Not to go OFF TOPIC here or hijack the thread...
Should someone new to whole idea of DIY Tube amps learn a little about SS designs?
The information contained in this thread and that which is certainly to come, has led me to believe that maybe I SHOULD take a look at SS.
Maybe my previous experiences with SS have "tainted" my opinion to something like "VT's can do no wrong"?
In all honesty the basic "simplicity" of VT designs has made it easier for me to grab hold of the concepts. Maybe the more technical SS circuits might give me a new insight?
Should someone new to whole idea of DIY Tube amps learn a little about SS designs?
The information contained in this thread and that which is certainly to come, has led me to believe that maybe I SHOULD take a look at SS.
Maybe my previous experiences with SS have "tainted" my opinion to something like "VT's can do no wrong"?
In all honesty the basic "simplicity" of VT designs has made it easier for me to grab hold of the concepts. Maybe the more technical SS circuits might give me a new insight?
I believe that no technology is worse or better than the other. Ignoring solid-state circuitry is like ignoring capacitors, resistors (or some other important component) just because you have a prejudice that “they’re bad”. Both tubes and transistors are components that can provide tons of possibilities and learning to understand and exploit them can only be beneficial. Learning SS circuitry will strengthen your understanding of tube circuits (especially those of hybrid kind and those that are not so "classic" designs) and vice versa. Basically it will widen your understanding of electronics in general. I don’t understand why anyone should limit himself or herself to a specific scope on this field. Just have an open mind and ignore the common myths about solid-state inferiority. The SS amps/circuits are only as good or bad as someone designs them to be.
Hi hkibbe. I can think of no better speed intro on tube physics than http://www.john-a-harper.com/tubes201/ . Guessing from the fact you're a student is it safe to assume your guitar amp will be driven into overload regularly? If so, SY's comment is the place to start. The input impedance of a common used guitar tube snaps from near-infinity to something approaching a resistor in series with a diode to cathode when the grid voltage exceeds the cathode voltage. The 'sound of tubes' in this application is the sound of an unhappy driver, excess plate current sagging the power supply, and a host of other conditions in which the hi-fi perspective has little relevancy. As others have hinted, your interest may lie more towards the sound of tubes abused.
I think a good case can be made that there is a tube sound. There didn't have to be, but people seem to have built similar circuits over and over again. Remember that most engineers start by surveying the literature, including schematics of their competitors. Economics also drives designs to have more in common than one might think. Same thing in the solid state world, and I can safely say a fair number have not been shining examples of good sonic practice. I'm not saying every amp in its respective family sounds the same, but I think there are some commonalities. Regarding a post above, didn't McIntosh build solid state amps with output transformers?
teemuk said:
It is like to compare 2 roses sniffing through a smelly rubber gas mask.
smoking-amp said:
Look at the loadlines. I haven't been able to find one that'll get the estimated THD below 6% with the 12AV7. You can get better THD estimates from the 12AU7. Another thing I haven't found is anyone saying that 12AV7s don't sound pretty bad. Mostly, it's described as very "dark" -- possibly too much h2.
Neither type seems to have been intended for audio use. Almost all the 12AU7 circuits seem to be RF or semi-digital. Not much use for 12AV7s since a 12AT7 would be better everywhere you'd consider a 12AV7.
Johan; loudspeaker stuff ...
There is loads of info on this.
http://www.peavey.com/support/technotes/hartley/chapter_2.pdf
http://mixonline.com/TECnology-Hall-of-Fame/1925-dynamic-loudspeaker/
keep listening
richy
There is loads of info on this.
http://www.peavey.com/support/technotes/hartley/chapter_2.pdf
http://mixonline.com/TECnology-Hall-of-Fame/1925-dynamic-loudspeaker/
keep listening
richy
Gordy,
Although I have a copy of this paper by GEC valve vs Transistor preamble, I have a suspicion this docu has been re-copyrighted so I would do wrong by submitting it.
Link given.
www.scribd.com/doc/6722472/AudioXPress-September-2
richy
Although I have a copy of this paper by GEC valve vs Transistor preamble, I have a suspicion this docu has been re-copyrighted so I would do wrong by submitting it.
Link given.
www.scribd.com/doc/6722472/AudioXPress-September-2
richy
This is the one generality that I would agree with for the most part. There is a fairly wide range in "tube sound", but that is the main commonality. Don't pay mind to the Pentode vs. triode BS that some people spew. Even though the curves look like those of transistors, Pentode amplifiers usually sound much closer to a triode than a solid state amp. It is mostly because of design technique. I've heard some really impressive pentode amplifiers.
Pair play. I've been in tubes for a long time but appearing from the solid state chaff during the early 1980's one couldn't ignore the very first Hitachi 100W MOSfet amp that came on the market.Jeb-D. said:
That was a big sound step (away from the then poorly designed bipolars). It had the simplest diff class A front end with complementary power types and good sounding. Interestingly the thd pattern was sim to a tube amp, but the sig/noise wasn't as good but it could also oscillate into a rough load, so can a badly designed tube amp.. So we have it both ways. As per other comments, these types of mosfet amps required massive global nfb to linearise the performance.
I agree with Miles 100%. When comparing tubes amps, hightime to separate musical inst guitar and HiFi amps into different catagories. The design is different although I've often used HiFi amps as bass bin amps. Here there is a better fit, as thd is lower and not so noticeable onset near clipping. A solid state amp forced into clipping on the bass end sounds repulsive.
Regarding the bass end, a tube amp working a large bass reflex cabinet tuned to a low frequency really profits from looser and lower global nfb gain relationship of the tube amp versus the Q factor of the loudspeaker cabinet. A high Q LS unit ported/vented cabinet can accentuate the bass whereas a small one will lack depth. Your choice. One gets used to one and not the other.
However there is more to this than meets the ear. The real problem is that no two loudspeakers and no two amplifiers sound the same be it with phase correlations or whatever. There is always something up. Invariably, the subject of loudspeaker science in my mind is never far away and is still relatively fresh.
I'm used to the sand filled panels.
richy
Perhaps some more remarks:
One must be careful here. Triode tubes have visibly curved characteristics, but semiconductors' graphs resemble those of pentodes. There are no devices similar to triodes in SS.
But those curves do not imply non-linearity. Without over-doing theory, amplification is achieved by analysing distances between where the 'curves' cut a load line. Simplifying, distortion is inherent to the distances between Ia/Va (or Ib/Ic) lines, not the curvature of each. (They are also curved in the case of pentodes/SS, just difficult to notice on a small scale.)
The next point is conditional. It is not so much that an enormous amount of NFB is essential to make SS amplifiers acceptable, it is more "man's vanity" .... SS amplifiers' transfer function can be 'shaped' to have less phase shift than that of the general tube amplifier - thus designers add NFB because they can! A figure of 0,0001% distortion looks better on paper than 0,001%, so it is done. Never mind whether there is any difference in audible performance (high order distortion is often worse).
Yes, usually more NFB is required to get them on par, but as an example (just for the record): My 80W SS amplifier does some 0,24% of thd before any global NFB, with some nested feedback here and there. Final figure is 0,008% with only 28dB of global NFB; not that much by today's standards. (The Mullard 5-20 tube amplifier has 30dB of NFB.) The basis for the umpteenth time: Start with a proper design before any NFB and then add your NFB.
It has been said; just to emphasize that tube sound mostly comes in for guitar artists, for the reasons offered by others. In hi-fi amplifiers there is no excuse for an audible difference, other matters being equal (I have designed several of both topologies).
coldcathode said:
One must be careful here. Triode tubes have visibly curved characteristics, but semiconductors' graphs resemble those of pentodes. There are no devices similar to triodes in SS.
But those curves do not imply non-linearity. Without over-doing theory, amplification is achieved by analysing distances between where the 'curves' cut a load line. Simplifying, distortion is inherent to the distances between Ia/Va (or Ib/Ic) lines, not the curvature of each. (They are also curved in the case of pentodes/SS, just difficult to notice on a small scale.)
The next point is conditional. It is not so much that an enormous amount of NFB is essential to make SS amplifiers acceptable, it is more "man's vanity" .... SS amplifiers' transfer function can be 'shaped' to have less phase shift than that of the general tube amplifier - thus designers add NFB because they can! A figure of 0,0001% distortion looks better on paper than 0,001%, so it is done. Never mind whether there is any difference in audible performance (high order distortion is often worse).
Yes, usually more NFB is required to get them on par, but as an example (just for the record): My 80W SS amplifier does some 0,24% of thd before any global NFB, with some nested feedback here and there. Final figure is 0,008% with only 28dB of global NFB; not that much by today's standards. (The Mullard 5-20 tube amplifier has 30dB of NFB.) The basis for the umpteenth time: Start with a proper design before any NFB and then add your NFB.
It has been said; just to emphasize that tube sound mostly comes in for guitar artists, for the reasons offered by others. In hi-fi amplifiers there is no excuse for an audible difference, other matters being equal (I have designed several of both topologies).
Why do you even build with tubes if that is how you think? Transistors should provide the same results, but cheaper and more efficiently.
By today's solid-state standards sure. But that is a crapload of feedback by today's tube amp standards.
Many amplifier designers forget that loudspeakers are far from linear. With that being said, is the best tool for driving them a perfectly linear amplifier?
Hi Jeb-D,
I agree, but I do it because some people want/prefer it. (I'm not on the open market, but make per request.) I come from the era
, perhaps some reckon I should be able to do a good job.
I reacted to the general idea that lots of NFB is needed for SS amplifiers.
But what is norm in 'today's standards'? With all the contemporary urban legends about what is supposed to be 'right' (whatever that might mean), I would rather take my lead from classic amplifiers: Williamson used 20dB; Leak TL12: 26dB; Mullard 5-20: 30 dB; McIntosh: >20dB (multiple loops, not sure of total amount) and so forth. Those figures were indicative of what was required in good designs to make distortion inaudible, and were applied with the necessary stability margins without eroding open loop gain to well within the audible range like many SS designs etc. etc., - so! I would not dare to quote any figure as a standard; it depends on the design (if not nowadays on taste!).
I suppose not, but then you are venturing into the field where amplifier characteristics must be the conjugate of the loudspeaker for any improvement on the present situation - how will that be achieved without the same manufacturer making both? What is the amplifier designer supposed to do about it, except for ensuring that his amplifier stays stable? User-adjustable controls in the hands of laymen for 'tuning' to loudspeakers would create chaos. The most practical solution seems to be the present, although you would know that e.g. motional feedback is used - but again, only in 'committed' amplifier/loudspeaker designs ... no room for choosing either independently.
Jeb-D. said:
I agree, but I do it because some people want/prefer it. (I'm not on the open market, but make per request.) I come from the era
, perhaps some reckon I should be able to do a good job. (referring to 28dB of NFB)
I reacted to the general idea that lots of NFB is needed for SS amplifiers.
But what is norm in 'today's standards'? With all the contemporary urban legends about what is supposed to be 'right' (whatever that might mean), I would rather take my lead from classic amplifiers: Williamson used 20dB; Leak TL12: 26dB; Mullard 5-20: 30 dB; McIntosh: >20dB (multiple loops, not sure of total amount) and so forth. Those figures were indicative of what was required in good designs to make distortion inaudible, and were applied with the necessary stability margins without eroding open loop gain to well within the audible range like many SS designs etc. etc., - so! I would not dare to quote any figure as a standard; it depends on the design (if not nowadays on taste!).
I suppose not, but then you are venturing into the field where amplifier characteristics must be the conjugate of the loudspeaker for any improvement on the present situation - how will that be achieved without the same manufacturer making both? What is the amplifier designer supposed to do about it, except for ensuring that his amplifier stays stable? User-adjustable controls in the hands of laymen for 'tuning' to loudspeakers would create chaos. The most practical solution seems to be the present, although you would know that e.g. motional feedback is used - but again, only in 'committed' amplifier/loudspeaker designs ... no room for choosing either independently.
Johan Potgieter said:
I question that assumption. Back in those days, speaker design was very much a "black art". The early designs may have required quite a bit of damping to tame squirrelly impedance v. frequency characteristics. Since the advent of the "Thiel Method" of speaker design, this is no longer the case. Modern speeks don't seem to have that need. I find that too much gNFB makes for a "solid statey" sound. When I tried ~20db(v) of gNFB the result was truly horrendous.
One of the biggest hoaxes in the studio industry is an easy way to manipulate the low end fundamental bass and push it to the 2nd harmonic to make the bass sound "richer" in small box loudspeakers. So when a big floor standing unit comes along the low end is overpowering. I have my doubts on original sound quality when console engineers can manipulate the job in such a way to improving artist sales.
When comparing loudspeakers, todays pop music is the worst medium one can use. Try instrumental Bluegrass banjo or classical. The sharp transients will show up any mid and top end problems.
Johan mentions; quote <Those figures were indicative of what was required in good designs to make distortion inaudible, and were applied with the necessary stability margins without eroding open loop gain to well within the audible range>
That's confusing and has to be re-phrased. Otherwise the math doesn't add up. Applying global nfb, the open loop gain must drop the overall amp gain to the value of feedback applied. Too much and one has an oscillator. It's getting late or perhaps of language I've missed the point.
richy
When comparing loudspeakers, todays pop music is the worst medium one can use. Try instrumental Bluegrass banjo or classical. The sharp transients will show up any mid and top end problems.
Johan mentions; quote <Those figures were indicative of what was required in good designs to make distortion inaudible, and were applied with the necessary stability margins without eroding open loop gain to well within the audible range>
That's confusing and has to be re-phrased. Otherwise the math doesn't add up. Applying global nfb, the open loop gain must drop the overall amp gain to the value of feedback applied. Too much and one has an oscillator. It's getting late or perhaps of language I've missed the point.
richy
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.