I think we need a hall monitor or something.
Robert have you ever built one passive RIAA and one very ordinary op-amp RIAA (carefully matched time constants) and nulled the outputs to each other? You would be surprised and might re-think your statements. An ambitious bystander might want to sim the two in LTSPICE where you get perfect resistors and capacitors.
An op-amp and associated feedback components reach a steady state condition, the non-minimum excess phase (true delay) is tiny (nsec). Note, there can be some contributed by distributed time constants in the base of transistors, this is still 10's of nsec but it will mask the simple speed of light numbers.
I have no interest in being dismissive or insulting but I am disappointed to see the feedback goes round and round yet again. I don't care who believes it it is wrong nonsense. Curious though I have only seen this by a certain group of audio designers are all the folks depending on MRI, CAT scan, pacemakers, ultra sound, etc. getting short changed?
Last edited:
Feedback reduces distortion, except in this thread.
So, op-amps do have more than one downside after all.
1 - They are painful when stood on.
2 - They are painful when sat on.
This thread turned out to be quite a revelation after all then eh 🙄
So, op-amps do have more than one downside after all.
1 - They are painful when stood on.
2 - They are painful when sat on.
This thread turned out to be quite a revelation after all then eh 🙄
Robert,
Let's think this through:
20 kHz sinusoid -> w = 2*pi*f = 125,663.7
Peak slope of a sinusoid is at n*pi. Easy one right at 0. 🙂
Let's assume 5 ns prop delay.
f(t) = sin(2*pi*f*t)
f(0) = 0
f(5ns) = 6.28e-4 (small angle approximation works nicely here)
And that error shows up in the feedback loop. So it gets attenuated.
Let's think this through:
20 kHz sinusoid -> w = 2*pi*f = 125,663.7
Peak slope of a sinusoid is at n*pi. Easy one right at 0. 🙂
Let's assume 5 ns prop delay.
f(t) = sin(2*pi*f*t)
f(0) = 0
f(5ns) = 6.28e-4 (small angle approximation works nicely here)
And that error shows up in the feedback loop. So it gets attenuated.
In the very beginning of Lounge R&D I built the basic active circuit out of Walt Jung's book. Last year I built the basic active JE990 (not the one with servo) phono stage with the Hairball Audio 990 board/kit. I really tried to to get the 990 to sound good - polystyrene caps, shunt reg power... It always had a slight compressed 70's sound - Even my wife could hear it on the other side of the house.
You certainly know our hot buttons.😀
You could try my experiment, in the very least you could prove to yourself that what is heard is not manifest in the two voltage vs. time signals.
Last edited:
It seems like the explanations having to do with op-amp delay, etc., largely exist to explain why passive RIAA circuits with ferrite core inductors may sound better. Even a CC resistor may find it's way into the circuit somewhere to make it sound slightly better. For people who truly want to believe what sounds better and more realistic must in fact be so, it calls for some explanation. Op-amp delay makes for good story (good for the telling - like good fiction - found by the reader to be coherent and complete, and without loose ends) for non-engineer types. Not so for trained engineers, though. Nonetheless, for those who with a need to believe that purity and realism is given only by inductor circuits, some other story may not suffice. If the explanation for why inductor circuits sound better is because they add in a little euphonic distortion due to saturation/hysteresis/ringing effects, that may be a more plausible story to the engineering types, but it might be seen as an attack on the beliefs of others who find the op-amp story more agreeable.
Lol.
If SHMBO comments from the other side of the house/kitchen, it's usually expression of displeasure.
Dan.
Don't forget ice, as I discovered this evening on an ungritted hairpin! I was exceeding grateful for an advanced feedback system to correct that.
D.
What amplifier bandwidth is required to get 5nS of delay? If I can build an amplifier that has open loop distortion of .1% and a small signal bandwidth of 10,000,000 hertz with a phase margin of 90 degrees gain of 1 at 10,000,000 hertz can you calculated the voltage delta % between the feed back signal and the input at a frequency of 6283 hertz?
As to the general tone here I think I just might go back to using the funny sprinkles on my ice cream. That seems to make as much sense.
ok, we'll keep trying just a little longer ...
🙂
here's some (pretty good actually) educational reading material:
http://linearaudio.net/sites/linearaudio.net/files/volume1bp.pdf
some wisdom gained over time:
0. never speak ill of mom
1. don't tug on superman's cape
2. don't f*** with mother nature
3. don't even try to go against the laws of physics (unless you are the Q or captain kirk)
😉
mlloyd1
🙂
here's some (pretty good actually) educational reading material:
http://linearaudio.net/sites/linearaudio.net/files/volume1bp.pdf
some wisdom gained over time:
0. never speak ill of mom
1. don't tug on superman's cape
2. don't f*** with mother nature
3. don't even try to go against the laws of physics (unless you are the Q or captain kirk)
😉
mlloyd1
I realize that folks need a story to go with their effects boxes but if you are deliberately constructing a circuit to be non-ideal (i.e. make an effect) anything goes why even bother discussing it?
😱 All i can say is - RUN!
I know of at least one bloke who bored his wife with audio so much that when asked does that sound better 'Yes Dear' does that sound worse 'Yes Dear'.
I need to spend the kids inheritance on these new speakers.......
Try it, you never know 😀
Because designing a good one is non-trivial engineering problem. Modeling of ferrite core distortion is still not all that accurate.
My wife is the other half of Lounge. She is taking on the manufacturing more and more. Soon I just will be locked in a room with my bench and food will be slipped under the door while I design. 😉😀
- Status
- Not open for further replies.