Surely its the other way round - if the amp is in danger of slew-limiting it can produce lots of harmonics across the board once it starts limiting.
Slowness of an amp isn't normally due to the bandwidth of the individual devices, but to low standing currents for the capacitances they need to charge. Another way to think of slew-limiting is as clipping of internal currents.
A fast amp has a better chance of correcting errors more successfully through GNF and less chance of slew-limiting.
Its pretty much universal in amps that the THD plotted against frequency starts to inexorably rise above a transition frequency, and this transition is generally lower for inherently slow amps - again this means slow amps distort more than fast amps, at the top end.
Last edited:
It still produced 12 watts @20 kHz without clipping. Total unclipped power was 40 watts @ 1 kHz.
This is tubes we're talking about. That kind of power was a lot for the era. The performance of this unit was actually better than most tube equipment.
The slowness of this amp was due to the output transformer. Transformers introduce a low frequency pole, as well as a high frequency pole.
You'll never find a transformer coupled tube amp with blazing speed. Plus 1% THD was considered good for the era. This is antique stuff I'm talking about.
Plus you must have missed the part where I said it only had 10 dB feedback @ 20 kHz. There's a reason for that
and it's not to generate euphonic sound.
This is tubes we're talking about. That kind of power was a lot for the era. The performance of this unit was actually better than most tube equipment.
The slowness of this amp was due to the output transformer. Transformers introduce a low frequency pole, as well as a high frequency pole.
You'll never find a transformer coupled tube amp with blazing speed. Plus 1% THD was considered good for the era. This is antique stuff I'm talking about.
Plus you must have missed the part where I said it only had 10 dB feedback @ 20 kHz. There's a reason for that
and it's not to generate euphonic sound.
Last edited:
A tube seems (correct me if I'm wrong) rather fast in terms of slew rate, depending on the grid capacitances, but the caps in use often seem to be the bottleneck made worse by feedback attempting to correct for it.
Member
Joined 2009
Paid Member
I do believe you and your colleagues heard these differences. Most interesting. A couple of questions,
1/ How do your results line up with common comments on these resistors, do you confirm what is the prevailing wisdom or did you find any surprises?
2/ Different circuit locations may be more sensitive than others to resistor quality, do you have experience or an opinion on which positions in the circuit should receive the most attention for resistor choice (solid state and tube) ?
The output transformer is the big bottleneck. Low frequency pole + high frequency pole must be dealt with.
This is what I tried to point out, a couple of times.
That position in the circuit is not that likely to influence the sound, because it's shunted by the source. Changing the resistors in the feedback network, or a filter network, would likely enhance the results.
I am not commenting on this test, but the most clearly obvious difference I have heard is in the feedback resistor for a current feedback amp.
The situation changes with higher voltage circuits, Vcr becomes the dominant issue. And many resistor manufacturers don't even spec that anymore.
EDIT: looks like several of us posted the same thing at the same time...
Member
Joined 2009
Paid Member
The feedback series resistor is typically like 20x as large in value as the shunt one. So any voltage- and tempco variations are much higher in the series resistor so will cause distortion.
The trick from Audio Precision is using the same resistor for series as shunt, but use 20 in series to get to the 20x. In that way, all tempco and voltco effects are the same in each resistor and cancel out at the feedback point.
Jan
The trick from Audio Precision is using the same resistor for series as shunt, but use 20 in series to get to the 20x. In that way, all tempco and voltco effects are the same in each resistor and cancel out at the feedback point.
Jan
CertainlyCan you tell me more about NP's "H2 circuit"? (I'm intrigued.
Thanks,
Andy
Thread here: H2
Article here: https://www.firstwatt.com/pdf/art_h2.pdf
Potentially relevant bit here:
So why is the phase important? Well, it's a subtle thing. I don't suppose everyone can hear
it, and fewer particularly care, but from listening tests we learn that there is a tendency to
interpret negative phase 2nd as giving a deeper soundstage and improved localization than
otherwise. Positive phase seems to put the instruments and vocals closer and a little more
in-your-face with enhanced detail.
This is in danger of dragging this thread way off topic, but what you describe makes sense, especially for speaker systems, where HD is much higher. When I first got the ability to measure and plot individual harmonic magnitude and phase vs freq in the early 80's (via Dick Heyser's software for the Techron TEF analyzer), I was stunned by how different various drive units were. On some models the harmonics changed phase polarity as they went from piston to direct radiation mode. But I had bigger fish to fry and didn't delve into it as deeply as it probably deserved.
Came here to post exactly this. I’m using this trick in the summing loop of a PSE EL34 amplifier I am building (per the Electra-Print schematic).
This is as stupid as it gets. A shunt input resistor cant change the signal, unless its overloading. I guess Ohms law is a "theory" if you know Zero knowledge in electronics.
And if theres one ignorant deluded audiophool believing nonsense theres a thousand. Lol religion makes you hear different.
Last edited:
Ohm's law is derivable from Maxwell's Equations. In other words, Ohm's law is a mathematical model of reality. Physical resistors are not perfectly linear, time-invariant, and or stationary. That said, good metal film resistors are pretty good for most audio purposes. Are they always perfectly inaudible if not 'overloaded?' Not everyone agrees on the answer to that.
I think a shunt resistor can effect the signal, a shunt volume control is discussed here Simple Shunt volume control by Potentiometer?
Not the same thing.
A typical output impedance for line level is 100 ohms (it does vary). A typical input impedance for line level is 10-100K (47K being "standard").
Now answer yourself this question. If you shunt a 47K resistor with 100 ohms, what happens to the noise generated by the 47K resistor? Which resistor's noise and nonlinearity profile will dominate?
That's right, and you can get some pretty unusual results driving line level inputs with high impedance outputs.
Anyone from back in the day that has tried to drive a solid state power amplifier with an "old school" tube preamplifier (which often had quite high output impedances) knows exactly what I'm talking about. You can get some very unwanted results. But, driving an old school tube amp (some of them had input impedances of 1 megohm) with a solid state preamplifier not only works fine, but reduces the noise profile of the tube amplifier considerably.
I think most of us know why this happens.
- Status
- Not open for further replies.