Columns Retired Columns & Blogs |
May 5, 2009 - 7:30pm
#1
SET Linear and Nonlinear Distortion Revisited
Loudspeakers Amplification | Digital Sources Analog Sources Featured | Accessories Music |
Columns Retired Columns & Blogs |
Loudspeakers Amplification Digital Sources | Analog Sources Accessories Featured | Music Columns Retired Columns | Show Reports | Features Latest News Community | Resources Subscriptions |
That would, I believe, be more consistently true than jj's remark.
I guess most tube amps have a high source impedance. However, the other thread was really concerned with distortion characteristics of SET amps and their perceived affects on the sound. In fact, you have actually raised an interesting new topic, and one not really confined to SETs.
There are some few speakers with a fairly flat impedance vs. frequency curve, but most of them do not. A high source impedance amp will not have a very flat response into most speakers.
And so the user of a high source impedance tube amp or any other high source impedance amplifier would be wise to use proper speakers with their amplifier. This was pointed out in the other thread.
What constitutes proper speakers for an amplifier?
For a high source impedance output - transformer coupled or OTL - I think most people would agree a high impedance load of 8-16 Ohms would be preferred. The load should be as consistent as possible with only minor alterations due to frequency and with minimal capacitance or inductance other that which comes from the driver(s) itself or a minimal crossover/filter network. Beneficial to the operation of the amplifier would be a speaker system with high electrical sensitivity and high overall system efficiency. Full range, single driver dynamic systems fit the bill or such drivers combined with a high frequency driver of consistent values.
If you'll look at any good SET webpage, you should find some links to speakers that suit such amplifiers. IMO it isn't a bad idea to use such speakers systems with most any amplifier. Why insist an amplifier must be capable of arc welding to sound good with some bizarre load of a speaker?
On the other hand, low sensitivity, low efficiency does not equate to unreasonable in certain cases;
http://www.symphonysound.com/articles/tubefriendly.html
The ones that someone thinks sound best with it?
Well, there are more speakers available than anyone could try out. How is one supposed to find out which ones will sound best? Let alone sound best using one's own tube amp? There has to be some way to cut down the numbers.
Yes, if one uses such amps, one needs to spend some time trying to find speakers the amplifier can drive and sound good with.
"IMO it isn't a bad idea to use such speakers systems with most any amplifier."
This presupposes one likes those tube friendly speakers. The little Rogers LS 3/5a would hardly meet my needs, and I have heard few high sensitivity speakers I like well enough to consider as main speakers. Then there is the problem of actually determining whether a speaker works well with tubes or not. It seems like an awful lot of extra trouble to me.
However, most speakers do not have a flat impedance curve, the minimum impedance of most speakers seems to be well below 8 ohms, and are not terribly sensitive, averaging may 87 dB or so.
"Why insist an amplifier must be capable of arc welding to sound good with some bizarre load of a speaker?"
Why not? Such an amplifier is capable of driving the vast majority of speakers. The performance should remain stable for a long time and there is no need to worry about changing tubes.
As well, an amplifier with a low output impedance will keep a very flat response into many different speaker loads. Maybe some speakers will sound better with some fixed EQ from the amplifier, but it seems to me that most speakers are designed with a low source impedance in mind.
To Whoever,
I found this quote from the RCA Radiotron Designers Handbook, written by 26 engineers, page 881 quite interesting.
"III) Procedure for "matching" loudspeakers to various types of amplifiers."
"When the load resistance is constanct, the only procedure necessary is to select a transformer ratio so that the resistance reflected into the primary is the correct load for the amplifier. When the load is a loudspeaker, the procedure is outlined below."
"CLASS A Triodes, either singly or in push-pull, may be treated very simply by arranging for the loudspeaker impedance at 400 c/s to equal the correct load for the amplifier. Thanks to the shape of the triode characteristics, the rise of impedance at the bass resonant frequnecy decreases the distortion, and although the power from the valve is lower than at 400 c/s this is counterbalanced by the rise of loudspeaker efficiency at this frequency. The rise of impedance above 400c/s results in a tendency towards a falling response, but loudspeakers specially designed for use with triodes are capable of giving fairly uniform response up to their limiting frequency. As a result, the designer of an amplifier with a triode output stage need not consider the impedance versus frequency characteristics of the loudspeaker, but only the response."
"The ratio of the nominal load resistance RL to the output resistance Ro of the amplifier is not unimportant. If RL/Ro is very high"
(low output impedance of amplifier/high feedback)
"the loudspeaker is being operated with nearly constant voltage at all frequencies. If RL/Ro is around 2-3"
(damping factor/higher output impedance),
"the voltage applied to the loudspeaker is slightly greater at frequencies where the loudspeaker impedance rises--this is generally an advantage."
I thought this was quite interesting.
On another note, SET amps, depending upon the design usually has most of its harmonic distortion concentrated in the lower harmonics, which is much more benign. However, the above comment has limitations depending upon the distortion characteristics of the previous stages. The total distortion at 1 watt (for 100db spl) in some SET amps is claimed to be approximately .01%.
Hope this helps.
What's that supposed to mean?
Can I "presuppose" you wouldn't buy any speaker you didn't like no matter the amplifier?
No one suggested you had to buy the 3/5a's. They are used as an example of a wonderfully "amplifier friendly" speaker that happens to also work very well on tubes since it makes no great demands of any amplifier other than available voltage. If you find another speaker with similar electrical characteristics - nominally high impedance with non-demanding electrical phase, then you can buy that speaker. How much of a curmudgeon are you going to be?
I would suggest you investigate the high sensitivty speakers being built today. You sound like you equate high sensitivity with Cerwin Veags. That is not what's being built now and, as an example, if you listen to pair of Lowthers, you'll find a 70 year old driver that sounds nothing like a CV. Most dealers don't stock such systems because they have a difficult time selling the music rather than the googaws that make up the system. Find a good dealer and listen. If you don't like what you hear, that's fine. But, if you haven't listened to a very good high efficiency system, then you shouldn't make such comments.
So? What makes that right? I would say that has become a major problem with too many of the speakers being sold today.
You didn't ask me what "most speakers" are, you asked what would work well with a high source impedance amplifier. I told you. I would think most anyone serious about SET's or OTL's would investigate what will mate synergystically with their amplifier. It might be a bit of trouble to find but then I don't expect to find a SET or my next speaker at Best Buy. How much effort did you put into your last speaker acquisition?
You are essentially arguing against tubes. If you don't prefer tubes, that's fine but some of us do.
Why not have an amplfiier capable of arc welding? Cost to begin with. Then complexity. You are unlikely to find a single pair of output transistors that can handle a 1.5 Ohm load with a 45 degree phase angle. So a huge power supply costs money while multiples of output devices tend, IMO, to sound less like music than a simple ciruit. Did you buy your last amplifier because it could drive any speaker or because it sounded better than anything else you could afford? Or because it measured OK?
You are again showing your bias against tubes or any amplifer with a autoformer/transformer coupled output. If you want to argue bias, then you are on your own. Why would I want to get on that ride?
I just told you what sort of speaker would not have a "fixed EQ" with such an amplifier because I thought you seriously wanted an answer. Now I see you just wanted an excuse to say unkind things about tubes, 3/5a's, high sentivity speakers, etc. or anything I suggested. Really, Xeno, if you want to insult tube amplifiers and all the rest, then you probably should go to another thread. I give you answers and you say they aren't what you want. Fine, don't buy any. No one is insisting you must. But, 1) you don't sound like you have much experience in this area to be criticizing what you don't know, 2) you are simply arguing your own biases and not making a rational argument for or against SET's and high efficiency speaker systems and, 3) I gave you the answers you asked for. As jj would say, you are "picking an argument".
So? What makes that right? I would say that has become a major problem with too many of the speakers being sold today.
As best as I can figure from all that muck, this is what your response comes down to ...
I didn't.
And as far as I'm concerned that is the end of this "discussion".
Yeah, you've written yourself into a whole, so I don't blame you.
You have fun with your tubes.
Jan, given how commited you seem to be to your vacuum tube ideology, I don't know how much sound quality actually matters to you.
Back in the day, and I still think today, there are vacuum tube amps that have respectible source impedances and don't produce a signficantly different-sounding audio signal at their output terminals depending on which speaker you hook them up to.
I'm not really up on the latest-greatest in the land of hollow state, but don't people like Manley, Conrad-Johnson, and McIntosh still sell vacuum tube amps whose fidelity to their input signals is actually pretty good with most speakers?
My recollections of HE2005 include that I'd walk into a room, see the glowing bottles and generally hear some pretty strange stuff. I remember walking into this one room, seeing the glowing bottles and my ears were not assaulted. The music was just alright! I looked around and saw that it was the Manley room. Go figure.
I don't know exactly what you are trying to say here other than, "Jan, given how commited you seem to be to your vacuum tube ideology, I don't know how much sound quality actually matters to you", to which I say, "Screw you!".
Once again I will repeat, I didn't ask.
I realize this is disrespectfully off topic and way too provocative for this thread, but....
Back in '75 my friend Ron Hamilton built a Mike Moffat designed SET good for about 18 watts per. Fed it signal from a modded HK Citation IV and to a pair of nicely tuned QUAD 57's.
It sounded effin' bitchin!
This is irrelevant to loudness growth.
And Buddha gets the 6 runs
This, in a nutshell, is the answer.
It is interesting how any number of obvious issues are being brought up here to argue around (it's hard to argue against the facts, so all you can do is dance around the side issues) the issue of distortion and loudness growth.
Output impedence is certainly a WELL KNOWN issue. Why anyone considers this relevant to the the distortion issue is unclear, unless a speaker requires a very low impedence to remain linear, in which case we have another problem.