Hi Folks,
Here is a very detailed but long review on the 14B SST done a number of years ago that you may find of interst.
Bryston 14B SST
A milestone in power amp design?
When I bump into various costly products and start to analyze them, the conclusion I reach is often:
”Yet another box where the manufacturer seems to have put most of the effort in design and marketing, and has the guts to charge extensively!”
When it comes to the very special manufacturer Bryston in Canada, I’ve never got that impression though – despite the fact that some of the products they make have remarkably high price tags. On the contrary, the feeling has rather been (when meeting Bryston people at trade shows, peeping inside the boxes, reading test results from Hi-Fi-magazines, not to forget listening to some systems at private users) that this is a manufacturer producing a range of exceptionally well engineered products!
I have for almost 15 years now tried to get hold of some Bryston-amplifiers to make a full test and review in MoLT, but unwillingness (and a strange unfriendly attitude) from the distributor has stopped those plans up to recently. The Swedish distributor was replaced with a Norwegian one, responsible for Scandinavia.
I like Norway, but it felt complicated to have to contact a person in another country – Norway, to perform at test in MoLt. On top of that, Norway is not a European Union member, and that complicates things even more. Frankly, I think the choice of a Scandinavian distributor (one that, according to their own saying, is not carrying any stock) in the only Scandinavian country not inside the Union, is somewhat strange.
All of a sudden, a while ago, I was contacted by an owner of a large Bryston amplifier, who had heard that I wanted to do a comprehensive test of one of the bigger Bryston amps. ”–You may borrow my amp to do the test!”, he said. So that´s just what I did! So now I know more about it – and here’s the report
Round 1:
First round with the amplifier was something of an anti climax. My hope and expectation was that this amplifier is something really extra, and the first hook up to a pair of speakers was also promising, telling that the amp was something extra. For a few days it was playing in my house, and it gave a very positive impression. Everything sounded comfortable and clean-clean-clean without sounding too clinical. More like a warm and pleasing sound, but perhaps the finest edge was missing.
Next connection was at my friend Gustaf Orest (this time also to a pair of speakers, not by using the A/B-listening equipment). This second listening environment gave the same impression, – clean, clear, effortless/pleasant, but maybe a somewhat warm and rolled off in the upper end. A brief look at the schematic (too brief as it turned out) showed that the bandwidth should be satisfactory, so we didn’t really know what to think about the impression of a rolled-off timbre.
Neither at my place nor at Gustaf´s was any A/B-listening performed between the amplifiers we are normally use and the Bryston amp (and definitely no F/E test) so we weren’t sure about the impressions. I’d like to add that my main object of comparison – an old Denon amplifier of 2 times 20 watts – does not pass the F/E test without errors, even though it performs remarkably well. If my memory still serves me right (the F/E test was made many years ago) the sound of the Denon was towards the warm end, and that made me react on the impression the Bryston was still more on to the warm side.
At Gustaf’s we compared it with a large power amplifier he has designed, which, I believe would probably pass an A/B-listening test without detecting significant colorations. But his power amp had not yet been A/B listened to, and the only conclusion we could draw so far was that the Bryston amplifier was sounding extremely clean and unstrained, but we had to proceed with the A/B-listening to determine if there were any timbral or other colorations to be found.
First A/B listening
And now the anti climax reaches its height – the amplifier was easily identified by the timbre – just in the way we thought – it put some colorations to the music that made the music sound somewhat darker than with a neutral amp. Furthermore, we identified it as slightly dynamic restrained, but from experience we know that this type of coloration can have its origin in the tonal character, even when it doesn’t sound like it is about timbre.
The coloration related to tonal imperfections are probably the most complex, mystic and difficult to analyze of all colorations, coming from an audio device playing music. Those who believe that all tonal colorations are perceived similarly as simple tone control knob shifts still have a lot to learn. J
Now it was time to take a second look at the schematics, this time a closer one, on the whole schematics – and now I discovered a rather peculiar designed output filter, and also that the intrinsic bandwidth was maybe a little bit too good...It is important that the bandwidth is not achieved at the expense of reduced slew rate margins, for example. Given the output filter component values, it seemed like it was dimensioned for a single amplifier, but this was an internally bridged design! My thought was, is this maybe a remnant from an older single amplifier design?
To the lab!
Measurements confirmed the coloration we had heard. The range above 10 kHz not only dropped a few tenths of a dB in amplitude in the dummy load, but I could also establish that the amplifier wasn’t stiff in the highest octave, i.e. the loss in amplitude was clearly load dependent. So, what to do now? My first step was to call the owner of the amplifier and explain the situation – including the ease in detecting the amp in the B/A-lstening. When having listened to my message, he said without further hesitation:
”Ok, please modify it for me to the best of standards!”.
He also mentioned another thing. He suggested I could have a chat with Bryston about it. He gave me a contact name and phone number, and I thought: ”Sure… of course they will take action when someone from the far away country of Sweden has a bunch of ideas on what to do... NOT!”
It turned out that I was all wrong. Bryston not only listened, they also, prestige less and in a very entertaining way, told me a lot of stories about various things, including how, when and why this output filter was designed. Furthermore, they immediately started redesigning it to solve the problems I had found. Very serious in other words. They asked if they could call back after having done some investigations – seen from my A/B-listening perspective, a perspective they seemed to like.
Altogether we probably spoke five or ten minutes over the phone, and I tried to explain as well as I could my view on the problems and what I believed was suitable solutions. It all ended up with them redesigning parts of the amp according to my suggestions. Not only this sample, but all amps in running production from that date. But to start from the beginning, at this time I had no amplifier from present production in my hands, the one I had was the older Swedish one borrowed for the test in MoLT.
Rebuild or new sample?
The owner of the amplifier still wanted me to rebuild, or upgrade, his amplifier to a factory standard “LTS-version” of the amplifier. In order to modify the amplifier I asked Bryston to tell the exact component brands and values, in order to have this amp identical with a factory produced amp – but Bryston preferred to send a completely new amplifier for this test (to make sure it would be correct). Very ambitious and serious, but really not necessary I thought, so I continued to ask for information on how to upgrade instead. However, they stuck to their opinion, and after a short while one new amplifier was on its way. Anyway, now it was time to start the testing again!
Round 2:
The thing I noted was that the cabinet was changed a little. In my eyes it was a change for the better, although the difference doesn’t make too much of an impression. It is easier to carry it around now due to more comfortable shapes of the handles, and the connections at the rear are now protected by two large hoops, that can be used as handles too when carrying the amp around. This time the testing started out in the lab, because I wanted to verify the amplifier was behaving as I hoped. It did! (By “hoped” I mean, “does not alter the signal in any way that makes me anticipate detection in the A/B-listening”, as was the case with the old amp. So I’m talking about MY hope for the amp. Other people might have different ideas and hopes regarding how an amplifier should behave, including different kinds of pleasant colours…)
Now the damping factor was very good and the frequency response was better than +/- 0,05 dB all the way up to 20 kHz, also when loaded with 4 ohms. Above 80 kHz the roll off was starting to get steeper, and a small ringing was seen. My guess was that this pole was there because of a little more feedback in combination with the altered output filter. Even more impressive was the distortion analysis. This is actually the best measured results I have seen, ever.
With the standard instrument settings we normally use, (100 dB vertical), it is possible to see distortion products down to –100 dB relative to the signal, i.e. 0,001% distortion. When the Bryston was put on the lab bench, the screen was blank... completely blank! Nothing but the fundamental (at 0 dB) was there. THD+n (the sum of all distortion harmonics and noise) was a little above –100 dB added together, but each harmonic by itself was well below –100 dB. That, ladies and gentlemen, is a very impressive result! At least seen from an engineer’s point of view.
How does it look, really?
The amplifier can be considered to be rather large in my opinion, but it is not giant. As a matter of fact, compared to some other High End amplifiers of 700 plus watts per channel (not cheating by running in class D or G), it is not big at all. It is built in the classic 19” format (the front panel width), i.e. 48cm wide, and the depth is just above 52cm (including handles on both front and back). The height is 20.5 cm, feet included. If my memory serves me right a Rotel 1090 is much bigger, at least height-wise. The NAD 208 is smaller of course, but not by that much. Quite a bit lighter, though.
The 14B SST has a heavy aluminum front plate with sturdy handles, sturdy enough to allow carrying the amp around quite easily, in spite of its weight. I am not sure if the handles are an option or not. This is how it looks on the Bryston homepage. Whether an amplifier is good looking or not is something that everyone has to determine by his/or herself, but it has definitely an unobtrusive, discrete look, especially compared to many rather ugly looking nasty over-designed amplifiers of today. Looking at it, it is obvious that all parts are high quality and quite costly, but still in proportion to the well engineered whole of the amp – by that I mean that my impression is that most of the efforts has been put into the inner parts and behaviors, and not the looks.
And how does it measure?
As mentioned earlier the distortion figures are overwhelming. But besides that, this is an amplifier with plenty of power output. Clipping level is just 10 watts short of reaching 700 watts in 8 ohms. With a short signal (100 ms long, 1/10 repeated signal) it delivers close to 730 watts per channel, so no one can deny this is a potent piece of equipment. As already mentioned the bandwidth was >80 kHz and with roughly a second order roll off. (I normally put the target low pass values to 2nd order @ 80 kHz , or a 1st order slope from 200 kHz). The damping factor was now more than satisfactory within the whole audio range (>100). The bandwidth at the low end after the modification, was as good as you would like every amplifier to have, with a low cut of below 0.3 Hz (the limit I believe every serious Hi-Fi apparatus should be designed to accomplish, with the possible exception of RIAA amps and loudspeakers). Looking at distortion levels alone, this is, as far as I can remember, by far the best power amplifier that has ever been hooked up to the test equipment of LTS! .
To make a long story short: For all normal levels (signal 20 – 20 000 Hz, output power from 1 nW and up to 700 W) all single distortion components are below –100 dB compared with the main signal. This is impressive in the perspective of how difficult it is for any manufacturer to achieve that level of performance.
How to connect?
The amplifier has quite some inputs and outputs, but as with all power amplifiers, it is easy to find the right ones to use. Normally there are only two input signals and two from the outputs to connect. The easiest way to go is to connect a pair of unbalanced interconnects from the preamp’s output, and a pair of loudspeaker cables from the outputs to the speakers.
What can go wrong here is if any of the two input selectors is in the wrong position, one for selecting the input and the other one for altering input sensitivity. On the test sample they were already in position “single ended” and “1 volt” respectively, and I left them in that position for the first test round (of course I checked the 2 volt sensitivity position as well and didn’t experience any problems or even any differences.) 1 volt means that 28.3 volts is produced at output for that level, and that is the THX standard for power amps. 28.3 volts RMS will result in 100 watts into 8 ohms, which of course is the reason for the standard being 28.3 times amplification. It could also be expressed as a gain of 29.0357… dB (or to be more precise, 29.0309 dB if one uses 100 watts rather than 28.3 volts).
Apart from the already mentioned switches, there is only one more that has to be regarded, and that’s the trigger/control voltage switch. The trigger signal can be anything from 4 to 12 volts, AC or DC! If you want to be able to start the amp by remote from another unit, put this switch in “external” position. I don’t. (In the switch position “internal”, the amp disregards any control signals coming in, but it can control other equipment, if you want it to.) By the way, it’s a pleasure reading the manual and it’s non-Voodoo attitude regarding recommendations of interconnect cables. It simply says: ”Use quality 100% shielded cables with gold plated connectors”. A wonderfully no nonsense statement!
One more thing regarding the rear side of the amplifier: It is also possible to down adjust the input sensitivity by up to 14 dB by means of two adjustable pots, (the trimmers that are accessible from the back). I cannot see any practical use for this feature, except maybe to compensate for a bad preamp volume pot with channel unbalance at low levels, or maybe to have the volume set higher to reduce effects of bad S/N after the pot in the preamp. Hopefully there is no need for any such tweaks.
So, how does it sound then?
To answer this question it is time to move on to listening test number 2! Now when A/B-listening a Bryston 14B SST for the second time (the new “LTS-modified” sample of course) our expectations were not high, neither low, more like....fuzzy! Would it show that everything we heard before is taken care of, or will we find new hidden characteristics (colorations) that would render the device detectable in the A/B-listening? Something not caught in the measurements we had done?
The outcome of the second A/B-listening be¬came different from ever before. To explain how, I have to start with explaining the normal procedure, and results: A “normal” A/B-listening test involves 3 to 6 listeners, listening to the two alternative signals, A (After) and B (Before) the test object. This is first done openly, i.e. all listeners knows which is A and B, for a longer period of time. During this listening, people are allowed to talk to each other about what they hear or think they hear. Thus, it’s possible to “learn from each other”, and vague characteristics that take a long time to identify for one listener, typically get identified very fast, due to the cross-communication of experiences.
When the blind ”verifying listening” thereafter takes place, it is still an option to talk to the other listeners, but of course it is difficult to draw any conclusions since it is no longer known which one is A or B. To sum it up, normally there are lots of views, ideas and opinions regarding the character of the tested amplifier after the open listening. That was not the case this time. We were sitting in open listening for well over one hour, and no one mentioned a single word about any differences they either imagined or heard. Actually, that’s the first time ever something like that has happened.
In an effort to bring clarity to the “problem”, (that it is difficult to find errors when no one has even an idea of what to listen for), we started looking for more difficult program materials for the amp to handle, but we did not manage to find anything that made anyone even imagine hearing differences when the signal had passed the amp during open listening. So - we had to throw in the towel and start the actual blind A/B-listening without having any clues regarding what to listen for.
The situation was very strange for A/B-listening since we normally have a lot of clues on what colorations to listen for from the open listening. After a long time during which none of us had presented even a guess on whether the amplifier was in the chain or not, someone suggested that probably something was wrong with the test setup? Maybe the amp had been omitted, and the switching was between B and B? Or maybe the switch didn’t really work? A thorough check was performed, and for sure it was correctly connected! Turning the amp off certainly killed the A.
We kept on listening without any changes. We switched between A and B sometimes rapidly, sometimes with longer intervals. We tried having the music running while switching, and to play shorter loops and switch so that identical sequence was heard on A and B. Nothing helped! The listeners looked at each other, and shook their heads. After half an hour without any guesses sane or wild, people were just laughing when prompted for an answer, saying “its impossible!”.
At that point we decided, in order to get any result, that the listeners had to guess - like it or not. Slightest indication of a difference - go with it! Not to any greater success though, after ten minutes of listening we got only in all five answers from two persons. One person made two of them, one wrong and one correct. The other left three answers, wrong-right-wrong.
So -after hours of listening we had not been able to reach any other conclusion than - we simply could not detect any coloration from the amplifier whatsoever. It was time for a new strategy:
Bring out the heavy artillery – the transient test signal from Studio Blue!
This signal is not a music signal, and for merciful reasons we do not normally use it when testing amplifiers, because almost no amplifier will pass it without altering it. It is more difficult than any other signal I have yet faced. The origin of the signal is something as simple as an old electronic metronome, and can be described as a “transient attack with tail and noise”. The intrinsic dynamic in the test signal is enormous, and that can be used (to play extremely loud) because the spectral character of the signal is nice to the ears, but very difficult for most electronic equipment.
The extreme playback level made possible by the ear friendly character of the signal, enables the listeners to hear artifacts very low down in level. It can’t be used with any precision, however, to uncover problems in the time domain. But when it comes to amplitude/amplitude and amplitude/frequency domain problems, it is phenomenal. This does not only apply to THD-problems, but also many different modulations, bandwidth limitations, microscopic changes in timbre, mains power/rectification leakage (signal dependent hum and noise for example) and a lot of other potential problems. All this will be disrespectfully revealed.
So, what happened? Same as before – no detection.
We A/B-listened for as long as we had energy. We tried using high levels and low through the test object, and very different listening levels as well. When evening turned into night we gave in. This is not the first time we failed to detect an amplifier with a reasonable statistical significance in A/B-listening, (it’s the second). But it is the first time in the history of LTS, that we, despite hours of testing, did not even get past five guesses, that also proved to be random. This is the first time that none of the listeners even believed that there was any audible difference whatsoever between input and output of the amplifier. That does not yield an increased scientific weight, but each and everyone has to judge for him/or herself if this indicates lack of coloration from this amplifier. Obviously the scenario does not affect a device with a lot of coloration.
Measurements as well as the listening tests show this device is some very exceptional piece of equipment, at least after the modification. Valued by the measured performance only this is by far the best power amplifier that has ever visited the LTS-tech-lab, and the listening test does not contradict this. Until the day I can afford to buy 20 of these power amps, to connect all 40 channels in series and perform another B/E-listening on the whole chain, estimating how far below the threshold of hearing the artifacts are can be nothing but a wild guess. Doing it would cost me less than 1,5 million Swedish crowns, so anyone wanting to sponsor such an experiment is more than welcome – call me, and I’ll give you my bank account number. J
The sound, part 2 – long term listening
As I had the amplifier available for many months this time I have spent many hours listening to it connected to my private equipment back home. The circle is in a sense now closed, since this is how it started with the first sample. It is difficult to judge a single device’s contribution to the experience when it is connected to a chain of other equipment. It is inevitable that the characteristics from all parts of the chain will affect the result. So what you can learn from a A/B-listening is a much more general and exact picture of how the device in test performs, and how it will behave with various other equipment.
But one thing can be said about listening to a test object in a complete system – it certainly makes the experience much more enjoyable than focusing on the difference between input and output signal as in a A/B-listening. In a system you can enjoy the music, since it is no longer about detecting colorations from the test object, but about listening to the recording – to the music! It is when you get home and are “listening as usual” to a variety of recordings, that you harvest the fruits of an amplifier, superbly transparent performing in the A/B-listening..In this case I can truly say – it sounds amazing, not to say remarkably fantastic!
If I compare with my memory of how the first sample sounded in my home, this new one sounds more fresh, open and clean, but I have my doubts about relying on memory when it comes to compare these small differences in sound. One impression that hits me though, imagination or not, is that it seems to be bigger differences among my CDs than I have ever heard earlier. And what I hear sounds more like real live music than ever before! The feeling I get, sitting with my eyes closed in my comfortable armchair listening to my chain, now with the Bryston at place, is that I at any time could stand up, step forward and shake hands with the musicians, or even bring my cello and play along with them! J
This is how fun Hi-Fi should be, always!
Value for money?
With all due respect to euphony, but is the amp really worth the value of a small car? Of course, that’s a question to which the answer involves taste, resources, priorities and maybe the size of the recording collection. If you have a system + recording collection worth, say, 3.5 million SEK the device has to increase the listening yield by 2% to be worth its price tag. If the system and collection together is worth 70,000 SEK however, a new device for 70,000 SEK has to double the listening value to be worth the money, and I can promise that won’t happen, unless the previous power amp was of appalling quality. I think it is better to focus on the number of recordings in the collection, and ask yourself how much extra for each CD an amplifier is allowed to cost, i.e. how much extra value the amp it adding to each disc. Whichever way you look at it, it is a subjective question. My opinion is that this amplifier will fit best in the best systems on the planet and together with huge recording collections, but might be misplacéd in systems that do not reach all the way anyway, i.e. where it would be more sensible to invest the 70,000 SEK another way, where the money would have a higher efficiency.
Warranty terms
One thing that makes Bryston different from most other Hi-Fi-companies, is the warranty terms. In Sweden the typical warranty is one year, even if the consumer laws say that units sold in Sweden and breaking within two years is the responsibility of the manufacturer, within limits. Some companies offer three years of warranty. Some even put in another goodwill year (often an instruction for the repair shops to use their judgement, and allow an extra year for common problems). When it comes to paint and rust on cars the promise goes way longer. The paint, they promise will not fall of for years, maybe five, and warranty against rust is often more than ten years, even if it only protects against through holes… Anyway – we are not spoiled with generous warranty terms to the point that you get a solid feeling that the manufacturer really believes in their product.
Bryston offers twenty (two zero 20) years of warranty on their amps!
I think this has to be a world record in the audio business. This is what they say in their information. How many manufacturers carry the responsibility for their products this far? Fewer than one in a thousand I’d say. Pretty nice these warranty terms anyway, if you are worried about having problems. On the other hand you have to remember to be prepared to carry the freight costs should anything happen, Canada is quite far away from Sweden...
Conclusions
For anyone dreaming of a truly transparent amplifier, this is really a very, very nice amplifier. Actually the nicest one I have ever come across! It’s not just nice, it is very powerful and expensive as well. In my opinion, it’s an amplifier that just doesn’t make sense in most systems performing less than superb, i.e. in most reasonable priced systems.
The reason is that, for most systems, a money input the size of a small car could be used more wisely than by investing in a Bryston amplifier. Not that it is something wrong with it (on the contrary – I know of no amplifier with better performance!) but most systems fall short in so many other ways, often due to the speakers or to room acoustic problems. Problems that you might like to address first if they are the dominating shortcomings of your music reproduction chain, and you have the money to spend. Or there might be new dimensions to expand the system in – maybe to a multi-channel capacity or with a video projector! J But – in a system that is “almost there”, close to the limit in every thinkable way (every way but the power amplifier) – a system where everything is wonderful, and the money already spent is in parity with this kind of investment – well if that’s the case, to my knowledge, there is just no other or better alternative than
The Bryston 14B SST power amplifier – the best amplifier I have come across, so far! It gets my highest recommendations without reservations!
Ing. Öhman
PS. Well, I forgot to mention – I bought the amplifier