In my view it would be the non-linear loading of the high Vas node impedance by the non-linear load attached to it, NOT specifically 'running out of current'.
If you shunt the Vas node with a resistor, any non-linear load current will have less detrimental effect on the Vas voltage because the non-linear load current 'works' against a lower impedance.
Jan
OK then we are now searching for by what means digital cable sound different to wit how are the signal altered in a way that we hear the change ?. We are in agreement on the analog cables. As for connector I have gone to rodium rather than gold less layers and am happy with the sound for me. Soldering of rodium is an adventure thou .
But now you need more input to make the same Vas voltage. You have to look at the complete signal chain in situ. This is all moot in any case, at say 10kHz, the non-linear displacement currents will swamp the base current especially considering the ft's of modern devices.
The complication is that the output stage may be quite linear in voltage terms, but very non-linear in current terms. Hence if driven from a current source (actually a transconductance, if you include the preceding stages too) it looks like heavy feedback is being used to fix up a very non-linear forward path. At HF the Miller cap will help by reducing VAS output impedance.scott wurcer said:
I think my problem is that I am used to thinking of circuit non-linearity as being just a mild perturbation on a linear circuit, but output stages driven from a high impedance source are very non-linear (and frequency-dependent) so should we be using a Volterra series to describe them? Can we just do Thevenin swaps or do these only apply to linear circuits (this is a genuine question).
The error current is supplied by the input stage in either case and the error voltage at the amplifier's input caused is the same in both cases. Except now you have added the need to drive the current in the shunt R and if the input voltage to current conversion has any non-linearity you have actually added distortion.
In a special case at some frequency could there be a fortuitous cancellation? Maybe. This can be observed on some data sheets where in a distortion vs frequency plot there is a suckout, this is usually just a curiosity and not useful.
A transient analysis will yield the correct answers for distortion if run long enough, the Volterra or harmonic balance analysis is just another way with the advantage that there is no initial transient to die out.
Let's take a look:
Let us suppose that GmR is the input stage gain. This includes the VAS. Now IF R(load) is nonlinear then the overall transfer function will be non-linear. Agreed?
Now what if we add R(linear) to the total?
Then the gain is Gm [R(L)paralleled with R(linear) and the gain will be lower, but won't it be more linear?
Let us suppose that GmR is the input stage gain. This includes the VAS. Now IF R(load) is nonlinear then the overall transfer function will be non-linear. Agreed?
Now what if we add R(linear) to the total?
Then the gain is Gm [R(L)paralleled with R(linear) and the gain will be lower, but won't it be more linear?
If of interest the Thorens TD 145 did marginally improve with added mass ( DL110 ). It might also be damping arm tube resonance using the Blu Tak weight. The bass was noticeably greater, cutting rumble more intrusive. I can live with that. Overall the Thorens is disappointing. A JVC direct drive of the more domestic type seems much better.
I put the Blu Tak just behind the head-shell joint. The counterweight taken back about to 40% of it's travel ( was 0% ) . Weight reset to 2g.
Top Tip. I remember from Linn days cleaning the belt with Pledge furniture polish. It did make a difference to flutter. The belt felt rough before that and the kitchen towel very black.
I can not describe how bad the awful RCA to DIN adapter was and how OK the replacement of equally awful looks was . A timely eduction in common sense. Cable mostly don't matter and connectors really do. Now with high grade DIN fitted.
I put the Blu Tak just behind the head-shell joint. The counterweight taken back about to 40% of it's travel ( was 0% ) . Weight reset to 2g.
Top Tip. I remember from Linn days cleaning the belt with Pledge furniture polish. It did make a difference to flutter. The belt felt rough before that and the kitchen towel very black.
I can not describe how bad the awful RCA to DIN adapter was and how OK the replacement of equally awful looks was . A timely eduction in common sense. Cable mostly don't matter and connectors really do. Now with high grade DIN fitted.
Distortion gets higher with VAS load. Only in very specific case the two nonlinearities may tend to cancel to some extent, I have experienced it with just one circuit (not this one).
Attachments
This is the crux of the problem 1mV of curvature on a 10mV signal is 10% non-linearity more or less. The same 1mv on a 1V signal is 0.1% non-linearity 100X "better" but no net gain. The resistor is just adding more (nearly perfect) current demand from the input stage, when enclosed by the feedback this is just a reduction in net gain.
EDIT - As PMA shows there could be a slight degradation in distortion because the input stage is never perfectly linear.
Last edited:
My olde Dual CS 604 has an even ligher tonearm and cartridge, but at the recommended pressure of 1.2 gr I really feel no bass loss. It's as deep as recoerded, as authopritative as any froman LP if present.
I must add that the Ortofon recommended capacitance of 400 pF is met almost exactly (off by +10 pF) and 47 kOhms ny my Luxamn C-03 preamp.And that Ortofon outputs 1.6 mV/cm, i.e. exactly as you say your Denon does, so no extra gain there.
I have a Linn LP12 Ekos in the loft with Lyra Helikon. Far too fragile for family use. I have never met anyone with an Ekos Denon DL103. I suspect it would shine. Would be great fun to build a new pre amp. I could go double inverting if Denon ( 100R lower gain arm ).
The mass I added to the Thorens arm was to drive Fo down ( Fo =1/ 2 Pi root /(MC) ).
Also a Garrard 401 is being sorted out. Looking for a Hadcock 228 arm. Think there is a rare S shaped one at Loricraft. I have an Ortofon ML25 calibration, I would like to use it. That has been in the Linn and was very good. It needed the tags splayed to get it in. It just aligns OK.
The mass I added to the Thorens arm was to drive Fo down ( Fo =1/ 2 Pi root /(MC) ).
Also a Garrard 401 is being sorted out. Looking for a Hadcock 228 arm. Think there is a rare S shaped one at Loricraft. I have an Ortofon ML25 calibration, I would like to use it. That has been in the Linn and was very good. It needed the tags splayed to get it in. It just aligns OK.
Telling SY not to overstep the line would kick start a new subthread on where the line is, and so forth, and, of course, you'd be asked for proof of its very existence.
And ultimately, what for? Professor van den Hul has been teaching metallurgy from about the time SY started going to the kindergarten, but as we have learnt here, that means diddley.
I'm perfectly happy to use his 102 Interconnect cables with some decent RCA plugs, his 352 Hybrid (said to be made of a combination of OFC and carbon threads) for myself (due to unusually long run of cables caused by room architecture, I need 5.5 sq.mm thick cabling) and 122 Hybrid (same as 352, only 3.5 sq. mm thick for shorter runs of up to 3 m, which is eanough for my wife's system). And I will continue to use them until I hear something better in my room and system.
I can't wait for SY to debunk prof. van den Hul and make a better cable at 1/5 of the cost. Because, after all, SY knows physics, while prof. van den Hul probably thinks phyiscs is a new type of sandwich.
Last edited:
Even in case of SE VAS working into CCS, distortion is increased if you add resistor VAS load.
Hi,
That old dragon of a uni-pivot? No, Nigel you do not want one of those.
If you want a good, well behaved uni-pivot go for one of Tom Fletcher's Omega Points.
(I think Tom's retired now but his products (read his engineering children) should still be around. He's also a man with big friendly audio heart, one that deserves a statue in British audio).
Best,
That old dragon of a uni-pivot? No, Nigel you do not want one of those.
If you want a good, well behaved uni-pivot go for one of Tom Fletcher's Omega Points.
(I think Tom's retired now but his products (read his engineering children) should still be around. He's also a man with big friendly audio heart, one that deserves a statue in British audio).
Best,
Van den Hul sells quack medicinal gadgets, preying on the desperation and weakness of the sick. That may be fine for you but in my view, it's pure evil. That separates him from mere frauds like Jack Bybee or Geoff Kait.
I should note that I made interconnects from carbon fiber at least ten years before he did. Unfortunately, I have scruples and abandoned it as a not-great idea instead of slapping a high price tag and a phony story on it and selling it to gullible audiophiles. Having a conscience and scientific integrity is often an impediment to commercial success. A better cable is available at <5% of the cost, assuming that by "better" you mean "audibly transparent" when using ears only to evaluate.
I should note that I made interconnects from carbon fiber at least ten years before he did. Unfortunately, I have scruples and abandoned it as a not-great idea instead of slapping a high price tag and a phony story on it and selling it to gullible audiophiles. Having a conscience and scientific integrity is often an impediment to commercial success. A better cable is available at <5% of the cost, assuming that by "better" you mean "audibly transparent" when using ears only to evaluate.
What "quack medicinal gadgets" are you talking about?
As for the pricing of his cables, I don't know how it goes elsewhere, but locally CS 122 sells for some €15 per meter, while 352 sells for about €25/m the last time I asked. That's hardly a rip-off price, given that one does not normally buy cables every year.
As for the pricing of his cables, I don't know how it goes elsewhere, but locally CS 122 sells for some €15 per meter, while 352 sells for about €25/m the last time I asked. That's hardly a rip-off price, given that one does not normally buy cables every year.
- Status
- Not open for further replies.