My 2 class Ds are EVS 1200; dual mono, IcePower AS1200 modules (no longer made) and a LSA Voyager 350 GaN fet amp from Underwood HiFi, which occasionally come up on the used market. Great deal of you're patient, then send it to Tweak Audio for big improvement over OEM, or Underwood has a cool amp based on IcePower AS2 1200, module. Retro design with cool looking meters: On sale now
Competitive class D amp suggestions
I have been Class D fun since a few years ago when i bought my first class D amp. I like the concept, in general, and all the attractive features of this class of amplifiers. I tried 4 different ones, currently i stayed with one of them that i consider to be the best among all four amps. I do enjoy and like it. At the same time, my 5 watts SET amplifier (with more than 100 times higher distortion according to the specs) gives more natural and (surprisingly) notably cleaner sound (THD of the class D amp is 0.001). The soundstage of the class D amp is not so bad but that of the tube one is still better.
I remain attracted by class D amps though.
Any fresh suggestions on reasonably priced class D amps (i mean excluding non-reasonably priced class D amps, e.g., Merrill amplifiers)?
Any comments on non-reasonably priced class D amps are also welcome (so far i was not able to audition many class D amps and am curious if there are some which could really compete with Class A).
223 responses Add your response
@rolox Thanks for your post. Looks and sounds interesting. Would be nice to try it. |
I’m very fond of the heavily modded Trends TA10.1 I use in my main system. With a nice linear power supply, it sounds marvelous driving the 500 to 20KHz range in my trip-amped, very high sensitivity system. If you don’t need much power, that TA2024 chip has something magical, somewhat closer to SET sound than the vast majority of other class D amps. Swap a few passive components inside, by-pass the volume control (there's a switch inside on the pcb), use it with a nice preamp and feed it some clean 12V, it really is lovely. I have zero experience with the higher end, more modern offerings but have a few small class D with various chips, some of them using great components, but they fall flat on their face compared to the old Trends amp. |
Just an update on my earlier post on my LSA Voyager GAN 350 map. Mr. Walter from Underwood hifi was so kind to replace it with another new one. So far the new one works without showing similar problems. Walter has also recommend an SPL Elector SS preamplifier, which I think is a good match, it compensates a bit a lack of the body and "inanimateness" in the sound reproduction of the Voyager. |
Today is a remarkable day in my audio hobby realm - I received a new pair of Atmosphere Class D monoblocks, sent with a care. I plugged in them a few hours before I started to audition them. During the first few minutes of auditioning, my initial impression was the presence of solid low and mid-low frequencies, perception of body and space and lack of mid-high and high freqencies. The latter issue somehow disappeared relatively soon. The amplifier is in its first 2 hours of break in now. I'm already enjoying the sound and I feel that it will get gradually better. I think there is a lot of more to explore with these monoblocks. I will carry out a more substantial analysis, also comparing the amplifier with a few other ones that I'm using after I return from my trip. I hope this will happen in about 3-4 weeks. Stay tuned. |
|
Most digital products can drive it directly with ease.
Its my opinion that the specs listed on the Line Magnetic website are suspect. At such high power levels its very hard to make an output transformer that will support that kind of bandwidth and be SET at the same time. At any rate, you'll likely appreciate the extra power, and if you are using a high efficiency speaker you'll appreciate the lower noise floor of the class D as well. My speakers are 98dB and I have to have my head in the midrange horn to hear any noise from the amps at all. That translates easily to better low level detail, assuming that the amp has all its other ducks in a row. Its not going to be as lush but its not dry. It won't be as lush because its distortion is so much lower. The advantage is there is more musical detail and it does this without brightness. Distortion obscures detail, and its a simple fact that whether they admit it or not, before the onset of clipping nearly every SET made makes about 10% distortion. What this means is that any any power level an SET can make, the class D is going to be well over 2 orders of magnitude lower distortion- and so at any power level will offer more musical information. Its not dry because the distortion it makes is very much like that of a tube amplifier- with the 2nd and 3rd harmonics being far more prominent than the higher orders, but of course all at a much lower level than you can get from a tube amp. That is why I ended the last paragraph the way I did- this stuff is easy to hear. |
If you will decide to try Atma-Sphere class D amp, you may start to control it with the volume control of your DAC of course. Then you will see if you like the result and then think of a preamp (I have been using a passive preamp with class D). I think the amp is worth of trying and it comes with 15 days of trial (I ordered it just recently). |
@veerossi The designer of the Atma-Sphere GaN amp @atmasphere is on here regularly and hopefully he’ll chime in here with some very helpful suggestions and insights. I believe he now uses his new GaN amp in his own system even over his very highly-regarded OTL tube amps so that may give you a clue, but hopefully Ralph can add more meat on the bones here for you. |
For the Atma-sphere class D amp: Is a preamplifier necessary if only connecting a streamer/DAC to it? The streamer/DAC has it’s own volume control (if I want to enable it). Lumin S1. Someone I came across (Ocean) mentioned this class D amp is definitely worth checking out. Also, which preamplifiers are recommended for it (that have a remote)? Other than heat and efficiency, what do you think are the pros/cons of moving to this amplifier from an 805 tube amp (Line Magnetic LM-805ia)?
|
@twoleftears , I wonder how this integrated amp sounds with this speaker impedance correction feature. E.g., the distortion at higher order harmonics (the pick at 1k and the following ones at 2k etc) is considerably higher than that of the first, 2nd and 3rd order harmonics, they should be audible. Given that, it is really interesting how this is reflected on the SQ. The diagram is "unusual" also because it shows a big difference between the two channels where a little distortion is observed in the blue channel as there are only two blue picks. |
Here's the measurement of the Technics SU-G700M2, which may not be exactly a class D amp, or a GaNFet amp, but which is something close.
As you'll see, some of the higher order harmonics are higher than one often sees, and normally you only want to see the second and third harmonics around that level. John Atkinson commented: "The fifth harmonic in both channels, the seventh harmonic in the left channel (blue trace), and the ninth harmonic in the right channel (red trace) were almost as high in level as the second and third harmonics. Although these harmonics are low in level, lying between –80dB (0.01%) and –70dB (0.03%), they could be audible at high listening levels in the region where the ear is most sensitive." |
I used the LSA GaN Voyager in my main system for a few days. Not surprisingly, it didn't sound worse in the main system, which is a good characteristic. At the same time, the main system revealed some technical problems related to the amp. I noted this before in my second system as well but I suggested that the problem could have been in the preamp. First, the amp just does not sound well with the balanced inputs - there is a quite notable lack of the middle and high frequencies with the balanced inputs ( i tried two different quality balanced cables). Fortunately, one of the RCA inputs has no such problem. Second, a quite strange thing happens. With both XLR inputs, the left channel is too weak. And with both RCA inputs the right channel is too weak. I mean, even if i stand in front of the speaker, the sound coming from the other speaker is dominant. The preamp in my second system has the balance control, so this issue was somehow "settled" by reducing the volume in the normally sounding channel by about 40%. The T&A DAC/preamp in my main system has no balance control, so it was just impossible to use the amp with either XLR or RCA inputs. Then i used the rear switches to put the right channel in XLR and the left channel with RCA modes . This was the only way to achieve a more or less balanced sound. Third, i think the amp has some phasing problem. Apparently, it is made out of phase. But when I swapped the positive and negative speaker outputs, the amp remained out of phase. Perhaps, this is not a big problem but something may also be wrong with this. Based on my small experience with the combined XLR-RCA auditioning in my main system (thanks to the rear amp switches that permitted me to use such an option), i think there is nothing particular to complain about the sound quality of the amp (judging analytically). At the same time, there is no point in comparing it with my SET tube amp that just sounds more real, full and alive. This can be subjective. But perhaps, not. Just take one of the CDs I used to compare the SQ of the two amps. It is Keith Jarrett's "Buy buy blackbird" (hey @jjss49 !). It is a particularly well recorded CD. Although it is a studio album, one can feel the studio environment and the musicians on each of the instruments, just as if it were a live recording. The SET amp gives this feeling, the Voyager does not (so far as it is now). Perhaps, the Voyager may still improve (it should have about 110 hours of burn-in). I will be able to judge once I get it back from Underwood Hi FI. I am going to ship it for repair today.
|
@niodari Red and blue track the two stereo channels. In a well engineered component, they should pretty much overlay; more than a little divergence is indicative of a problem with components, construction, design, etc. |
I have a small update on the LSA GaN 360 Voyager amp. it should have now a bit over 100 hours of burn-in (I used it more than usually in my second system basically leaving it on most of time). I did not note any essential improvements after 80 hours or so. Perhaps, it may still get a bit better. I can say now how the Voyager compares with the Cherry Megaschino in my second system. The sound of Cherry is cleaner, more open, less "forced" and more alive and natural (even with half-working right channel, I adjust the channels taking an extreme balance position). To me, Cherry does not sound like a traditional class D amp, while Voyager sounds like a good class D amp, perhaps the best one I have heard so far except Cherry. No doubt that it is a very good value performance/price. I suggest that Atmosphere and AGD amps may sound somewhat similar to Cherry, but cannot be sure of course. For me, Cherry just sounds like a good class A tube amp, point. So pity that Tommy passed away! Soon I will move the Voyager to my main system. I will let you know about my impressions in a few days or so. @twoleftears and @atmasphere , thanks for additional explications! I did not reilase, from the beginning, that different pics were different harmonics, a very useful image. But now I do not understand what the blue curve stands for.
|
The distortion can be quite low! The tricky bit is knowing what is meant by 'warm'; to some this is a bit of extra mid bass energy. I see it as 2nd harmonic and of course it can be both. I think you'll find though if you delve deeper that those comments do not apply to all. Nelson Pass is aware of how distortion affects tonality so he is careful in his designs to make sure he's got a bit of 2nd going on to help out in that regard. |
They sounded pretty darn good and musical when I heard the older 120W/ch version. Horses for courses I suppose??
And as Ralph says here:
The speaker (and amp) had a quietness about them. Historically I have usually found it rare to find quiet speakers. It is always a sign that something magical is happening. I only have found that happening when the speakers and amp have a low distortion, so I have no reason to suspect anything else than what @atmasphere is saying about “the ear” using the harmonics to determine loudness as being likely true. Personally I do not likely opt for musicality via high distortion over having low distortion electronics and drivers in a speaker that is also not plagued by resonances. While both can sound good with most music… it just seems easier to get good piano and vocals out of a system that is lower distortion and more free of resonances. But, then again… I have only heard a limited number of systems, and never heard Atmasphere gear in person. I am intrigued by his electronics… so it is (IMO) a low risk decision. |
@niodari The first harmonic is the fundamental frequency of the note itself. So that's the big spike to the far left. Thus the other harmonics begin with the second. The reason why the same note/frequency played on a violin and a clarinet sounds different is--simplistically--because different instruments produce different combinations of harmonics... rather like different models of amplifiers! |
Let's be clear about something: 'masking' is where a louder sound (like a 2nd harmonic) can obscure a quieter sound (like a higher ordered harmonic). "Supress" suggests reducing the harmonic. So the lower orders do no suppress higher orders, but they can mask them if their amplitude is high enough in relation to the higher order(s) to be masked. The higher orders are the 5th and above.
No. I am saying that with lower distortion there is greater transparency; distortion obscures detail. @twoleftears Interesting. That 18th harmonic (9KHz) looks increased. There also appear to be some intermodulations present (inharmonc; not positioned directly on a graticule or directly in the middle of two graticules). |
Ralph your tutorials are helpful! At lease i understanding now better what is or can be happening with the sound delivered by audio equipment (why knows about electronics may also understand why is this happening). I still am not sure about some basic things. Now i suspect that the "higher order harmonics" are not second and third order harmonics but 4th, 5th etc order harmonics (just to clarify, in my earlier posts i referred to "higher order harmonics" as the second and the third order ones!). Do 4th, 5th etc order harmonics also exist? Is the goal then to suppress 4th, 5th etc order harmonics with the second and third order harmonics, or also suppress the first order harmonics ? Why a high distortion in the first order harmonics cannot suppress the 4th, 5th etc order harmonics? Independently of what is the answer, it is clear that an average of some order harmonics and hence THD makes no sense (see my earlier posts) merely because it does not take into account the rules of acoustics.
I did not understand that. Are you saying that by obscuring the details greater transparency is reached ? @twoleftears this diagram, a proof of a nearly the same behavior of the second and third harmonics is nice to see.
|
Most of them tend to have unmasked higher ordered harmonics at a low level, and because the ear uses the higher harmonics to sense sound pressure, its keenly sensitive to their presence! Also because the ear assigns a tonality to all forms of distortion, this causes most of said amps to sound harsh and bright even though they might have less higher ordered harmonics than a tube amp! This is how important the masking that the 2nd and 3rd provide can actually be. The reason so many class A and AB solid state amps have this issue is they lack the required Gain Bandwidth Product to support their feedback at all frequencies. This is why they can play good bass, since at those frequencies the feedback is enough to do the job. But at some higher frequency the amp runs out of GBP and so the feedback decreases: and with it the distortion at those frequencies rises. That is why distortion vs frequency is an important measurement. The distortion simply should not rise at all if brightness and harshness is to be avoided. There are a few A and AB amps that appear to have enough GBP to get around this problem and I think most of you know of one of them- the Benchmark- but you still have the issue of the distortion spectra itself. Most of the measurement guys that poo-poo subjective listening will tell you that the distortion is inaudible if about 100dB down. I personally doubt that; if it were true there would be no controversy around the ’sound’ of these amps. Since getting the distortion that low (< 0.005%) is pretty hard you can see that the distortion spectra still plays a role. One thing to keep in mind is that the distortion spectra at 1 Watt doesn’t tell the whole story! A classic example is SETs, which at lower power really do quite well. But when you get over -6dB of full power the higher ordered harmonics become more prominent. This causes the amp to sound ’dynamic’ at first (as power is increased) since the power needed is on the transients. So the ear reacts to the loudness cues on the transients, converting the result to ’dynamic’. I’m convinced that when audiophiles use the word ’dynamics’ to describe what they hear in an amplifier or stereo, that you can safely replace that word with ’distortion’ and not change the meaning of the conversation at all. In many class A and AB amps what happens is not quite the same- as power is increased the distortion spectra changes but the result is more harshness and brightness, since that’s a problem with them even at lower volumes.
|
Ralph, your distortion discussion is very helpful. What can you say, generally speaking about the distortion make up for class A and highly biased Class a/b solid state amplifiers? Are they generally the higher distortion 2nd harmonic type or the lower distortion 3rd harmonic type.. or have I missed the mark completely??? |
In unrelated news, Topping has recently released a new version of its D90SE DAC, called the D90LE, and that is unremarkable in the current context. What may be of interest, perhaps for educational purposes only, is that included in the new firmware is the option to simulate either tube harmonics or transistor harmonics (or neither). Specifically, just 2nd and 3rd harmonics, fixed for either option. Scroll over three quarters way down here to "Sound simulation", and there are images of the fundamental and its two prominent harmonics spikes (sadly the images are not very clear).
|
In a nutshell, if there's enough lower ordered harmonic distortions, they can mask the higher orders. The higher orders otherwise cause harshness and brightness. Either the 2nd or 3rd can predominate, but if the 3rd dominates then it has to be at a lower level to begin with- about 1/10th what you might see if the 2nd were dominant in the design. For example almost any SET makes 10% THD at full output- most of which is the 2nd harmonic. If you have an amplifier design that makes the 3rd instead of the 2nd (which can happen if the amp is fully differential and balanced, since even orders are cancelled throughout the design) then the THD will and must be much lower. This isn't that hard to achieve since a design that exhibits a dominant 3rd harmonic will tend to have less distortion overall anyway. On top of that the distortion cannot rise as frequency is increased, which is a problem with most solid state amps made (and with tube amps employing feedback). Class D offers a solution for this- you can have the distortion vs frequency look like a straight line, and you can run a lot of feedback at the same time. In this way you can achieve distortion levels that are several orders of magnitude lower than you might see in a tube amp ye the actual distortion signature (2nd and 3rd vs the higher ordered harmonics) can look quite similar to a tube amp. This allows greater transparency (since distortion obscures detail) yet the same kind of smooth mids and highs you associate with a really good tube amplifier- IMO, the best of both worlds. In case you don't get my drift, I'm drawing a very straight causal line between the distortion measurements and how the amplifier actually sounds; IOW its predictable based on sufficient measurement! The tech to do this really didn't exist 40 years ago but it does now. A lot of audiophiles don't know that yet; like so many industries tradition plays an enormous role in this. |
the numbers might get you in the right theater but you will only find the perfect seat by trial and error because the amp has to mate with the source or preamp and speakers in a way that sounds great to you.
|
So the situation is not completely hopeless. Ralph, how these distortion parameters reflect the distortion of the harmonics of each degree? I think that now i know that i am looking for an amp with small first order distortion but large second and third order distortions, roughly. But, as I have also mentioned, i don't really know how large they can or should be, no idea at all. |
Engineers have other measurement tools: distortion vs frequency (rising distortion is an indication brightness and harshness may be present), distortion spectra at one Watt and also at 06dB of full power and distortion spectra measured at different frequencies. If you know what you are looking for, these measurements can tell you how the amp will sound. I agree THD isn't a good marker. |
@noske So far, for me this is a most relevant forum to discuss this since am not aware of other possibilities and i know nothing about electronics. As i have already mentioned, i was not aware of these things before i had a look on the article which seems to be very useful and well written. An average is no good if you wish to have an information about the distortion of the harmonics of a particular degree, unless you adjust the weights respectively,. but then you will have different functions, i.e., using current terminology different "THD" for different harmonics. If ones wishes to have a single universal measure, we may give a large weight to the gain and relatively small or 0 weight to the second and third degree harmonics. But then we will practically ignore the higher order harmonics, which we know are good for may people (who likes tube stuff)! Or a manufacturer may decide which of the "THD"s specify in their products. @snapsc , you may be correct about the design strategies of the Cherry amps, Tommy also liked tubes. I think these are second and/or third order harmonics that made me enjoy so much the Megaschino, and my tube SET amp as well. I suggest that it is precisely this tube-like effect of the higher order harmonics that makes the sound so enjoyable. Creating alternative "THD" measures seems to me to be a best simple solution to reflect a more precise picture of what is really happening. In fact, we have here a kind of multiobjective and threshold optimization problem. Our first objective is to minimize the first order harmonics that is also referred to as the gain if i am not mistaken, and also we aim at somehow maximize (!) the distortion in higher order harmonics. Or we may try to maximize the distortion of 2nd and 3rd degree harmonics where we wish to keep the distortion of the first degree harmonics (the gain) no larger than a certain claimed threshold (for example, 0.001%). Redefining THD in this way, the producers of the tube amplifiers will bravely be able to put real distortions that their equipment have. I don't know about electronics and the limits of the distortions, i.e., would a very large distortion of the 2nd and 3rd order harmonics be good, and what is an ``ideal'' balance between the second and the third order harmonics? At this point, with my humble knowledge, i suggest that all this THD (and perhaps other spec parameters) theory is just a b*****t (created by engineers who do not know much about mathematics). It is a non-relevant and non-informative parameter which does not disclose actual balance between the harmonics of different orders. Electric engineers succeed to develop exceptionally sounding equipment, but the existing parameter for measuring their ``distortion'' seem to me to be useless. |
@ghdprentice, maybe I should have stated my point a different way. I know too many people who when doing comparisons think along the lines of, "Well it cost $100,000 so it must be better...". To your point, I do go for value, but I go for the highest quality of sound I can attain as well. In addition to moderately priced components I have some very expensive components in my system. All components in my system were chosen for their value, regardless of their price, and because they contribute to the overall enjoyment of my system. |
@clio09 ”Too much out there at a fraction of the cost that is a better value in my opinion.”
Yes you pointed out you are a value shopper and it is not a good value. Some people are going for the the best sound they can get. This is their market. But for them it is completely worth the cost. Hence, fr them it is a good value. Values are personal. |
Yeah, and SINAD. This is probably not the right forum to discuss issues of specific weightings as it is quite a mathmatically beguiling issue. This, I think, is understood. I look at a threshold metrics. Should they be met, then a visual of what else is presented is worthy of further examination - this is perhaps three or four other parameters, on a good day. And this is becoming reasonably well understood by many inquisitive folk, with thanks to a small number (one?) of .educational resources that did not exist until recent years, Plus valued contributions by exceptionally knowledgeable people.
|
| Post removed |
I have just realized that i don't really know which order harmonics THD measures. My rough believe now is that if it is high for 2nd and 3rd order harmonics then its fine (i found the article from your latest link particularly useful and well argumented).
We will now never know that... I suggest this can be a best way to deal with the distortions! Such an SS amp could also sound like a tube amp. |
@noske We recently documented that our noise emissions meet EU directives; we always had intention of selling overseas. |
@mglik, the Berning Quadrature Z are $33k list. I'm not one of those people that feels the higher the cost the better the sound. While the Berning amps are quite good, and I respect David Berning as an innovative audio engineer, I'd never pay $33k for them. Too much out there at a fraction of the cost that is a better value in my opinion. So yes, the Atma-Sphere Class D were right there with them, so much so that I packed the Berning Quadratures up today to get them ready for sale, to be followed by the Atma-Sphere M-60s. I am keeping the Atma-Sphere Class D, as well as my Music Reference OTL-1, RM-10, and RM-9 SE amps. |
This can be done with digital signal processing, or something similar. I think its a good idea - that is I think what Nelson Pass used on test subjects to reach his optimal tweaking of included harmonics. I don't know how difficult it would be to realistically implement in an analogue amp intended for home use.. Probably would be expensive - just a wild guess.
|
The interesting thing from the March Audio review/test is that although the distortion is very low, the remaining distortion is mostly 2nd and 3rd. The question becomes "although it has great bass and an extended top end, is it musical"...or is possible that the distortion possibly too low for optimal musicality? Of course "musical" means different things to different people...from lush harmonics to no fatigue to open and airy...and then there is the whole issue of how much influence is being put forward by the specific speakers and room. The important thing is that with more of these tests, people can begin to draw their own conclusions at to what might be more appealing to them individually. In my case, I never really understood what CherryAmp was doing to get a less sterile and more musical/harmonic/lush sound...now, I have a suspicion that since Tommy was an admirer of Nelson Pass, he got the 2nd and 3rd harmonics where he wanted them while driving total distortion low.
|
