Yes. That is his design approach. He had wide latitude from my basic direction. Such as inherently linear CMA topology and high slew rate etc.
It is very linear before any over-all feedback.
Dr. Arto Kolinummi's book ends with an inherently linear, ultra low distortion design topology built that has zero over-all neg feed back. But adding error correction for the output stage.
THx-RNMarsh
Last edited:
That's one of my favorite Diy audio-based articles. Content and writing style.
Steel core with copper plating of the central conductor, or copper core?
Most RG-59/RG-59U I have seen was steel core with copper plating.
Hi Robert,
Are you kidding me? That was in the 80's and 90's. I won't remember them, but I'll try. Some of Triumphs stuff sounded pretty good - close to the actual sessions. While not off the charts for super sound quality in a classical sense, those recordings were true to the performance. There were a coulpe children's books of spoken word. A few Symphonies and other pop music. I don't drag them out to listen to because they were true to the performance. I play music in general depending on my mood. I'm pretty sure that goes for most of us here. Oh yeah, some Russian Military bands and some organ pieces. Pretty much anything except RAP, Country & Western, Opera and stuff like that. Damn few folk pieces, but some Celt stuff and similar.
Basically, it was a group mix of various music that took place over a month or so and a group opinion that was in my favor (lucky me). It was nearly the least expensive system with the best sound overall. I did get a fair amount of work from that experience, so I was pleased bout that.
-Chris
Are you kidding me? That was in the 80's and 90's. I won't remember them, but I'll try. Some of Triumphs stuff sounded pretty good - close to the actual sessions. While not off the charts for super sound quality in a classical sense, those recordings were true to the performance. There were a coulpe children's books of spoken word. A few Symphonies and other pop music. I don't drag them out to listen to because they were true to the performance. I play music in general depending on my mood. I'm pretty sure that goes for most of us here. Oh yeah, some Russian Military bands and some organ pieces. Pretty much anything except RAP, Country & Western, Opera and stuff like that. Damn few folk pieces, but some Celt stuff and similar.
Basically, it was a group mix of various music that took place over a month or so and a group opinion that was in my favor (lucky me). It was nearly the least expensive system with the best sound overall. I did get a fair amount of work from that experience, so I was pleased bout that.
-Chris
Plain copper single core, plain copper braid. Garland Cables CC59 Coax Cable, RG59B/U, 23 AWG 84% Braid 75 Ohm
View attachment 706198
This stuff is pretty much the cheapest coax around, I got this reel from a recycling center for a few dollars so for fun that's what I'm using for interconnect and speaker lines for now. 🙂
This cable is 'pure' copper (no platings, no foils) and is exactly what I was looking for, the giveaway price was a nice bonus.
Dan.
Last edited:
To buy in to Putzeys one has to buy in to the idea that open loop bandwidth doesn’t matter.
Open loop bandwidth doesn't matter, unity gain bandwidth does. If it behaves like an integrator over a larger bandwidth, then, great, it probably enjoys better performance at lower frequencies for the same unity gain.
That was simple. Glad we can move on.
That was simple. Glad we can move on.
Do you remember how I abused my Maplin Mosfet Amp? A little memory test for us both 🙂
Let me enlighten you. Putzeys promotes class D which is severely bandwidth limited. With a 1st order integrator the amount of NFB in the audio band would be very low compared to a linear amplifier. So he says you can fix this by using a 6-pole (I think) Bessel filter in the forward path which rolls off the gain at an extreme rate with extreme phase shifts.
IOW he is effectively saying that the performance of a slow amplifier can be as good as that of a very fast amplifier by using filters.
What is wrong with this picture?
I remember thinking you should have been arrested for it.
Was it reducing the closed loop gain?
You are a true maverick.
Why not reduce the open loop gain below 1 to get even more feedback? 😛
Well that was a dumb mistake. I meant "reduce the closed loop gain below 1".
Now my cunning suggestion makes complete sense, right?
Might be a little too quiet, was fine for even moderate listening levels. C'mon we gotta be realistic here......what's it all about?? 😕🙂
Derfy,
You certainly have the ability to try M. Schroeder's experiment to determine sensitivity to phase shift. You generate two sine waves and vary the phase between them making sure the envelope peak remains the same. Typical would be to have one say -15 degrees and the other +15.
Both Schroeder and Neve concluded 5 degrees at 20 kilohertz was the threshold of perception. Neve based his observation on a defective mixer channel and traced it to different in input transformer performance so it could have been an amplitude change.
So you might want to allow peak level changes.
Try 6,000 and 18,000 hertz test tones and see what you can hear for phase shift differences and then relate that to band width.
Then we can talk about amplifier high frequency response.
As a hint many folks use a simple input low pass filter to improve the perceived amplifier performance.
You certainly have the ability to try M. Schroeder's experiment to determine sensitivity to phase shift. You generate two sine waves and vary the phase between them making sure the envelope peak remains the same. Typical would be to have one say -15 degrees and the other +15.
Both Schroeder and Neve concluded 5 degrees at 20 kilohertz was the threshold of perception. Neve based his observation on a defective mixer channel and traced it to different in input transformer performance so it could have been an amplitude change.
So you might want to allow peak level changes.
Try 6,000 and 18,000 hertz test tones and see what you can hear for phase shift differences and then relate that to band width.
Then we can talk about amplifier high frequency response.
As a hint many folks use a simple input low pass filter to improve the perceived amplifier performance.
Ed
I am deaf to 18kHz and quasi deaf to 6kHz, so I won’t do that test.
I only want to ask, does one would have to check first for IMD performance of the audio chain to be utilized (using the same 6k &18k stimulus) in that test before concluding for his/her audibility capabilities?
George
I am deaf to 18kHz and quasi deaf to 6kHz, so I won’t do that test.
I only want to ask, does one would have to check first for IMD performance of the audio chain to be utilized (using the same 6k &18k stimulus) in that test before concluding for his/her audibility capabilities?
George
Maybe we are not as far apart as appeared at first. My own view is that the way forward is not added distortion, but reducing whatever distortion it is which is currently limiting the more realistic illusion while not increasing other distortions which we already know about.mmerrill99 said:
I was speaking of the electrical effect on varying feedback. Subjective preference is irrelevant. You were asserting that the effect of feedback itself can easily be distinguished from the tone changes it will also produce. This is simply false, however many people believe it and make money from convincing others it is true.Destroyer OS said:
Where did I say that? If you can find it please let me know and I will instantly retract it, because it is certainly not what I believe. I do believe that 0.1% distortion is better than 5% distortion, but that is quite a different statement.
Sorry, I have absolutely no idea what you are trying to say here. Buffers usually have more feedback than amps.
Let me enlighten you. Putzeys promotes class D which is severely bandwidth limited. With a 1st order integrator the amount of NFB in the audio band would be very low compared to a linear amplifier. So he says you can fix this by using a 6-pole (I think) Bessel filter in the forward path which rolls off the gain at an extreme rate with extreme phase shifts.
IOW he is effectively saying that the performance of a slow amplifier can be as good as that of a very fast amplifier by using filters.
What is wrong with this picture?
I can't see anything wrong. The maths works, the product works (and sells well and wins awards). Hypex make good money. The audio world is better for Bruno having done all this.
I didn't realize that we were taking a generalized concept of "open loop bandwidth", which, if I'm not mistaken, is specified as the -3 db (-6?) point where the open loop gain drops below its DC value, and treating it for very specific use cases that fit a narrative we want to create.
E.g. if we look at something like modern, unity-gain stable, opamp like the AD4898, it has an open loop bandwidth around 3 kHz (100 MHz UGBW, ~130 dB at DC). Is this, when used in a low gain (lets say <40 dB as "low") going to have any significant phase shift? MAYBE some of the slower older opamps run at high gain, e.g. a TL072 run at a G=40 dB will have some roll-off at the top of the bandwidth. But I'd argue you're using the wrong part for the job.
Bruno's high order class D topologies have nothing to do with that question and it's disingenuous to conflate the two. The bandwidth of our present class-D output stages necessitate that the output-inclusive loop will need to have high order roll-off if you want large loop gain through 20 kHz. Funny that's going to require a wide bandwidth amplifier/control block. Hmmmmmmm.
P.S. Ed, I'm more in the business right now of stuffing my ears with expandable memory foam to protect whatever HF sensitivity I have and making sawdust. Plus I'm in the final throes of a nasty, nasty cold so I'm still permanently stuffed up. Doubt my hearing, much less my mental acuity, is there for your low phase angle experiment.
E.g. if we look at something like modern, unity-gain stable, opamp like the AD4898, it has an open loop bandwidth around 3 kHz (100 MHz UGBW, ~130 dB at DC). Is this, when used in a low gain (lets say <40 dB as "low") going to have any significant phase shift? MAYBE some of the slower older opamps run at high gain, e.g. a TL072 run at a G=40 dB will have some roll-off at the top of the bandwidth. But I'd argue you're using the wrong part for the job.
Bruno's high order class D topologies have nothing to do with that question and it's disingenuous to conflate the two. The bandwidth of our present class-D output stages necessitate that the output-inclusive loop will need to have high order roll-off if you want large loop gain through 20 kHz. Funny that's going to require a wide bandwidth amplifier/control block. Hmmmmmmm.
P.S. Ed, I'm more in the business right now of stuffing my ears with expandable memory foam to protect whatever HF sensitivity I have and making sawdust. Plus I'm in the final throes of a nasty, nasty cold so I'm still permanently stuffed up. Doubt my hearing, much less my mental acuity, is there for your low phase angle experiment.
Last edited:
- Status
- Not open for further replies.