Klipschguy Posted October 28, 2001 Share Posted October 28, 2001 Hey guys, I keep reading here on this bulletin board that solid state amps have "tons of harmonic distortion at the microvolt level" and are therefore inferior for super efficient speakers which often operate in this range. Does any body have any OBJECTIVE evidence to support this claim? I'm not trying to precipitate a difficult thread here. The claim just keeps getting posted here and I'm really wondering if it's true. Warm regards from a fellow Klipsch fan, Andy Quote Link to comment Share on other sites More sharing options...
Kevin S Posted October 28, 2001 Share Posted October 28, 2001 My recollection of the many distortion graphs that I have seen over many years in audio magazines would not bear out the high distortion at low wattage statement. But I am making this statement from memory, and there may have been exceptions that I did not take note of. ------------------ L/C/R: Klipsch Heresy II Surround: Klipsch RS-3 Subwoofers: 2 HSU-VTF-2 Pre/Pro/Tuner: McIntosh MX-132 AMP: McIntosh MC-7205 DVD: McIntosh MVP-831 CD Transport: Pioneer PD-F908 100 Disc Changer Turntable: Denon DP-72L Cassette: Nakamichi BX-1 T.V. : Mitsubishi 55905 SAT/HDTV: RCA DTC-100 Surge Protector: Monster Power HTS-5000 Quote Link to comment Share on other sites More sharing options...
djk Posted October 29, 2001 Share Posted October 29, 2001 In 1978 Yamaha came out with their second model line of receivers.They introduced the concept of NDCR, noise-distortion-clearance-range.All of their products would do less than 0.01% THD+noise from 100mW to full power.They would typically exceed this by 10dB.These were the last of the discrete transistor models.By 1980 they were using ICs and modules.Despite a 20dB improvement in specs, some of these models did not sound as good as the ones they replaced.Some models could be switched into class A at reduced power.THEY DID NOT SOUND ANY DIFFERENT.In another post on the Klipsch BB I explained the some of the major whys of the easily percieved differences between tube and solid state gear.Five years ago I designed a 25W/ch stereo amp using a $0.28 IC, four $0.05 driver transistors, and four $0.39 output devices.All Radio Shack quality parts.The output stage ran class B, no bias at all.The entire stereo amplifier(except for the power transformer)was on a 4 X 5 perf board($1.99 from RS).This included the pre-amp, heatsink for the power amp, and the filter caps(two, about half the size of your thumb).The speakers used for the audition were 95dB/W proto-types of a $10K a pair speaker.Everybody liked the sound.Somebody wanted to hear what the speakers would do with a 'real' amplifier.First came the $1000 Adcom, Carver, etc.Next came the $3000 Bryston, Muse, etc.Next came a $7000 Mark Levinson.Last came a $11,000 Rowland.While my 'thing' could not play as loud(after all the Rowland was 350W), everyone was surprised at how much louder the other amps couldn't play(by comparison)considering that they were(on average)rated at 10dB more power.But what stunned and amazed them was that the 'thing' sounded 'cleaner' and had more 'slam' than any of the other amps.The 'thing' was designed to be the heart of a computer sound system to be sold for $350(MSRP)with two front, two rear speakers, an image processor, and a powered sub.Two companies went into bankruptcy over legal issues concerning the image processor(a long story).The point of relating this is that it is possible to design and build inexpensive solid sate equipment that sounds good.IF YOU KNOW HOW TO.I am currently doing L-band RF for spread spectrum military radios because it pays. Quote Link to comment Share on other sites More sharing options...
jdm56 Posted October 29, 2001 Share Posted October 29, 2001 Perhaps one reason no tube fan has actually produced any proof of the "lower distortion at low power level" mantra, is because to do so would require objective measurement, which is not an area they want to go. Particularly in light of the fact that at these same low power levels at which they claim sonic superiority for tube designs, the fact is that noise is the dominant problem, not harmonic distortion. The distortion products of both designs tends to be masked by noise. Particularly the tubes, which are generally noisier than ss circuits. ------------------ JDMcCall Quote Link to comment Share on other sites More sharing options...
Kevin S Posted October 29, 2001 Share Posted October 29, 2001 Klipschguy, You may find my post uner "Tube Amp Neophyte" of interest. Then again, maybe not. ------------------ L/C/R: Klipsch Heresy II Surround: Klipsch RS-3 Subwoofers: 2 HSU-VTF-2 Pre/Pro/Tuner: McIntosh MX-132 AMP: McIntosh MC-7205 DVD: McIntosh MVP-831 CD Transport: Pioneer PD-F908 100 Disc Changer Turntable: Denon DP-72L Cassette: Nakamichi BX-1 T.V. : Mitsubishi 55905 SAT/HDTV: RCA DTC-100 Surge Protector: Monster Power HTS-5000 Quote Link to comment Share on other sites More sharing options...
Colin Posted October 29, 2001 Share Posted October 29, 2001 I am still digging for the charts on my Pioneer Class A reference amp to see how bad THD is below 1 watt. Like any solid state amplifier, the Total Harmonic Distortion (THD) is a U shaped curve. It is .005% at 1 watt into 8 ohms at one end, dips below .001% at 15 watts, curving back up to .01% at 30 watts at the other. In other words, ultra clean response. Technically the amp can put out more than 30 watts, but the THD climbs like a rocket once output is higher than 22 watts. Hence the M-22 designation. Below one watt, the THD curve is just as high as it is above 30 watts! (though I do not have the number in front of me) Stereophile magazine formally defines clipping as the 1% THD+noise figure. So the "higher THD at micro-watts for SS amps" arguement may only hold for the cheap low cost amps that clog the audio mainstream - I'll have to search for some specs on this, but I am quite willing to give up the battle if the arguement makes no sense ... ------------------ Cornwalls & Klipsch subs; leather couch & feet up; lights out & tubes glowing! Quote Link to comment Share on other sites More sharing options...
sunnysal Posted October 29, 2001 Share Posted October 29, 2001 I want to try djk's little amp!...I do have a question for you though djk, what would you have changed in that amp if you wanted to improve its performance? if you wanted to bring it up higher end? I am curious because guys like mr. pass and others have been telling us for years that you can build SS amp at modest prices with great sound...regards, tony Quote Link to comment Share on other sites More sharing options...
eq_shadimar Posted October 29, 2001 Share Posted October 29, 2001 My HK AVR 510 is rrated at 0.03%@1 watt stereo mode and 0.05%@1 watt 5 channel mode. The specs are listed as THD+N taken at 1KHz into 8 ohms. They also listed the 4 ohm spec but they were the same. Seems pretty low to me. Laters, ------------------ ...wake up to find out that you are the eyes of the world... My Home Theater Page Quote Link to comment Share on other sites More sharing options...
Ray Garrison Posted October 29, 2001 Share Posted October 29, 2001 Andy, The best measurement technique that I've seen that can give you a visual idea of how the distortion and noise components vary with output power, and how different amps produce different "looking" levels at very low power levels, come from Stereophile. They have a test they run, using a device called the "Miller Audio Research Amplifier Profiler", that shows the response of the amp to non-continuous signal "bursts" - a few cycles of a 1kHz tone, then off for several milliseconds, then another burst, then off, etc. This signal (IMHO) more closely represents the kind of signal an amp is going to see when being fed music than does a steady state tone. I've attached a few graphs. (I hope the Klipsch magic file uploader keeps the file names the same...) Here's what they were measuring: wla_gemini_2A3 - the Wave Length Audio Gemini amp, using a single-ended 2A3 output tube; leg_starlet - the Legend Audio Starlet 6550 based amp; omega - the Classe Omega behemoth solid state "state of the art" Casse amp; MF_NuVista - the Musical Fidelity NuVista amp using the nuvistor small signal tubettes. Take a look at the distortion curves. You'll see that the levels at higher power of the tube based amps are ridiculously high, compared to the solid state amps. However, if you look at the way the noise and distortion varies at very low levels, you'll see that the tube based amps have a very, very smoothly decreasing level of noise and distortion as the power level drops, while (these two, anyway) solid state amps have a very "hashy" low signal level output. Note that I picked these two solid state amps 'cause they show a rather extreme case of hashy low signal level performance with impeccable high powere level performance. Check out the Stereophile archives site and you'll see these graphs for a lot of other products. The Mark Levinson ML383, for example, appears to have NO DISTORTION of ANY KIND at ANY LEVEL (at least, at the scale the graph is reproduced at ). Do these graphs give you a valid reference of how the amps will SOUND compared to each other? Damned if I know, but it's certainly intresting. Ray ------------------ Music is art Audio is engineering Quote Link to comment Share on other sites More sharing options...
Ray Garrison Posted October 29, 2001 Share Posted October 29, 2001 ...only let's me upload three at a time, here's the graph of the Wavelength 2A3 based amp... ------------------ Music is art Audio is engineering Quote Link to comment Share on other sites More sharing options...
Klipschguy Posted October 29, 2001 Author Share Posted October 29, 2001 Great info guys. By the way, while I am big fan of tube AND solid state equipment (own some of both). I'm the biggest of fan of high quality sound, performance and reliabilty, especially for the dollar. Man, nothing is worse than buying an amp, be it solid state or tubes, and being disappointed with the sound. I value your opinions here because, like me, you guys all have a passion for horn speakers and you know what you enjoy driving them with. The "best first watt" philosophy rings home with me and is the reason for this and another post (about the best solid state for the dollar). I am seriously considering the Monarchy SM-70 Pro solid state amp, and I am also VERY seriously considering building a tube amplifier. As a matter of fact, I need to start a separate thread to get your takes on a DIY tube amp. Thanks again, Andy Quote Link to comment Share on other sites More sharing options...
Kevin S Posted October 29, 2001 Share Posted October 29, 2001 Ray, Any idea how to calculate the solid state db figures into percentages to correlate to the tube figures? My recollection is that noise 75 db down (which I think is the worst case for both solid state amps, I can't really make them out very well) is a very small percentage. If I'm right, I think that all of the amps you have shown have inaudible distortion at 1 watt and below. And if the distortion is below audibility, my guess is that it doesn't matter whether it measures "hashy" or "smooth". Inaudible is inaudible I would think. ------------------ L/C/R: Klipsch Heresy II Surround: Klipsch RS-3 Subwoofers: 2 HSU-VTF-2 Pre/Pro/Tuner: McIntosh MX-132 AMP: McIntosh MC-7205 DVD: McIntosh MVP-831 CD Transport: Pioneer PD-F908 100 Disc Changer Turntable: Denon DP-72L Cassette: Nakamichi BX-1 T.V. : Mitsubishi 55905 SAT/HDTV: RCA DTC-100 Surge Protector: Monster Power HTS-5000 Quote Link to comment Share on other sites More sharing options...
Colin Posted October 31, 2001 Share Posted October 31, 2001 great stuff Ray - the curves for the tube amps look so much smoother, but the levels seem to be too miniscule to be significant, why does the chart for the Wavelength amp looks so different? Tweaking audiophiles know that the human ear has a measured listening range from 20 to 20,000 Hz at loud volumes, and that most music is generally 200 Hz to 3000 Hz. But we also know that sounds as low as 5 Hz can be sensed, adding to the realism of reproduction. Just as we know that while the harmonics of musical instruments do not extend much higher than 10 kHz, it is important to reproduce sounds far beyond that range. In fact, DVD-A format can extend the high end up to 45,000 Hz and many audiophile speakers now reach past the previous 20,000 Hz ceiling. The point is that perhaps seemingly inaudible measurements may indeed make a difference. The smooth THD curves of the tube amps might just be significant. But what I wonder about is the two-dimensional fashion in which we test audio equipment. Every graph I have ever seen of a musical note is a Grand Canyon of sharp peaks and valleys. The harmonics spike up and down like Ping-Pong balls bouncing out of control. The low 40 Hz bass note on a Fender electric guitar, for example, has audible harmonics that bounce both lower and higher - way beyond the initial string pluck. Harmonics bounce down below 14 Hz, while the 17th spike extends all the way up to 700 Hz. (Stereophile's Test CD 2, May 1992) FFT analysis shows that reproduction of a musical note is not a two dimensional function at all. Instead, it is a three dimensional one. It has amplitude, attack and decay in both frequency and time. Perhaps simple output and THD is not what we should be measuring for sound quality at all. Ralph Karsten of Atma-sphere OTL amps, for example, always includes the output section rise time in his specifications, indicating how quickly his amps respond (600V/micro-sec at http://www.atma-sphere.com/s-30.htm). I wonder how well various speaker and amp combinations can track and reproduce the hills and valleys of musical notes? Maybe we should be measuring how closely various combinations can reproduce the each musical instrument with something like the FFT analysis? The aggregate of the tests could be summed up with a correlation coefficient. It might, for example, say a certain combination was "89% for reproduction of a 40 Hz note from a piano". Isn't the exact approximation of the musical note more important than the flat measurement of output power and Total Harmonic Distortion? Where are the measurements that say how closely we come to reproducing the original? ------------------ Cornwalls & Klipsch subs; leather couch & feet up; lights out & tubes glowing! This message has been edited by Colin on 10-31-2001 at 12:24 PM Quote Link to comment Share on other sites More sharing options...
Ray Garrison Posted November 2, 2001 Share Posted November 2, 2001 Kevin, Sorry it took me a few days to respond to your question. Every once in a while (usually several times per day) someone asks a simple question about a very basic concept I think I understand, and while attempting to answer it I find out I don't really know as much as I thought I did. Relating distortion specifications measured in percent to specific dB ratios is a good example. I thought I had a pretty good understanding of the relationship, but when I sat down to explain it, I realized I wasn't really sure. I've been doing a lot of poking about, and it was harder than I expected to find any kind of definitive statement on how to convert dB to percent distortion. After much head scratching, and finding several seemlingly contradictive statements in various references, I stumbled across a formula published by Electro-Voice in the technical specifications for the EVI-12 Full Range Speaker. According to Electro-Voice, this is the formula: Percent Distortion = 100 * 10^(-difference in dB/20) or, % distortion is 100 times 10 raised to the negative value of the ( difference in dB between signal and distortion divided by 20 ). So, if we say that the third harmonic is a 80dB down from the fundamental, we get: % = 100 * 10 ^ (-80/20) or % = 100 * 10 ^ -4 or % = 100 * .0001 or .01% If there's anybody reading this who has a different mechanism for converting between percent and dB of distortion, please post it. I'd appreciate any feedback from anyone who can tell me whether the above formula is right or wrong? Thanks, Ray ------------------ Music is art Audio is engineering Quote Link to comment Share on other sites More sharing options...
Kevin S Posted November 2, 2001 Share Posted November 2, 2001 Ray, I am sorry to have put you thru so much trouble. I hope you are right, because I am not a good enough mathematician to know if you are wrong! But I did kind of remember from reading somewhere that -75db was a very low percentage of distortion. Thanks for an effort beyond the call of duty! ------------------ L/C/R: Klipsch Heresy II Surround: Klipsch RS-3 Subwoofers: 2 HSU-VTF-2 Pre/Pro/Tuner: McIntosh MX-132 AMP: McIntosh MC-7205 DVD: McIntosh MVP-831 CD Transport: Pioneer PD-F908 100 Disc Changer Turntable: Denon DP-72L Cassette: Nakamichi BX-1 T.V. : Mitsubishi 55905 SAT/HDTV: RCA DTC-100 Surge Protector: Monster Power HTS-5000 Quote Link to comment Share on other sites More sharing options...
djk Posted November 3, 2001 Share Posted November 3, 2001 Ray, your answer is correct.10%THD(1/10th the voltage) = -20dB, 3.16% = -30dB,1% = -40dB, etc. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.