Thanks, one of those transfer functions is a pretty horrible non-physical crossover region with discontinuities that have singularities in the derivatives. I don't use LM324's why should anyone?
I agree with Jon here: when building something for high end potential flaws have to be examined under microscope. Like, apply 1 MHz triangle on full swing to see close how it behaves on 20 KHz sine. If it is subtle on oscilloscope and hard to measure, it may be well audible as "something is odd". But if it looks nice, fronts of squares are equal, that means at least stereo image would not be blurred by phase intermodulation.
That's why I don't like symmetrical amps in this regard: single ended amps can't have them at all. But in order to save trees and the world from global warming I use push-pull topologies.
I saw, this adds an element of contrivance however small. In any case a transfer function this bad would hardly be expected when comparing reasonably well designed products.
DUH, John, and they EAT electricity!
I bought some cheap PCB with 4x100W class D amp, going to try for woofers and subwoofer... Have no time now though, wife is coming from Florida (conference), need to clean up in the house... Bought 4 boxes of such drivers, $1 each, need to find the place for them in my barn... Going to try tall open baffle line arrays.
An externally hosted image should be here but it was not working when we last tested it.
Hi,
Actually, they have severe crossover distortion, worse than Bipolar, in terns of static crossover (Gm related). However, it happens at higher levels than with bipolars and mosfets have much reduced problems during switchoff compared to bipolars.
But that is not the same as not having problems.
Ciao T
Actually, they have severe crossover distortion, worse than Bipolar, in terns of static crossover (Gm related). However, it happens at higher levels than with bipolars and mosfets have much reduced problems during switchoff compared to bipolars.
But that is not the same as not having problems.
Ciao T
Wavebourne said:
Been there, done that. About the only thing I proved was how line arrays can limit the vertical beam. Handy ability for a club PA. Other than that, it sounded like a bunch of 1$ speakers. Best of luck. I assume you have read Sigfired's suggestions on open baffle.
I got that z.3 amp the other day. What you don't get in a box 1/4 the size of normal is a heatsink or power supply filtering. For my intended use the former is not a problem, but I may have to do something about the filtering. Second worst of any I have. The small foorprint is a really big help. A peak under the covers, they did not scrimp on component count.
I have a bit different experience. Currently I use 8 of 4" Chinese drivers per side in wall. They don't sound like 16 cheap Chinese speakers with plastic cones. They sound very clean -- high sensitivity, low excursion, low distortions.
The same about my experiment wit PA speakers: they don't sound like 16 of 6.5" drivers with fiberglass cones, and they compared to 15" JBLs sound like studio quality sound in theater.
My conclusion is, speaker arrays is a very good way to go.
I am scratching my beard too, about power supply... It wants something about 20V, I have nothing similar. I have a bunch of 24V signal transformers, but it would mean too high loss on a voltage stabilizer.
I mean this one: 4x100W @ 4 Ohm TK2050 Class-D Audio Amplifier Board 320-302
Seems the correlation between ancedotal subjective, quality audio out of the complex reactance transducers (loudspeakers) vs. the quantifiable measurements of the amplification stages from microphone to loudspeaker is tenuous at best.
Some brillant designers weighed in on different topologies and technologies, various hidden variable, e.g. not measured using std. protocol, theories ( thermal loading of junctions etc.) discussion of G(t) or G(w) absolute phase, THD IMD local feedback loops, FET,MOSFET,Triode, opamps, but no Pentode, and various other subjects such as bandwidth, noise floor even/odd harmonics were discussed, but only a couple notes on actual possible experimentally verifiable suggestions.
I thought the dc loading of the std. THD /IMD test was a quantifiable result, and I downloaded the macscope software and ran some tests using a macbook and a , shudder, rock band usb microphome. Actually the usb mic and macscope was very interesting, from 400 hz to 3khz less than .8 % THD was my comfort limit even on single tone test.
Personally I wanted to understand why a zero measurable distortion citation 16 preamp was not my solution in front of an HK citation II. driving a modified Focal 7k/ T90/zobel pair incorporated into a Rectlinear III. A Heathkit SP2 running at low/moderate was preferred by everyone, at least at volumes below .8% THD as measured by newly downloaded Macscope being driven a usb mic in front of the slightly modded Focal 7KDBL. (~80-85 dB)
Something in the 'air", the horns, the cymbals and the "pop" in the bass and the leading and trailing edges of vocals was "distant" or muted. My gut feeling is there is a low level nonlinearity in the vacuum tube class A toplogy that "puts back" or fake some low level detail. But that doesn't explain the by EF86 phono section, I am not alone in "liking" it, and it doesn't help me tighten the LF output from the SP2. It also doesn't explain why the musicians I know were 100% favoring the CITII to a Ultra low distortion japanese (denon) 90's amp with, I believe FET class a preamp circuits. I guess they didn't know enough to hate the topology and pentodes
Decided I will take a stab at modeling the last stage of the citation one (it has a reactive feedback loop in an anode follower) in spice and see if I can accomodate the 250V B+ of the SP2 and use that as the output to the 12BY7A splitter for the CITII, that of course requires dynamic balancing and fresh tubes but still is "musical" to my ears.
Some brillant designers weighed in on different topologies and technologies, various hidden variable, e.g. not measured using std. protocol, theories ( thermal loading of junctions etc.) discussion of G(t) or G(w) absolute phase, THD IMD local feedback loops, FET,MOSFET,Triode, opamps, but no Pentode, and various other subjects such as bandwidth, noise floor even/odd harmonics were discussed, but only a couple notes on actual possible experimentally verifiable suggestions.
I thought the dc loading of the std. THD /IMD test was a quantifiable result, and I downloaded the macscope software and ran some tests using a macbook and a , shudder, rock band usb microphome. Actually the usb mic and macscope was very interesting, from 400 hz to 3khz less than .8 % THD was my comfort limit even on single tone test.
Personally I wanted to understand why a zero measurable distortion citation 16 preamp was not my solution in front of an HK citation II. driving a modified Focal 7k/ T90/zobel pair incorporated into a Rectlinear III. A Heathkit SP2 running at low/moderate was preferred by everyone, at least at volumes below .8% THD as measured by newly downloaded Macscope being driven a usb mic in front of the slightly modded Focal 7KDBL. (~80-85 dB)
Something in the 'air", the horns, the cymbals and the "pop" in the bass and the leading and trailing edges of vocals was "distant" or muted. My gut feeling is there is a low level nonlinearity in the vacuum tube class A toplogy that "puts back" or fake some low level detail. But that doesn't explain the by EF86 phono section, I am not alone in "liking" it, and it doesn't help me tighten the LF output from the SP2. It also doesn't explain why the musicians I know were 100% favoring the CITII to a Ultra low distortion japanese (denon) 90's amp with, I believe FET class a preamp circuits. I guess they didn't know enough to hate the topology and pentodes
Decided I will take a stab at modeling the last stage of the citation one (it has a reactive feedback loop in an anode follower) in spice and see if I can accomodate the 250V B+ of the SP2 and use that as the output to the 12BY7A splitter for the CITII, that of course requires dynamic balancing and fresh tubes but still is "musical" to my ears.
Pano,
I don't know.
Has any evidence been presented that "measures good" in the traditional sense correlates reliably with "good sound" as perceived?
I have not noticed it. I think until it is (or we get a formal admission of the contrary) the thread remains relevant and open.
Indirectly many of the seemingly "off topic" posts address this, even all the way to the differences in educational systems*.
Transistors may or may not have nasty crossover artefacts. They may or may not accentuated by looped feedback and other interactions and produce a system that measures well statically, but sounds bad...
Ciao T
*FWIW, I spend the first 3 month of my 3 years degree level "polytech" electronics course practical modules soldering hedgehogs (veroboard with wires, wires had to be precise length, correctly un-isolated and correctly soldered) and making a key out of a solid piece of steel using dull metal saws, dull files and generally bad tools (most actually the throwaways of people who actually did real work) and either module was an absolute pass requirement, including the fact that the key had to unlock the door to the johns....
I cannot remember how many times I have since given thanks for these 3 Month (and the 6 month when I did nothing but calculate transformers in any conceivable way and some never before and since conceived), definitely more often than I did curse the mindless and stupid way of teaching us while I was filing a key and soldering Hedhogs...
Such "getting your hands dirty" education as was common in the eastern block does give one a totally different angle on many things compared to the college/uni boys turned out by the western systems.
I don't know.
Has any evidence been presented that "measures good" in the traditional sense correlates reliably with "good sound" as perceived?
I have not noticed it. I think until it is (or we get a formal admission of the contrary) the thread remains relevant and open.
Indirectly many of the seemingly "off topic" posts address this, even all the way to the differences in educational systems*.
Transistors may or may not have nasty crossover artefacts. They may or may not accentuated by looped feedback and other interactions and produce a system that measures well statically, but sounds bad...
Ciao T
*FWIW, I spend the first 3 month of my 3 years degree level "polytech" electronics course practical modules soldering hedgehogs (veroboard with wires, wires had to be precise length, correctly un-isolated and correctly soldered) and making a key out of a solid piece of steel using dull metal saws, dull files and generally bad tools (most actually the throwaways of people who actually did real work) and either module was an absolute pass requirement, including the fact that the key had to unlock the door to the johns....
I cannot remember how many times I have since given thanks for these 3 Month (and the 6 month when I did nothing but calculate transformers in any conceivable way and some never before and since conceived), definitely more often than I did curse the mindless and stupid way of teaching us while I was filing a key and soldering Hedhogs...
Such "getting your hands dirty" education as was common in the eastern block does give one a totally different angle on many things compared to the college/uni boys turned out by the western systems.
Hi,
Find a 24V/20A SMPS on E-Pay, modify the sense to produce 20V. Easy.
Given you are using class D amplifiers any possible pretense to quality is lost anyway...
Ciao T
Find a 24V/20A SMPS on E-Pay, modify the sense to produce 20V. Easy.
Given you are using class D amplifiers any possible pretense to quality is lost anyway...
Ciao T
You're prejudices are showing. Showing badly.
Still don't see what this has to do with measuring sound quality.
- Status
- Not open for further replies.