If the amplifier does not follow them, they become longer in duration.Vinyl ticks and pops deliver the highest slew rate figures, considerably beyond
program material. I am not at all certain that I want to bother reproducing
them with perfect fidelity.
You and I discussed it once, and agreed that 10X peak program material was
probably an adequate margin - that puts us at something like 15 V/uS for a
100 watt amplifier.
Could that make them more audible?
The harmonic phase structure of classical musical instruments is well documented. Not so for electronic music. So yes you can crank square waves through an audio amplifier. Even if you don't blow the tweeters I don't want to listen to it.
However as recorded music is band limited that also puts an upper bound on slew rate.
My largest tweeter array had to do 90 dB, 8,000 Hz at 800' using multiple 112 dB per watt horns. It was driven by Crown MA5000 amplifiers and the amplifier limits did come into play.
The issue being raised is about the difference in using CFAs and if it is due to slew rate.
My suspicion is that it is a matter of wider bandwidth and some inherent distortion compensation mechanisms. Not slew rate as in domestic use there are multiple techniques to achieve more than adequate performance. Most of the other folks doing large scale sound systems only aim for 8,000 Hz high frequency response. I am able to get 12,000 Hz by a few tweaks which are part of the secret sauce used here.
When I measured the basic non complementary differential input to single transistor voltage gain stage to darlington power output stage, the results showed significantly lower distortion from the non-inverting input. The folks circa 1970-80 who used the inverting input mostly went away. It could well be that there are better topologies now used and the difference is moot.
Rubbish! If I can do it, up to a point, imagine what all of us could do. I did it for fun, as a small experiment just to satisfy my curiosity, but 1 will get 10 if somebody like Damian Martin, John Curl, Brad Carson, not to forget Nelson Pass, sat down hard and wotked on it in an organized way, that would be something altogether different.
Will we now get CFA lovers start to fear for weakened foothold? Yet this would be hardly any revolutionary news, much of what we use daily now resulted in such criss-crossing of methods. What I did is something of a minor miracle in view of the fact that I know next to nothing about CFA topologies. A fluke. My ONE AND ONLY solid knowledge is that I listened to Andrew Russel's attempt and found it sounds very good to me.
I have design of a "CFA" with the high impedance - input, so it could be said it is CFA VFA combination. It behaves exactly as a CFA but with no drawback of the global feedback network eating to much power and needing high power resistors.
The link to the thread where it is: http://www.diyaudio.com/forums/solid-state/253039-unique-cfa-120-230w-amp-16.html#post4428619
Damir
From my vinyl observations, with MM cartridges (2.5-5mV @5cm/s), strong-sharp scratches produce signals with a slope of 0.3-1mV/ns at the cartridge terminals (unamplified-unequalized signal).
Ticks and pops due to dirt may produce stronger output but the signal is not that abrupt as from a scratch.
The worst -in terms of slope- signals I’ve observed, come from static discharging (up to 2.5mV/ns at the cartridge terminals)
George
This is the 'H' bridge amplifier it behaves like a CFA but is considered to be a VFA by the semi industry - I decided not to argue with the smart guys on that one.
You can read my little write-up here CFA vs. VFA: A short primer for the uninitiated – Extensively Updated January 2014
You can read my little write-up here CFA vs. VFA: A short primer for the uninitiated – Extensively Updated January 2014
Last edited:
No, it has nothing to do with MM versus MC, the programme content itself lacks high slew rate, as does typical digital content.
The paper you posted recently seems to lack this measurement for programme material. It's easy enough to verify............
What actually matters for power amps is current slew rate capability; run some sims for current demand using programme material as a piecewise source into realistic model loads.........it's the load which presents the issue not the programme material.
The vinyl defects don't see pre-emphasis, some of the worst stuff I've seen had synthesized chimes added to an already hot film soundtrack. A real torture test for a pre-amp. I also seem to find Italian engineers mix a little hot on the highs.
Damir, the low impedance feedback is imposed by the parasitic capacitance of the emiter of the input transistor (feedback target). You are obliged to keep-it low, if you want no phase turn in the feedback signal at high frequencies.
If you do the same for a VFA, sacrifying the balance of impedances between the two inputs, you will increase the bandwidth very near the one of the same king of amp in CFA mode.
An understanding of the CFA operation is useful to see its niche in the over-all scheme of things and to blend that knowledge with VFA knowledge for the best of all. --- which seems some are doing already.
As you said.... more tools.
The low z circuitry of CFA means operate with currents rather than voltage which give us some benefit in both higher SR but also lower distortion from device and stray C's. A lower operating/signal voltage swing across these non-linear c's means lower distortion.... esp at higher freqs.
Esperado's point is well taken.
THx-RNMarsh
Last edited:
Sounds like something to put a bug in the ear of Bruce Hofer about
Best to let the sleeping dogs lie.
Sounds like something to put a bug in the ear of Bruce Hofer about
If you consider the square wave case sin(x) + 1/3 sin(3x) ... As the worst case then you only look at odd harmonics and for most recorded actual music material then you need only analyze 6667 Hz. As slew rate increases linearly with frequency.
HF transducers without zobels are mostly inductive. Slew rate is measured with resistive loads. There are some folks who do have complex passive networks ahead of the transducer but in practice removing the passive network and doing the processing in DSP is far more effective.
I'm only half-kidding about Audio Precision. There are a number of professionals who have reservations about even the top-of-the-line 555 machine, and recently I've had to familiarize myself with its capabilities.
I considered teasing Hofer a bit (I've only met him once, in LA some years ago) about the analyzer bandwidth in the new instruments. As switchmode amp aficionados may know, the older instruments had difficulties dealing with significant amounts of out-of-audio-band energy, the AES brickwall filter options notwithstanding --- it was a front end issue. In the oldest instruments there was only so much the ~5534s could manage, and the errors were unpredictable. So Ap came up with a nice low-distortion passive filter box, developed by Duke Aguiar. I may have mentioned it when we were gnashing our teeth about potential distortion from ferromagnetic materials.
But the new analyzers don't require that filter anymore. Except, with the fairly recent Levinson amps, the relatively-low out-of-band energy extends to at least 4MHz, according to what I've heard. Back to the drawing board? Or just a kinder gentler higher-frequency filter box?
Of course this has almost nothing to do with slew rate.
I considered teasing Hofer a bit (I've only met him once, in LA some years ago) about the analyzer bandwidth in the new instruments. As switchmode amp aficionados may know, the older instruments had difficulties dealing with significant amounts of out-of-audio-band energy, the AES brickwall filter options notwithstanding --- it was a front end issue. In the oldest instruments there was only so much the ~5534s could manage, and the errors were unpredictable. So Ap came up with a nice low-distortion passive filter box, developed by Duke Aguiar. I may have mentioned it when we were gnashing our teeth about potential distortion from ferromagnetic materials.
But the new analyzers don't require that filter anymore. Except, with the fairly recent Levinson amps, the relatively-low out-of-band energy extends to at least 4MHz, according to what I've heard. Back to the drawing board? Or just a kinder gentler higher-frequency filter box?
Of course this has almost nothing to do with slew rate.
- Status
- Not open for further replies.