Cambridge CXA81 MKII Amplifier Review

[Haskil]

I can only concur with your comments.

I think this is a decent integrated amplifier with many inputs, and showing more than sufficient performances to make happy owners.

Preamp output will perform well with the vast majority of amplifiers. Future owners now know why to avoid insensitive amplifiers, should they need more power. Thanks to ASR / @amirm.

Like you, I find that as it stands, this amplifier will give great satisfaction to its purchaser in most cases of use.

But in my opinion, the problem posed here by Amir's measurements and position on the subject is this:
- what would have been the retail price of this amplifier if, instead of listening to the song of the resistors and the toroidal transformer chosen by ear, Cambridge Audio had taken better care of the DAC-analog outputs part of this device to make it achieve performances superior to those of the 16/44.1?
I'm not at all sure that the price would have been higher, the panther would have been playing golf and the buyers would have been happy... even if those who buy this device are probably not regular readers of our ASR pages...

 

[Scytales]

I'm afraid that many final amplifiers are not that flexible, both class AB and D. The gain cannot be randomly adjusted as it defines the depth of the feedback (feed-forward linearization -> THD improvement) and the phase/amplitude balance for stability at high frequencies (for example, a very deep feedback may give you oscillation somewhere between 100kHz-100MHz).

Agreed.

But I think that in the context of his message, amirm used the term "gain" in its broader meaning of "level change" rather than the more narrow meaning of "amplification factor" applied in active circuit engineering.

Simple output (source, preamp) or input (preamp, power amp, intermediate devices between those two) level adjustments can be use to adjust gain structure.

In fact, serious preamplifiers or power amplifiers incorporate that kind of adjustment, generally passive attenuators independent of the amplification factor of the inside active circuitry, though sometimes, level adjustment is implement through active stage having variable amplification factor (rare).

 

[Scytales]

When the RCA plug became widespread, Europeans weren't too happy: their DIN plugs seemed better... In any case, a single cable could pass a stereo signal and a common ground for both channels, as well as a symmetrical signal. I remember articles in French technical magazines saying how the RCA plug was less good, less practical... And there was also the problem of input levels for tape recorders... But the RCA plug was not standardized, whereas DIN was accompanied by standards, again if memory serves: it was German and DIN was also a series of high-fidelity standards...
And a plug for amateurs that connects the hotspot first is still a pain in the ass... hum...

Your idea of standardizing output and input levels is a good one, but let a Frenchman who lives in a country that has imposed standards on everything it produces since Colbert, the minister of Louis XIV, tell you that if Colbert's standards initially boosted the quality of French products, which became the best of their time in many fields... in a second phase, they were their undoing, because as techniques evolved, standards didn't keep up and decline set in...

But in the current state of technology in the analog field, there's no reason for that to change, so your 2 volts/4 volts is a good idea. And your idea of a new type of connection is particularly interesting: IT is moving a long way in this field, but domestic hifi is still stuck with RCA, the wrong jack plug and banana plugs (banned in Europe for professional use because of the high voltages present at the output of amps, and hifi equipment meeting the standard is sold with plugs in each banana plug that you have to pop out to connect the plug). Or maybe hifi is lagging behind computing... how many devices have an RS232 serial port with the old plug to communicate with a computer... which in turn no longer has one...

and therefore switched to USB... Cables that let you use USB as a serial port have never worked for me: I have a TACT 2.2X, so I bought a second-hand Del XP laptop with RS 232 and its big screw pin...
So yes, I vote for a new intelligent cable, and also, oh so much, for a Wisa-type wireless protocol to enable amplified speakers to communicate with the electronics feeding them with signals...

Not taking into account the famous problem that plagued every attempt to create the one standard to rule them all :

 

[Kosimo]

A missed opportunity to improve the product to make it competitive. That's all. I agree with @amirm , just too much marketing and too little product innovation.

 

[restorer-john]

In fact, serious preamplifiers or power amplifiers incorporate that kind of adjustment, generally passive attenuators independent of the amplification factor of the inside active circuitry, though sometimes, level adjustment is implement through active stage having variable amplification factor (rare).

Pretty much every power amplifier I own (and that's quite a few) has independent level controls for each channel. All of them use pots up front to attenuate from their standard 1.0V~1.5V sensitivities for full rated power. Some have fixed/variable inputs with switching/relays which is handy.

My various preamplifiers all have rated outputs of 1.0V~1.5V and maximum outputs or 10V~15V+ or more.

Actual switchable gain power amplifiers are not unheard of, but rare. It's pretty much always done up front.

 

[Yevhen]

Agreed.

But I think that in the context of his message, amirm used the term "gain" in its broader meaning of "level change" rather than the more narrow meaning of "amplification factor" applied in active circuit engineering.

Simple output (source, preamp) or input (preamp, power amp, intermediate devices between those two) level adjustments can be use to adjust gain structure.

In fact, serious preamplifiers or power amplifiers incorporate that kind of adjustment, generally passive attenuators independent of the amplification factor of the inside active circuitry, though sometimes, level adjustment is implement through active stage having variable amplification factor (rare).

I also agree, but this solution is not suitable for the perfectionists and high-end religious minimalists

Potentiometer at the input will add some thermal noise and the unnecessarily high gain will boost this noise even further. In the end, the SINAD (mainly SNR) will probably drop by 1-2dB. And that's a tragedy/dethrone for those who was on the top of Amir's list of tested equipment

 

[restorer-john]

Potentiometer at the input will add some thermal noise and the unnecessarily high gain will boost this noise even further.

Many power amplifiers have ~50K pots up front for variable inputs. Many also have fixed inputs which bypass those pots.

 

[LTig]

[..]
Not that this would ever happen in this industry but the real solution would be to do away with RCA connectors. Come up with new connection that has a single wire digital communication path that would let the source and destination negotiate the best gain structure with ability to dynamically change it. HDCD has a crude version of this with signaling buried in the digital stream.

IMO the best solution is to feed active speakers with bit perfect digital signal and let the speaker manufacturer design the DSP to handle room EQ and crossover such that the DACs feed the power amps with the best combined SINAD. Volume control could be implemented digital or analog, via a remote.

 

[Bob101]

They won't even use it for that, since it already has a sub out. Virtually no one who buys this will use them for anything since it's a stereo amp. Yamaha A-S701 is a similar product, with similar measurements, which got a good review but did not include any preouts. ¯\_(ツ)_/¯

All A-S (below the 1000's models) Yamaha's have a mono sub out (summed?) with a non-defeatable, fixed 100 Hz low-pass filter on it, but with surprisingly high output (3.5V). Don't know how steep the filter is, but 100Hz cut-off is unnecessarily low. I am not aware of any even half-decent stand-alone sub that doesn't have it's own variable cut-off. From the 301/501/701 manual:

I'd rather have a full bandwidth stereo 3.5V pre-out than sub out. Plus a variable, defeatable high-pass filter for the speakers... we all have dreams, I guess.

From the A-S1200 and up manual/

Main in sensitivity is 1 Volt. As is pre output-level.

I've owned the A-S501 before an A-S1100 (eerily similar sound, almost indistinguishable to me in my room. The A-S1100 was more natural, confident, relaxed. Listening for longer periods, the 501 though not stressed at all, would sound always sound a little nervous).
Then I switched to the Adiophonics Purifi, for space and efficiency: an A-S1100 draws 75W idling, and my amp is on 18 hours a day. The purifi idles at 6W. Making the purifi draw 75W playing music material longer than an hour is just an open invitation to angry neighbours, cops or both. Both the Purifi and the 1100: eerliy similar sounding, again. At low volume it beats the 1100 hands down in every department esp bass control. Going louder the and detail of the purifi becomes very, very clear. The 1100 was far from vague, just more relaxed. The if the 1100is confident, Purify humiliates it. Don't ever doubt that. It never sounds like it's actually doing any work. Just driving your speakers as loud a syou want/they can. It's kind of weird having that much super-clean power for the first time.
I use a Topping D70pro sabre as a pre-amp. I only listen to digital sources, so that covers all my input needs. I use balanced out to the power amp, and SE out to the sub. Speakers Q acoustics 5040's (replacing Lintons), sub: Q's 'very average' 12inch closed sub nonetheless.
The gain of the amp in the sub was insufficient to match the sensitivity of the 5040's or the Lintons with the 1V out from the A-S 1100 preamp (Never used a sub with the 501). Now with 2.5V SE out on the Topping, I'm at 80% volume on the sub.

I really like the 5040's, despite their objectively 'strange' measurements. I like bright, plus they sound wide and deep in my small apartment. The Lintons are the better speaker, but essentially too big for my appartment, size-and sound-wise. They are on kind loan to a aquaintance. Should never have sold the A-S1100, though... we all have regrets.

 

[Hipster Doofus]

Dust in the wind, all we are is dust in the wind. …Kansas

Well I guess if you can buy a $500 anything you feel better if it’s in the top 10 percent and that the company is trying…but some folks do not like small equipment and want the weight / per dollar factor.

You can go your own way….Fleetwood mac

 

[ryanmh1]

I did not harshly rate the unit and certainly not just for the 1.6 volt limitation. It was collective measurements of the DAC which showed many areas lacking refinement. The collective results were "meh" for me and that is the overall rating I give it. A stereo, dedicated product to music, needs to do better than an AVR and it didn't. Much like a lot of its competitors. So some pressure is merited to lift the standard by which these companies design against.

I took "not recommended" and claiming they "made no improvements", perhaps, to be a bigger indictment than you intended. Integrated amplifiers are primarily used as as amplifiers, so I'm going to stick to that. So far as doing better than an AVR, nothing with a linear power supply (toroidal or EI core transformer for those following along and wondering what that means) will ever do meaningfully better than the better AVRs. The McIntosh MC462 is one of the quietest linear supply amplifiers I've ever seen measured. At 1W it is about 87dB. That improves by maybe 5 to 7dB at your 5W level, so say perhaps 92 to 94dB. Probably not coincidentally, the best you've ever measured appears to be the Accuphase E-270, which came in at... 94dB. That is arguably the reasonable limit with a linear supply. It's not unlikely no linear supply device will ever be in your blue range. Denon has (almost inconceivably) gotten their receivers down to about 88dB. That is the difference between McIntosh and Accuphase at 20 parts per million and Denon at 40 parts per million. The industry average for AVRs and integrated amplifiers with a linear supply is well over 100ppm.

Cambridge actually did "do better than an AVR" (Denon) by about 4 or 5 parts per million, at the amplifier output, and drastically better than industry norms. And they utterly smoked that Denon at 15kHz by at around150ppm from, like, oh, er, 1W on up. At full output the Cambridge just smashes the Denon by about 400ppm. Far more linear amplifier with a wildly better feedback implementation. That's perhaps some vestiges of Doug Self coming through still. Since this is a THD+N measurement, this substantially reduces the likelihood the amplifier is doing something naughty that might be audible. That's why we do those unrealistic measurements of full scale high frequency (ideally 20kHz or 19+20khz) tones: To maximally stress the amplifier. Cambridge crushes the Denon receiver here by a country mile.

Oh, and it warrants mention that Accuphase pulled off its results by allowing you to disconnect the preamp. When you hook up the preamp, it's -86dB, or 50 parts per million. In other words, this $1200 Cambridge beats a $5000 Accuphase in this whole SINAD drag race schtick. No idea how linear the Accuphase is since that was not measured back then. Probably fine, since Accuphase tends have good designs without too much audiophile "no feedback" nonsense. But Cambridge is giving you a proper feedback design on the cheap!

My problem is saying something cannot be recommended, despite it being one of the best products of its type. If amplifiers with linear supplies cannot be recommended due to poor SINAD (despite using this as any figure of merit has zero scientific basis), then just stop measuring them. But you cannot go fairly recommend some things with a linear supply and say they are fine, and other things not, despite the core performance being similar. Not on a website whose calling card is objective measurement, and certainly not when the Cambridge really is a lot better than the competition in a lot of measurable ways. The recommended Yamaha A-S701 has similar measurements, yet that too (like the Denon) has a higher likelihood for slewing induced distortion or other transient nasties since the high frequency measurements are less linear. (Audible? Who knows and who cares since our concern is technical performance, apparently, since the whole SINAD chart isn't exactly keyed to any reasonable metric of audibility either, as you full well know having taken the distortion audibility tests at Klippel's site.) The Cambridge is a full 7dB better, or about 150 parts per million better if we want to count it that way. For my money, I've always hedged my bets on linearity possibly being a much bigger deal than 1kHz distortion. As long as I'm continuing to pile it on (with my apologies), there was no attempt to test the Cambridge's "DAC performance" through the amplifier output, a courtesy the Yamaha did receive with the note that it was "how most users will use the product." And the Yamaha, on the digital input, didn't exactly do great by ASR SINAD standards (even if it was probably more than good enough, in my view).

In the end, the Yamaha was recommended based on measurement results, but the Cambridge not, and if you carefully compare the measurements, it's pretty inexplicable why that would be, except for penalizing the disliked implementation of a feature (preouts) which the Yamaha does not include at all. If there is some sort of metric that products cannot be recommended when they have a less than ideal implementation of a feature that is unlikely to be used much, that ought to be a consistently applied standard. As far as I can tell, it's not. But as an amplifier, this Cambridge is remarkable for a commercial product.

Don't worry, I'll tire of this shortly, and once again disappear and stop hassling you like I did a few years ago. And then I'll hopefully be delighted to return and see a much improved measurement regime, as I did this time. The Powercube measurements were great when they worked right, and the addition of 15kHz THD and 19+20kHz IMD tones these last couple years really is highly appreciated. Very, very valuable measurements for those of us who care about best possible circuit design and ensuring that a product isn't still adhering to 40 year old shibboleths about feedback. Manufacturers that sell stuff which could have orders of magnitude lower distortion by addition of literally $.10 worth of parts to implement a better compensation scheme really should be called on the carpet for inexcusably crap amplifier design. And the measurement updates finally do allow one to infer much of that, which is awesome. Looking at you, Denon!

Keep up the good fight.

 

[Tom C]

It just crushes the spirit to see something like this offered for sale in today’s competitive market.

 

[amirm]

I'm afraid that many final amplifiers are not that flexible, both class AB and D. The gain cannot be randomly adjusted as it defines the depth of the feedback (feed-forward linearization -> THD improvement) and the phase/amplitude balance for stability at high frequencies (for example, a very deep feedback may give you oscillation somewhere between 100kHz-100MHz).

Not talking about changing the core amplifier but the buffer/gain stage. It also doesn't need to be continuously variable.

 

[amirm]

So far as doing better than an AVR, nothing with a linear power supply (toroidal or EI core transformer for those following along and wondering what that means) will ever do meaningfully better than the better AVRs.

??? We are talking about the DAC stage, not the amp although the latter should also be revamped with current offerings in class D department.

 

[amirm]

My problem is saying something cannot be recommended, despite it being one of the best products of its type.

The "type" is stuck in the mud, completely oblivious to major advancements in both DAC and amplification by newer entrants. We are not talking about a bit behind. We are talking a mile behind:

I put those lower and upper bounds in there for a reason. A company wanting to convince a customer to get a dedicated stereo system needs to play in the region of the dashed blue. That is the "type" they need to be instead of producing what they have for literally decades.

On the DAC front, they can't possibly keep being this bad:

The dashed blue line is a $99 stereo DAC that came out some 4 years ago! They can't match it at their optimal output level let alone 2 volt.

We need to bring transparency here to customers. This class of audio gear needs a major performance boost. Our pestering of them has partially worked in AVR market. The hope is that it will work here as well. If not, it won't be long before the newer players gain name recognition and start eating their lunch.

 

[pma]

I put those lower and upper bounds in there for a reason. A company wanting to convince a customer to get a dedicated stereo system needs to play in the region of the dashed blue. That is the "type" they need to be instead of producing what they have for literally decades.

I bet that in a DBT test with properly matched level nobody would tell a difference from the Purifi, if played below clipping.

 

[Haskil]

I bet that in a DBT test with properly matched level nobody would tell a difference from the Purifi, if played below clipping.

Agree with you. But the problem really isn't there, it seems to me...

The problem is that we cannot support the lowest technological bidder in the field of high fidelity from the moment when the best technology does not cost more and therefore allows for superior performance... when well even they are not audible, in most cases.

I don't hear the difference between 16/44.1 and 24/48 or more, but I'm very happy that there are sound recordings made in 24/48 or 96... I don't differentiate between a signal/ratio noise of 90 dB versus 110 dB, but I am very happy that the more the years go by the closer we get to total transparency... I have known professional tape recorders with 3% harmonic distortion close to 0dB and a signal-to-noise ratio of 80 dB and the magnificent recordings made with it... but I'm happy that these performances have been shattered since...

A sector which does not evolve driven by the search for the best possible performance stagnates, then retreats, because from tiny renunciations to tiny renunciations, each time inaudible..., we end up offering mediocre products but crowned with great glory. audiophile... since then there is only magic to sell.

And ASR is rational, anti-magic and the little excesses that we can sometimes read in this area, and even sometimes in Amir's comments, are nothing in the ocean of disinformation that governs high fidelity... Until and including unfortunately in the words of Cambridge Audio which descends into idiophilic bullshit when the brand describes its work to develop the V2 of this integrated... made on a device that could be improved and in reality not "finished"...

The readers are old enough to sort things out and the discussions that follow the tests are interesting, stimulating and anyone of good will will learn things. And then, the relationships between audibility and measurements are discussed very, very often...

 

[Balle Clorin]

+1.6v issue is incomprehensible. This isn't some start-up company

Maybe because possibly the amplifier , as many do- reaches full rated power at 0.3V or so?
I do not understand why integrated amps often have a gain that cause clipping at so low level… 0.3V, maybe it is a volume control thing ?

How many integrated amps are really designed for 2V input ??

 

[Yevhen]

Yamaha A-S701 is a similar product, with similar measurements, which got a good review but did not include any preouts. ¯\_(ツ)_/¯

Yeah, I also think the this amp is not that bad.

Yamaha 701 and NAD 320 class AB amps are above Cambridge on Amir's list with SINAD around 90dB, but if you check distortion at frequencies higher 1kHz, Cambridge looks better.

 

[ryanmh1]

The "type" is stuck in the mud, completely oblivious to major advancements in both DAC and amplification by newer entrants. We are not talking about a bit behind. We are talking a mile behind:

I put those lower and upper bounds in there for a reason. A company wanting to convince a customer to get a dedicated stereo system needs to play in the region of the dashed blue. That is the "type" they need to be instead of producing what they have for literally decades.

I was going to address this at length, but I'll just end this with this: Ars gratia artis is one thing. I fear you're coming close to advocating mensura gratia mensurae, which has never been why objectivists claimed the subjectivists needed to measure. While the measurements have expanded, the analytical framework being applied often is so reductivist that other obvious product flaws are often missed, and exemplary performance is overlooked. Cambridge's 15kHz THD is incredibly low. So low that the amplifier is likely to be at least as "blameless" as a Hypex, while having the reliability of a linear supply. Who thinks the same way I do (to a degree)? None other than the designer of the 1ET400A:

"So an amplifier with a massive amount of feedback at bass frequencies might actually have quite little at high frequencies, because that’s where it gets difficult. What you get is high frequencies that are much more distorted than the low frequencies. Well, maybe that explains this very brittle, thin, choked kind of sound that is typically associated with negative feedback. Maybe the real cause isn’t there being too much feedback at low frequencies, but too little at high frequencies."
[...]
"That simple example shows that, very often, the standard measurements might simply cover something up that is very measurable and very glaring, but just happens to be under the radar of regular test methods."

In my view (and I may be wrong), you're measuring a lot of the right things, but often focusing on the wrong measurements. I'm just trying to gently urge a focus on some other aspects in my prior post, which are exemplary. You're trying to refocus on signal-to-noise or SINAD, which just isn't terribly important as a figure of merit.