this forum has but publicising reviews like this can be very damaging for any brand. Of course, if your measurement hold true with a second copy of the unit then the brand has to take the consequences. In fact, it would be huge because all their amplifiers are based on this topology, the smaller brother the TDAI-2170, their power amplifier, the SDA-2400 and also their multichannel product. And I believe the Lyngdorf products are more a less a trickled down versions of the Steinway Lyngdorf products, so their reputation is at stake as well. In general (Steinway) Lyngdorf products are very well reviewed and have a good reputation.

I simply disagree with you on ranking this unit as a DAC. The DAC subsystem is very on par with let's say DAC's in the 2K-3K region but only when you use it as it was designed, as an integrated amplifier with digital inputs. Lyngdorf would even say that there is no DAC inside the TDAI besides the one that controles the analogue outputs. This is an off the shelf implementation and only intended to feed subwoofers. Those who seriously use the TDAI as a preamp, use its digital output.

I have another question: how serious are you investigating the product at hand before you start your measurements? I ask because of before-mentioned ICC function. But I also like to know if you checked if you had the latest firmware version and - more important - checked if you disabled the Wifi and Bluetooth modules. These modules are build in for convenience (to create the ultimate all-in-one) but they are heavy noise injectors when enabled so not advisable for critical listening.

 

[MZKM]

Where are you getting this figure from? 1% is a joke.

It is the most common spec for power amps (0.08% is common for surround receivers), some examples (using the Stereophile spec page):
* https://www.stereophile.com/content...tion-stereo-10-power-amplifier-specifications
* https://www.stereophile.com/content...2008-monoblock-power-amplifier-specifications
* https://www.stereophile.com/content/ps-audio-stellar-m700-monoblock-power-amplifier-specifications
* https://www.stereophile.com/content/air-tight-atm-300r-power-amplifier-specifications

The Benchmark amp I mentioned previously uses a much lower threshold as THD shoots right up after, but even on their product page it says they don’t use the common 1% THD spec.

The onset of visible waveform clipping.

Clipping isn’t the actual issue, it’s audible distortion.

 

[dkinric]

We need to consider that the markets for luxurious items do not know recessions, never ever ...
That's true for the arts, jeweleries, cosmetics, cars, hifi stuffs, among dozens of others ...

They (the folks @ Lyngdorf in this particular case) simply do not care if their finished gear does not perform at the top level in this or that area, being this particular machine full of features and so sold as a super-luxurious Hi-Fi Swiss knife, they look at its average performance.
Mr. Lyngdorf, who is an extremely well respected businessman in the Danish audio industries, for sure knows and would have the resources to do a much nicer job instead, his goal is simply different.

Ohh, last but not least, within the Hi-Fi shop auditioning rooms, this type of gear is plugged up within chains composed of as much expensive sources and speakers, which render an overall quality that makes the audiophiles - looking for luxurious brands, getting that very satisfaction they were looking for.

I agree with your point as to the value of the name brand and what it represents. Also, features, cosmetic appeal, etc. all have a value that may be out of synch with the price for many, and that's fine if there continues to be a market for that.

But I think the main point is what this website has shown us is it's not all that expensive to make a well measuring and performing product. Surely within that $6500 price there is room and resources available to apply basic engineering principles to achieve good/great measurements in addition to everything else?

If I splurge on a Ferrari, I know I have a beautiful car. But, I don't want to get beat at a light by the Corolla next to me (not that I race cars at stoplights, but you get the point). Amir's measurements are bringing this out into the light for all to see wheras it wasn't really obvious before.

 

[Krunok]

Clipping isn’t the actual issue, it’s audible distortion.

Clipping is the cause of the distortion.

 

[garbulky]

You need to peruse more specifications, as that statement is largely incorrect.

Much of the better quality gear made through the 80s and 90s specified their THDs typically around the 0.003%-0.005%, 20Hz-20KHz, both channels driven from 250mW to full rated power. Those numbers were well below the so-called 'knee' and pretty much at the bottom of their downward slope.

I stand corrected! Thank oyu John

 

[garbulky]

0.1% translates to a SINAD of 60 dB so yes, that is definitely not a useful target to me.

Any extra test I have to run takes away time from reviewing something else. So strong justification is needed for me to do more work. The information you want is on the graph.

Not an extra test. Just a blurb from the graph. Makes for easier reading. From what I gather you are looking for higher standards here therefore you aren't going to do 0.1% so I guess this own't work. However you are going to have a review of an amp without mentioning that it will produce a whole lot more power than what is stated in text. That's the trade off.

 

[amirm]

Not an extra test. Just a blurb from the graph.

That is not possible. The graph steps the input voltage at certain increments. That will invariably jump over 0.1, 1% or whatever number you pick (every amp is different with respect to gain/distortion). I will have to create a new test where the analyzer hunts for the specific number. And do that twice if you need it at two THD+N values.

 

[amirm]

The owner questions if you had a faulty unit, I seriously do and also people over at another forum with a lively Lyngdorf section do. My stereo setup with the TDAI-3400 in the chain is as transparent and resolving as the direct headphone output of my Chord DAVE.

The owner is welcome to send it in and have it checked out. I am not here to stop here. I am here to give my opinion based on years of test and repairing electronics, and measurements at hand. On that basis, nothing is broken. The specifications from Lyngdorf is actually pretty close to my distortion numbers for the amplifier for example (in one channel).

What I don't want to do is to alarm the owner, have him spend money shipping things back and forth and be without his unit for a while. But again, if he still likes to do it, he of course can.

 

[amirm]

In having a science approach, you know as no other that N=1, especially when you encounter these abnormal results, is very poor. I don't know how many followers this forum has but publicising reviews like this can be very damaging for any brand.

I test a lot of products that don't do well. Some manufacturers like Benchmark contact me, send me a second unit and help resolve the issue. Some don't care or don't know. In latter cases, it is assumed it is not an issue for them.

The least costly path by the way is for the manufacturer to run the same tests and publish them. Since my analyzer is industry standard, there is good chance they have similar units and we can help resolve differences. Unlike magazine reviews, my tests are interactive and I am happy to continue to test and help get to final answer. That is what I say at the end of every review.

Until then, yes, it is a sample of one. It is the nature of reviews of expensive products that we test only a single sample. If you or someone else owns the same unit, feel free to send it to me and I will test that.

Finally, I want to make sure you know that this is not "abnormal results." What is there is switching amplifier noise that is too close to audible band. This is the nature of the design. It is alarming to people maybe because until now, they did not know this since no one had measured any units. I consider the results disappointing but not abnormal.

 

[Deleted member 65]

... Of course, if your measurement hold true with a second copy of the unit then the brand has to take the consequences. ...

I can understand your disappointment given the good reputation Lyngdorf has, however this is the first independent measurement I've come across of their equipment. Since this is a user driven community and considering your questions whether the DUT was faulty or not I suggest you offer your unit being measured @ASR, of course @amirm needs to agree.

 

[MZKM]

I will have to create a new test where the analyzer hunts for the specific number. And do that twice if you need it at two THD+N values.

You can’t just use the xy coordinates? Like how Audioholics does it for instance, example (they report both 0.1% and 1%)? I don’t mean to make extra work for you, just telling you why I don’t like the hockey stick knee method.

 

[amirm]

You can’t just use the xy coordinates? Like how Audioholics does it for instance, example? I don’t mean to make extra work for you, just telling you why I don’t like the hockey stick knee method.

The hockey stick by definition rises quickly over a sample or two. If you want a rough number from that, I can do that but can't officially say, "this is power at 1%" when it could be 2%. Here is the actual results at hand:

There is one step past the knee and then jumps to max at 0.4% distortion.

I would have to hugely increase the resolution here to have more points which makes the test much slower and also keeps the amplifier in clipping region much longer which may damage it.

Anyway, if there is strong feeling about this I will create the correct test and run it.

 

[Samoyed]

Wow. I’m glad I didn’t buy one. Recently, I considered it.

 

[Soniclife]

Anyway, if there is strong feeling about this I will create the correct test and run it.

I like the current way, it encourages you to look at the graph and not a single number.

 

[graz_lag]

Wow. I’m glad I didn’t buy one. Recently, I considered it.

$6,000 or so saved for some better shop ...