That's it, John.
The amp is back to bipolars becvause with them, I can obtain better open loop performance, which simply means that I have yet to learn a lot about using FETs. I realize that I am the problem, not the devices.
However, perhaps that's for the better, because I have added a FET buffer. In principle, I like the idea of buffers - they do complicate things a bit, but they also afford me the freedom to do as I like after them almost limitelessly.
As shown, it is capable of working continuosly into 2 Ohm loads, at least until the overheat protection kicks in. I reckon that if it can work into 2 Ohms steady state, I should not have any problems negotiating anything remotely resembling a real world loudpeaker.
As ever, the protection is a concern. Well, I think I've pinned it down fairly well. You can remove all of the overvoltage/overcurrent protection circuit, a classic two slope design, without any repercussions, absolutely NOTHING changes. That's because its threshold of operation has been set for 1.8 Ohms, rather low, but then those power devices are well neigh indestructable, I think I have earned the right to say that after working with them for over 20 years.
The rest is more or less standard fare, nothing exciting about it.
The current stage power supply is also tentative. Using a total of 69,400 uF per channel, with a solid 1,000 VA toroidals transformer made to my specs, is more than enough for even the worst case. I drew that one quickly, so there may be a mistake or two, I have yet to review it carefully.
The inidicator board is a "maybe". Ever since the days of SAE, I've been kind of nuts about LED peak meters, and using two per channel could give me a dyinamic range of almost 60 dB - no mechanics, no coils, no inertia, but let's face it, that kind of look is dead and buried. On the other hand, I do like my amp to be able to "talk" to me, so this is the most probable version to be used.
It offers a classic "Power on" LED, but also a "Signal present", "0 dB" and "Clip" LEDs. O dB LED is set for 28.5 Vrms output, and "Clip" will start to blink at 32.2 Vrms. That's not the actual clipping point, but I feel it should start warning me that I am dangerously close to real clipping BEFORE it actually happens.
That would give me a total of 6 LEDs per channel, which is a managable number, it can be arranged so that it's both visible and hence useful, but not to be intrusive or overbearing or flashy.
The only thing missing from this collection is the Shunt regulator. In a subsequent post, I'll provide the details of such a unit with an almost illustrious history, along with its schematics - it has been in public domain for a few years anyway, and was designed by my good friend and associate Ollie, who does my PCB artwork and recheks my work - it doesn't leave me until Ollie okays it, I trust him that much, to the hilt.
As ever, any suggestions are most welcome, many eyes will see what the two may have missed.
Oh yes, lest I forget - it's based on Otala's principles.
Well, no The Shuntyy today, I'm afraid.
I suspect that's because of the file size, about 450 kB in PDF format. I tried to upload it twice, but for some reason, it would not budge.
Basically, that's the best series/parallel, low noise shunt reguator I have ever seen or heard. But, it's rather complicated, The file contains the print artwork, so you can make it straight away without further ado.
Since it won't upload, if anyone is interested, just drop me a line to my e-mail box, which is:
dvv@beograd.com
and I'll be happy to send you your own copy directly.
When it was first designed, in 2006, copies were submitted to (Papa) Erno Borbelly and Nelson Pass, with a request to take a look and state their objections, suggestions, etc at will. Both requests came back with essentially the same comment - that's it, don't touch it any more, just make it.
Since then, I made a few of them and all I can say is that if anyone doubts the impact a power supply can have for an audio circuit, use this one and you will become a believer.
Not easy to make, not cheap either, but it works as advertised - and then some.
Ollie and I were planning on a "Mark II", but the current economic crisis has stopped us dead in our tracks, simply because we need to survive first, in a poor countzry with an unemployment rate of over 40%. It ain't easy, but it doesn't leave us much time for fun and games.
I suspect that's because of the file size, about 450 kB in PDF format. I tried to upload it twice, but for some reason, it would not budge.
Basically, that's the best series/parallel, low noise shunt reguator I have ever seen or heard. But, it's rather complicated, The file contains the print artwork, so you can make it straight away without further ado.
Since it won't upload, if anyone is interested, just drop me a line to my e-mail box, which is:
dvv@beograd.com
and I'll be happy to send you your own copy directly.
When it was first designed, in 2006, copies were submitted to (Papa) Erno Borbelly and Nelson Pass, with a request to take a look and state their objections, suggestions, etc at will. Both requests came back with essentially the same comment - that's it, don't touch it any more, just make it.
Since then, I made a few of them and all I can say is that if anyone doubts the impact a power supply can have for an audio circuit, use this one and you will become a believer.
Not easy to make, not cheap either, but it works as advertised - and then some.
Ollie and I were planning on a "Mark II", but the current economic crisis has stopped us dead in our tracks, simply because we need to survive first, in a poor countzry with an unemployment rate of over 40%. It ain't easy, but it doesn't leave us much time for fun and games.
The carbon film is connected via two crimped connections. Those can become loose if the board is subjected to vibrations. Like... falling gear from the back of the truck while unloading for the next gig.
Sims cannot take that into account.
Sound Quality Vs. Measurements?
Heck, no, this tread shold be called "everything that can be typed here"
Last edited:
The carbon film is connected via two crimped connections. Those can become loose if the board is subjected to vibrations. Like... falling gear from the back of the truck while unloading for the next gig.
Sims cannot take that into account.
Sound Quality Vs. Measurements?
Heck, no, this tread shold be called "everything that can be typed here"
That is the main fader this is the gain trim pot.
102 kHz music
I just stumbled across this by complete accident . The spectrum of musical instruments far above what is commonly thought .
There's life above 20 kilohertz! A survey of musical instrument spectra to 102.4 kHz
I just stumbled across this by complete accident . The spectrum of musical instruments far above what is commonly thought .
There's life above 20 kilohertz! A survey of musical instrument spectra to 102.4 kHz
You might want to follow up the reading with this:
Peripheral perception mechanism of ultrasonic hearing.
Ultrasonic masker clarifies ultrasonic perception i... [Hear Res. 2003] - PubMed - NCBI
Peripheral perception mechanism of ultrasonic hearing.
Ultrasonic masker clarifies ultrasonic perception i... [Hear Res. 2003] - PubMed - NCBI
But for HiFi what counts most is up till what frequency a human being can perceive sound. Elefants and whales can hear into far infrasound and bats and loads of birds of different feather way into infrasound. They would need different sound systems to replicate their reality, for my kind 20-20Khz is ok.
vac
Scott,
The issue is not when things work right, it is real world abuse. Assume the 1000 uF capacitor has a leakage of say .01 uA per uF x V. For even an 8 volt unit that would be 80 uA across 100K that would be .1 volt and across 10K .01 volts. The gain at maximum would be 60 dB. at minimum maybe 30 db. so that .09 volts would become 2.844 volts of pop! Now if the gain structure were set wrong you might be able to damage a speaker. But if the gain structure is set so you clip around 7 to 12 volts, then you will hear noise but most likely no damage.
E.S.
NE 5534 boosted op amp
[/IMG]
D V V . I think the boosted op amp idea is great . Alas my notes on my circuit are lost although so much like what you did to be not worth showing ( No 3 ) . This version would drive a 4 ohms speaker . The results were what one would get working from theroy alone . All the better rules applied I should say ( decoupling , gain etc ) . This was a class A design running approximately 1 A . It used simple diode and resistor bias ( vbe was tried for bias also ) .
Notice how both op amps are loosely speaking generics of 741 . Notice also how 5534 is unhappy . Better op amps were tried with startling differences . The 5534 problems are easy to cure with comp cap . However.....
The LF 351 N was kept as control during tests . It was much better than expected . TLO71 measured almost the same , it sounded different and to me less good .
The test also shows how the 5534 marginally had a higher output . I reran these tests and can say they are representative . Note all op amps had a surplus of current to drive the output stage .
The amp works well with no output stage feedback to the op amp . This is shown with feedback .
I was going to design a discreet op amp after this and never got far . A blameless TDA2050 if you like .
These are just my workshop notes .
D V V . I think the boosted op amp idea is great . Alas my notes on my circuit are lost although so much like what you did to be not worth showing ( No 3 ) . This version would drive a 4 ohms speaker . The results were what one would get working from theroy alone . All the better rules applied I should say ( decoupling , gain etc ) . This was a class A design running approximately 1 A . It used simple diode and resistor bias ( vbe was tried for bias also ) .
Notice how both op amps are loosely speaking generics of 741 . Notice also how 5534 is unhappy . Better op amps were tried with startling differences . The 5534 problems are easy to cure with comp cap . However.....
The LF 351 N was kept as control during tests . It was much better than expected . TLO71 measured almost the same , it sounded different and to me less good .
The test also shows how the 5534 marginally had a higher output . I reran these tests and can say they are representative . Note all op amps had a surplus of current to drive the output stage .
The amp works well with no output stage feedback to the op amp . This is shown with feedback .
I was going to design a discreet op amp after this and never got far . A blameless TDA2050 if you like .
These are just my workshop notes .
Last edited:
Oxford
Hi John . Shame I lost the circuit , it was simple and nice . The buffer proved very useful for comparing op-amps . I was supporting D V V and saying it is a good idea . Two modest op amps representing in miniature the whole debate . It is 2 minutes work to try another op-amp and opinions on op-amps is another matter .
Last night by semi-chance I visited the streets where two of our greatest audio people lived . Plantation Road and Walton Well Road . Two minutes walking distance apart . Percy Wilson and Michael Gerzon . I hope they knew each other ? How facilitating to have been there if they had . Michael I am sure would have written on this forum if he had lived into the age of broadband . As many will know his patents cover much of how modern digital works ( loss-less encoding ) .
Packing It All In
BTW. Alan Blumlein was dyslexic . I doubt he would have been on this forum . Michael is likened to Blumlein .
http://community.klipsch.com/forums/storage/3/520848/Horn Theory and the Phonograph.pdf
Hi John . Shame I lost the circuit , it was simple and nice . The buffer proved very useful for comparing op-amps . I was supporting D V V and saying it is a good idea . Two modest op amps representing in miniature the whole debate . It is 2 minutes work to try another op-amp and opinions on op-amps is another matter .
Last night by semi-chance I visited the streets where two of our greatest audio people lived . Plantation Road and Walton Well Road . Two minutes walking distance apart . Percy Wilson and Michael Gerzon . I hope they knew each other ? How facilitating to have been there if they had . Michael I am sure would have written on this forum if he had lived into the age of broadband . As many will know his patents cover much of how modern digital works ( loss-less encoding ) .
Packing It All In
BTW. Alan Blumlein was dyslexic . I doubt he would have been on this forum . Michael is likened to Blumlein .
http://community.klipsch.com/forums/storage/3/520848/Horn Theory and the Phonograph.pdf
Last edited:
Michael Gerzon was a genius, on the level of Richard Heyser. In fact, I got the two of them to get along with each other, about 30 years ago. Some tension, at first. Two very smart guys suspicious of each other, but as I knew them both, I knew they would get along, once they personally met each other. In my life experience, they were the two most innovative minds that I ever knew personally. Alas, both died young, before their time.
I still don't follow you. I don't see how you would get more than 10's of mV across the cap.
He might have been a great engineer but that is a bit of an overstatement. The concept of signal entropy came from Shannon, and in any case the predominant lossless format (FLAC) is free.
Michael Gerzon
You know if Jesus had made a secret visit to earth I would believe it to be Michael . He was the kindest man I ever knew . Very dignified and giving . If you are an atheist one should still accept how a person is said to be important .
My teacher for many years was another Oxford audio man . Unknown to all . 20 Lathbury Road which is also near the other two . I would guess Michael knew him .
Oxfordshire Blue Plaques Scheme
You know if Jesus had made a secret visit to earth I would believe it to be Michael . He was the kindest man I ever knew . Very dignified and giving . If you are an atheist one should still accept how a person is said to be important .
My teacher for many years was another Oxford audio man . Unknown to all . 20 Lathbury Road which is also near the other two . I would guess Michael knew him .
Oxfordshire Blue Plaques Scheme
Hobby
Audio was Michael's hobby gone wrong . I had the privilege of mildly discussing maths with him . A friend and I were convinced Michael would be the greatest mathematician who every lived . Alas this bloody hobby we all love got hold of him .
Audio was Michael's hobby gone wrong . I had the privilege of mildly discussing maths with him . A friend and I were convinced Michael would be the greatest mathematician who every lived . Alas this bloody hobby we all love got hold of him .
- Status
- Not open for further replies.