Hugh, remember the TV commercial for a chocolate bar from the 1970's ? The jingle goes "the Milky Bar Kid is tough and strong, the Milky Bar Kid just can't go wrong..." as this kid in a cowboy suit bursts into a bar with guns drawn. Francec was making a comparison to a previous post.
BTW, your comments at the recent MAC meeting were right on the money. I hope the 100+ people present understood enough of what you said to take something away from the meeting in a different light to what they may have had previously.
Through experience, one's own ideas change. Mine certainly have over the last 35+ years of my hi-fi journey.
Is there one facet of power amp design that you would like to share which you think is really important? (I've read and digested Self, Slone, Pass, and designed my own amps etc, so you can wax lyrical about technical aspects)
Not only can some of our Members not read, it seems some of us can't see either.Didn't you guys see the email circulating with Kevin Rudd's likeness on a Milky Bar wrapper ?
SandyK
Too true!
BTW just returned from listening to Hugh latest amp, which is really a Hi Fi starter project, simple to build, stereo amp and power supply all on one PCB, the Naksa. Just supply your own box, transformer, switches and connectors and off you go. Very impressed!
Happy Birthday Hugh, hope you sell loads!
The coffee was OK too!
BTW just returned from listening to Hugh latest amp, which is really a Hi Fi starter project, simple to build, stereo amp and power supply all on one PCB, the Naksa. Just supply your own box, transformer, switches and connectors and off you go. Very impressed!
Happy Birthday Hugh, hope you sell loads!
The coffee was OK too!
Thanks Jon, Lyn, Andrew and Sandy,
I had not noticed the KRudd likeness on the Milky Bars, interesting, isn't it? We have a genuine bureacrat here, in love with process...... fascinating!
Glenn,
The MAC meeting did not precipitate any discussion on the forum group as I'd hoped. I deliberately took aim at a number of sacred cows and frankly expected heckling, but none was forthcoming, and I was very surprised. I am by nature a rebel, and love to take a swipe at the high priests of audio design (not all, but many of them) and have been quite vigorously attacked at times - entirely to be expected.
My strong opinions notwithstanding, I'm actually very surprised that I did not prick a response from you, to be frank; I know your website, your journey, and your many good designs, but I can promise you my approach is quite a bit different, not necessarily better or worse however, just different.
Technical matters: I don't normally play these cards, as the debates they create are interminable, but here's a few:
#1 We should strive for monotonic decreasing distortion spectrum, if possible with odd orders at least 20dB lower than the preceding even order artefacts, and H2 no higher than -70dB to start out. I'm also very interested in transfer at 20KHz.
#2 Increasing evidence points to the phase of the artefacts having influence on sound quality. LTSpice lists these phase relationships; they are fascinating. Phase shift varies for the distortion artefacts for reasons not entirely clear.
#3 Phase shift at 100KHz should be less than 10 degrees. 3 degrees is an excellent figure to strive for. This influences imaging focus and spatial integrity.
#4 Any coupling caps should always have an appreciable DC biasing voltage across them. This puts the dielectric into its more linear strain region and the result is a much better sounding transfer.
There's four to begin with..... hope this prompts some debate!
Cheers,
Hugh
I had not noticed the KRudd likeness on the Milky Bars, interesting, isn't it? We have a genuine bureacrat here, in love with process...... fascinating!
Glenn,
The MAC meeting did not precipitate any discussion on the forum group as I'd hoped. I deliberately took aim at a number of sacred cows and frankly expected heckling, but none was forthcoming, and I was very surprised. I am by nature a rebel, and love to take a swipe at the high priests of audio design (not all, but many of them) and have been quite vigorously attacked at times - entirely to be expected.
My strong opinions notwithstanding, I'm actually very surprised that I did not prick a response from you, to be frank; I know your website, your journey, and your many good designs, but I can promise you my approach is quite a bit different, not necessarily better or worse however, just different.
Technical matters: I don't normally play these cards, as the debates they create are interminable, but here's a few:
#1 We should strive for monotonic decreasing distortion spectrum, if possible with odd orders at least 20dB lower than the preceding even order artefacts, and H2 no higher than -70dB to start out. I'm also very interested in transfer at 20KHz.
#2 Increasing evidence points to the phase of the artefacts having influence on sound quality. LTSpice lists these phase relationships; they are fascinating. Phase shift varies for the distortion artefacts for reasons not entirely clear.
#3 Phase shift at 100KHz should be less than 10 degrees. 3 degrees is an excellent figure to strive for. This influences imaging focus and spatial integrity.
#4 Any coupling caps should always have an appreciable DC biasing voltage across them. This puts the dielectric into its more linear strain region and the result is a much better sounding transfer.
There's four to begin with..... hope this prompts some debate!
Cheers,
Hugh
Last edited:
Hi Aksa,
in response to 4.
If the electrolytic does not have a DC bias, then add one, seems to be what you propose.
Does that mean that audio signals that need to pass a big capacitor should always use two bigger capacitors series connected back to back and apply a DC bias to the cap junction? Or is there another way?
in response to 4.
If the electrolytic does not have a DC bias, then add one, seems to be what you propose.
Does that mean that audio signals that need to pass a big capacitor should always use two bigger capacitors series connected back to back and apply a DC bias to the cap junction? Or is there another way?
Andrew,
As a general rule, the shunt signal caps are not as critical, though they do like a DC bias. I was more referring to the input caps, any series signal cap is very touchy and likes some DC.
A better solution would be to redesign the circuit to supply the bias, rather than use two in series with a step DC voltage in between......
Hugh
As a general rule, the shunt signal caps are not as critical, though they do like a DC bias. I was more referring to the input caps, any series signal cap is very touchy and likes some DC.
A better solution would be to redesign the circuit to supply the bias, rather than use two in series with a step DC voltage in between......
Hugh
Hugh,
In the interests of enjoyable meetings, the MAC committee has suggested that members don't heckle the presenter. Even if it might be justified, it can be embarrassing for the presenter, the audience and even the heckler. Every one is there to have a good time. Even if the sound is not that good or the presenter is not that interesting, people want to take away positives and not bear witness to sour grapes.
Its OK to confront the presenter after the meeting one-on-one, but criticism or polarised debate in front of an audience doesn't go down well.
I felt like putting my hands up a couple of times, but it would have been to say "I agree" anway, and a supportive comment about output stage design or long tail pairs versus other topologies would have gone over most peoples head.
I know. That's why I asked. All designers tend to be a bit coy. This is testing the waters.....
Agree, significant H3 and odd orders adds harshness. They are "compressive" distortions.
Yes, it may indicate why some amplifiers sound better than others, despite similiar "traditional" specs on paper. I think you are in the lead pack of "smarter" designers.
Unfortunately most back-yard designers don't even look at this. There is a tendency to just go for fairly wide bandwidth, throw in compensation caps, and if it doesn't oscillate then that's good enough.
I see your point. Kind of like cross-over distortion for caps, where the cross-over is the point of passing through zero polarisation. I've never liked the idea of an electrolytic cap in the bottom leg of the NFB chain. Its critical position for sound quality and there is very little bias present (a few mV). The popular alternative of adding a servo increases complexity, which most people shy away from.
Me too. Do you think further discussion should be continued in that crowded forum known as "solid state" ? Its much quieter over here in the AKSA forum.
In the interests of enjoyable meetings, the MAC committee has suggested that members don't heckle the presenter. Even if it might be justified, it can be embarrassing for the presenter, the audience and even the heckler. Every one is there to have a good time. Even if the sound is not that good or the presenter is not that interesting, people want to take away positives and not bear witness to sour grapes.
Its OK to confront the presenter after the meeting one-on-one, but criticism or polarised debate in front of an audience doesn't go down well.
I felt like putting my hands up a couple of times, but it would have been to say "I agree" anway, and a supportive comment about output stage design or long tail pairs versus other topologies would have gone over most peoples head.
I know. That's why I asked. All designers tend to be a bit coy. This is testing the waters.....
Agree, significant H3 and odd orders adds harshness. They are "compressive" distortions.
Yes, it may indicate why some amplifiers sound better than others, despite similiar "traditional" specs on paper. I think you are in the lead pack of "smarter" designers.
Unfortunately most back-yard designers don't even look at this. There is a tendency to just go for fairly wide bandwidth, throw in compensation caps, and if it doesn't oscillate then that's good enough.
I see your point. Kind of like cross-over distortion for caps, where the cross-over is the point of passing through zero polarisation. I've never liked the idea of an electrolytic cap in the bottom leg of the NFB chain. Its critical position for sound quality and there is very little bias present (a few mV). The popular alternative of adding a servo increases complexity, which most people shy away from.
Me too. Do you think further discussion should be continued in that crowded forum known as "solid state" ? Its much quieter over here in the AKSA forum.
Glenn,
Again, I'm most gratified, thank you for being so very positive.
If these arguments are presented in the SS section where there are thousands of lurkers, there will be long debate, explanation, and some angst. Generally, the better the argument, the more heated it becomes, since the world is dominated by conventionalists, bless 'em. Then the math guys will move in and scotch all the suggestions anyway without even switching on a soldering iron. Then the finger pointing, one or two sinbinned miscreants, and the cycle will start over.
I really think the only way to do this properly is to write a white paper for the website, which people can then read and ignore, knowing not many will be influenced, or indeed even see it. It is the way of the world, no less, and I ain't complainin'.....
I think NP's approach is probably the best. No confrontation, white papers used for new ideas (and old ones), carefully aimed, pithy comments in the forums, no acrimony of any kind, just a diffuse admiration from all. A very clever man, particularly in the way he handles people.
I love a fight, but less now as I near 60. It takes it out of me now, I can't afford to elevate my blood pressure without actually walking the park. Slowly I'm beginning to understand that new ideas are never welcomed; there needs to be marketing before new ideas are out there; and there is enormous vested interest in the status quo across all commercial spheres and no shortage of clever people to shore them up.
BTW, most of the true advances in technology come from back yard guys. They will try anything, and to keep costs down are very inventive, you included!
Cheers,
Hugh
Again, I'm most gratified, thank you for being so very positive.
If these arguments are presented in the SS section where there are thousands of lurkers, there will be long debate, explanation, and some angst. Generally, the better the argument, the more heated it becomes, since the world is dominated by conventionalists, bless 'em. Then the math guys will move in and scotch all the suggestions anyway without even switching on a soldering iron. Then the finger pointing, one or two sinbinned miscreants, and the cycle will start over.
I really think the only way to do this properly is to write a white paper for the website, which people can then read and ignore, knowing not many will be influenced, or indeed even see it. It is the way of the world, no less, and I ain't complainin'.....
I think NP's approach is probably the best. No confrontation, white papers used for new ideas (and old ones), carefully aimed, pithy comments in the forums, no acrimony of any kind, just a diffuse admiration from all. A very clever man, particularly in the way he handles people.
I love a fight, but less now as I near 60. It takes it out of me now, I can't afford to elevate my blood pressure without actually walking the park. Slowly I'm beginning to understand that new ideas are never welcomed; there needs to be marketing before new ideas are out there; and there is enormous vested interest in the status quo across all commercial spheres and no shortage of clever people to shore them up.
BTW, most of the true advances in technology come from back yard guys. They will try anything, and to keep costs down are very inventive, you included!
Cheers,
Hugh
Last edited:
Yeah, the SS forum is a big melting pot. Its quite high volume with a big variety of posts, but I its increasingly difficult to find really interesting discourse. The same can be said of the ChipAmps and Tubes forums.
Yeah, good idea. At least it makes the author think deeply about what they are writing, and they tend to answer a lot of their own questions in the process.
Heated debates are generally not worth it. Most people are resistant to change and have formulated their own ideas, and don't have a lot of time and money to explore new ones.
I would never build something exactly from a published design. I like to deconstruct the essence and reconstruct something fresh. The hours and hours of learning and construction challenges can be more rewarding than listening to the thing for hours and hours afterwards. Its usually a case of "that was fun to build, I'll use it in my system for a while. What's the next project? (or more likely, what existing project needs to be finished)".
Do you think ChipAmp drivers have any future in hi-end?
I've used a predecessor (LM4702 = stereo LME49811) to fit a 60Wpc stereo amp on a 8x10cm prototyping board (with 2 pairs of TIP142/147 hanging off the end). The circuitry is relatively simple but I'm not convinced the sound is particularly good. Maybe its the darlingtons that are letting it down.
Glenn,
Chipamps are cost driven, and redolent with protection circuits. Often the protection circuits compromise the sound quality, I find.
The other issue is component density. Music is pretty dynamic, so thermal effects ripple through the chip, altering the operating points. l think you have more chance of high sound quality with discrete devices precisely because these thermal effects can be isolated. Gerard Perot (Lavardin) has quite a bit to say about this, I think he is right.
Thanks for the posts!
Hugh
Chipamps are cost driven, and redolent with protection circuits. Often the protection circuits compromise the sound quality, I find.
The other issue is component density. Music is pretty dynamic, so thermal effects ripple through the chip, altering the operating points. l think you have more chance of high sound quality with discrete devices precisely because these thermal effects can be isolated. Gerard Perot (Lavardin) has quite a bit to say about this, I think he is right.
Thanks for the posts!
Hugh
Hugh,
Once again you prove yourself to be a true independent thinker & I admire you sir. I don't pretend to grasp the technical detail of your 4 points but I can recognise wisdom in the significance of your insights. I wish I was at a technical level to appreciate the full ramifications of what you say & I too bemoan the fact that these 4 points would unlikely be given a real airing & honest debate in the SS forum.
As you again wisely point out, real seekers of truth are a rare breed indeed & the general consensus usually rules. A lot of energy & effort is expended in how to manipulate this general consensus & the truth becomes the victim in this battle. Unfortunately this seems to be insidious in all aspects of modern life.
When I started writing this post I didn't think it would take that twist - hope it doesn't seem to end in a rant
Once again you prove yourself to be a true independent thinker & I admire you sir. I don't pretend to grasp the technical detail of your 4 points but I can recognise wisdom in the significance of your insights. I wish I was at a technical level to appreciate the full ramifications of what you say & I too bemoan the fact that these 4 points would unlikely be given a real airing & honest debate in the SS forum.
As you again wisely point out, real seekers of truth are a rare breed indeed & the general consensus usually rules. A lot of energy & effort is expended in how to manipulate this general consensus & the truth becomes the victim in this battle. Unfortunately this seems to be insidious in all aspects of modern life.
When I started writing this post I didn't think it would take that twist - hope it doesn't seem to end in a rant
So true. And advances in technology need not rely only upon advanced technology, as demonstrated by a brilliant countryman of yours (and backyard guy), Hugh...
YouTube - The D-Drive Infinitely Variable Geared Transmission
Cheers,
- Todd
John,
thank you for your post, very nice of you!
This is the classical description of an adversarial system, widely regarded in the West as the best way to do things. Does it remind you of the justice system?
I think that creativity for its own sake is often not found in large corporations which are geared solely to economic returns. There are exceptions, but by and large this is true. Most of the true advances in automobile engine technology, for example, originally come from the motorcycle companies.
Todd,
I looked at the D Drive, thanks for the link. In 1:1 it is reasonably efficient, but the weak point is that a torque input ranking with the input motive torque must be supplied in contra-rotation to manipulate the ratios. This could introduce losses, and certainly poses problems in terms of drive because the speed must vary with the ratio chosen. In a sense, this is merely transferring the problem. I do not share the enthusiasm for it; it is flawed.
If you can, find a RV Quad vehicle and remove the transmission cover. Between engine and chain wheel you will find an infinitely variable, friction drive using conical sheaf V belts and centrifugal control. This system was pioneered by DAF in Holland in the sixties. These are extremely efficient drives - life of the V belt is 10,000 kms on a quad (probably about 50,000 kms on an automobile) - and they deliver astonishing performance.
I drove a V12 BMW a few years ago with a seven speed transmission. Starting from the lights, I accelerated very gently with the engine at 1400, and as it cycled through the gears the car increased speed to 105 kph at this engine speed, with variation little more than 1350-1450 rpm. Incredible. Virtually an infinitely variable gearbox. They are here already!!
Today I serviced the autotrans on my old Cressida. The pan only took 2 litres of ATF!! This is little more than that required for lubrication, let alone power transmission. Yet fluid drives the torque converter (which holds about two litres more, as I understand). Modern auto transmissions are an engineering marvel.
Cheers,
Hugh
thank you for your post, very nice of you!
This is the classical description of an adversarial system, widely regarded in the West as the best way to do things. Does it remind you of the justice system?
I think that creativity for its own sake is often not found in large corporations which are geared solely to economic returns. There are exceptions, but by and large this is true. Most of the true advances in automobile engine technology, for example, originally come from the motorcycle companies.
Todd,
I looked at the D Drive, thanks for the link. In 1:1 it is reasonably efficient, but the weak point is that a torque input ranking with the input motive torque must be supplied in contra-rotation to manipulate the ratios. This could introduce losses, and certainly poses problems in terms of drive because the speed must vary with the ratio chosen. In a sense, this is merely transferring the problem. I do not share the enthusiasm for it; it is flawed.
If you can, find a RV Quad vehicle and remove the transmission cover. Between engine and chain wheel you will find an infinitely variable, friction drive using conical sheaf V belts and centrifugal control. This system was pioneered by DAF in Holland in the sixties. These are extremely efficient drives - life of the V belt is 10,000 kms on a quad (probably about 50,000 kms on an automobile) - and they deliver astonishing performance.
I drove a V12 BMW a few years ago with a seven speed transmission. Starting from the lights, I accelerated very gently with the engine at 1400, and as it cycled through the gears the car increased speed to 105 kph at this engine speed, with variation little more than 1350-1450 rpm. Incredible. Virtually an infinitely variable gearbox. They are here already!!
Today I serviced the autotrans on my old Cressida. The pan only took 2 litres of ATF!! This is little more than that required for lubrication, let alone power transmission. Yet fluid drives the torque converter (which holds about two litres more, as I understand). Modern auto transmissions are an engineering marvel.
Cheers,
Hugh
Yeah, the economic reason for natsemi and others producing these chip amps is to get them into high volume consumer equipment where quality is not paramount. However, 47Labs unintentionally brought in a whole new wave of diyer's with LM1875, 3875, 3386 etc ( I compare them at LM4780 / 3886 / 3875 / 1875 Comparison ). Some claim the extremely short signal path length and feedback path length is a good thing.
LM4702 and later chips of this ilk are a step forward because at least they can go above ~60W and the OPS can be determined by the designer.
You mentioned thermal effects. I imagine these can dynamically alter the behaviour of the amplifier with fluctuating signal level. Effects that can't easily be measured but impact the overall sound quality. I have one more of my chipamp based power amps to build (the 4th), and then I might call it a day, and go back to fully discrete designs. Those glowing bottle things deserve more attention too, I have a good collection of interesting new types to try out. I have something special in store for the MAC December GM! (I think 7 months will be enough time to finish it).
Your newer power supplies have a series resistor between two caps (ie. CRC), if I am not mistaken. I can see some reasons for doing this, but what are your main arguments for this design choice?
I concede to your engineering aptitude, my friend.
I have no related education so can only gauge by a mechanical intuition about these things, and I'm not getting a clear signal from this complicated baby; too many transfer points for my meager intuition to glue together. But it's certainly ingenious as a backyard contraption. But the inverted cone transmissions don't intuitively work well for me. Tensile strength of the belt plus the grip losses seem (intuitively) to be its limiting weaknesses. (My intuition has been proven inadequate many times in the past though.
)..Todd
Glenn,
My apology, distracted, busy day.
You are right, CRC supplies. I do it to quarantine the rectifier charge pulses from the speaker earth return currents. This reduces intermodulation; I can clearly hear an improvement. It also looks very cool, and means you can buy more caps of less size.
Recommended!
Hugh
My apology, distracted, busy day.
You are right, CRC supplies. I do it to quarantine the rectifier charge pulses from the speaker earth return currents. This reduces intermodulation; I can clearly hear an improvement. It also looks very cool, and means you can buy more caps of less size.
Recommended!
Hugh
What is this MAC thing? Sounds interesting!
There's so much information flowing like plasma from the bowels of the internet and there's just no way I can keep track of it all. Paul Bysouth keeps sending me things to read! AARRGH!!!
And then I only have an algebra 2 level education, so I can barely understand the technicalities of this stuff. On top of that, we've only got a few years till 2012! I hope I can get my Soraya up before then!
- keantoken
There's so much information flowing like plasma from the bowels of the internet and there's just no way I can keep track of it all. Paul Bysouth keeps sending me things to read! AARRGH!!!
And then I only have an algebra 2 level education, so I can barely understand the technicalities of this stuff. On top of that, we've only got a few years till 2012! I hope I can get my Soraya up before then!
- keantoken
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.