Thanks Lars, I'll do that. This was one of the 'issues' I was going to address in the real amplifier and my main concern is DC stability with the gain stage as it is. The present component values shouldn't be taken to literally as they were just thrown-in for the purpose of simulation without much regard for reality
In partitures isn't any distortion prescribed and any amp isn't musical instrument. Many generaions of technics was struggled for low distortion and you can now opposite. Is here anybody, who was listened some realy good amp with ultralow distortion in comparison with similar s...ts ? I mean that nobody. Distortion is not beautiful, although many peoples can " musical amp ", distortion is distortion. If you are sitting in opposite to the player with accoustic guitar, are you listening any distortion ? Certainly not - why you can listen it at home ?
I fail to see how a low cost amplifier board will result in a 'ultra low cost amp design'.
Lets assume we can lower the cost of the amplifier circuitry to 0. How much have we lowered the overall cost of the completed amplifier? 10%? 20%? The cost of manufacture, whether diy or commercial, is not in the amp itself, but rather the power supply, chassis, heatsinks, connectors, etc...
Low cost amp printed circuit boads are currently available, ESP, etc...
Certainly, the home builder can lower the costs of ancilliary components by using defunct, surplus, whatever, components. But we are left with the fact that a properly sized transformer (for a linear supply) is going to cost more than the amplifier circuitry. Same for the chassis, heatsinks....
With all the engineering expertise hereabouts, cannot someone come up with a 'proper' low cost switching power supply?
If one is seeking an 'ultra low cost amp design', one would be far better off using an existing amp manufactured in the far east, and depending on one's subjectivist proclivities, modding.
For those of us in the States, the Samson Servo, Art SLA-1, provide a lower cost base for DIY.
http://www.zzounds.com/item--SAMSERVO260
http://www.8thstreet.com/prod.asp?pid=6941
And above a certain power output, one inevitably realizes that given the costs of power supplies and heat sinking, a class 'D' amplifier offers sustantial cost benefits. (this is no surprise considering the commercial interests of the thread originator)
Lets assume we can lower the cost of the amplifier circuitry to 0. How much have we lowered the overall cost of the completed amplifier? 10%? 20%? The cost of manufacture, whether diy or commercial, is not in the amp itself, but rather the power supply, chassis, heatsinks, connectors, etc...
Low cost amp printed circuit boads are currently available, ESP, etc...
Certainly, the home builder can lower the costs of ancilliary components by using defunct, surplus, whatever, components. But we are left with the fact that a properly sized transformer (for a linear supply) is going to cost more than the amplifier circuitry. Same for the chassis, heatsinks....
With all the engineering expertise hereabouts, cannot someone come up with a 'proper' low cost switching power supply?
If one is seeking an 'ultra low cost amp design', one would be far better off using an existing amp manufactured in the far east, and depending on one's subjectivist proclivities, modding.
For those of us in the States, the Samson Servo, Art SLA-1, provide a lower cost base for DIY.
http://www.zzounds.com/item--SAMSERVO260
http://www.8thstreet.com/prod.asp?pid=6941
And above a certain power output, one inevitably realizes that given the costs of power supplies and heat sinking, a class 'D' amplifier offers sustantial cost benefits. (this is no surprise considering the commercial interests of the thread originator)
pmkap: You are absolutely right!
80% of the cost of the total amp is capacitors, transformers and enclosure.
However i have been looking at the current designs for DIY amplifiers based on chip amps, here on diyaudio.com and trought: this can be done better. DIY'ers seem to be very interested in low cost amplifier designs like the Gainclone, with it's inherent limitations. (Like Voltage, Current, Global Feedback Loops etc).
You can make better amplifier designs than using a chip amplifier and at low prices, however i would agree with you, it would probably be better and maybe even cheaper to buy a low cost ready built amplifier, and tweak it.
80% of the cost of the total amp is capacitors, transformers and enclosure.
However i have been looking at the current designs for DIY amplifiers based on chip amps, here on diyaudio.com and trought: this can be done better. DIY'ers seem to be very interested in low cost amplifier designs like the Gainclone, with it's inherent limitations. (Like Voltage, Current, Global Feedback Loops etc).
You can make better amplifier designs than using a chip amplifier and at low prices, however i would agree with you, it would probably be better and maybe even cheaper to buy a low cost ready built amplifier, and tweak it.
Lars,
No disagreement here.... And for those in the objectivist persuasion, given the measurements of these typical 'pro', low cost amps, diy simply does not make sense. Given this, I wonder why some of these objectivists post here, unless there is some gaping hole in their lives that is fufilled by taunting others....
As a 'flawed' subjectivist, I agree that the topology and implementation of an arbitrary amp (as well as the components chosen, subjectively) can have a profound effect. This is why many of us choose to obtain kits from the likes of yourself, Hugh Dean..... as we consider it a fair transaction to pay for that expertise and the 100s of hours of subjective component evaluation.
But as the amp circuitry itself is, at most, 20% of the total cost of an amplifier, how does one produce a truly Ultra Low Cost Amp Design? The obvious answer, whether Class A, AB, D, G.... is the linear power supply. I believe that near term EU mandates might well constrain (if not preclude) the standard transformer-bridge powersupply, and I assume that others besides Rowland, Linn, etc.... are persuing the switching powersupply route. While current off the shelf switchers might be as expensive, if not moreso, than the equivalent linear supply, the availability of switcher & PFC ICs, as well as the material costs of ferrite transformers should reverse that in the near term. Those very same constraints as to powerline harmonics and noise will ultimately benefit the DIY crowd. It is just a question as to when, and who with the requisite expertise, will make this knowledge available to our motley crew.
No disagreement here.... And for those in the objectivist persuasion, given the measurements of these typical 'pro', low cost amps, diy simply does not make sense. Given this, I wonder why some of these objectivists post here, unless there is some gaping hole in their lives that is fufilled by taunting others....
As a 'flawed' subjectivist, I agree that the topology and implementation of an arbitrary amp (as well as the components chosen, subjectively) can have a profound effect. This is why many of us choose to obtain kits from the likes of yourself, Hugh Dean..... as we consider it a fair transaction to pay for that expertise and the 100s of hours of subjective component evaluation.
But as the amp circuitry itself is, at most, 20% of the total cost of an amplifier, how does one produce a truly Ultra Low Cost Amp Design? The obvious answer, whether Class A, AB, D, G.... is the linear power supply. I believe that near term EU mandates might well constrain (if not preclude) the standard transformer-bridge powersupply, and I assume that others besides Rowland, Linn, etc.... are persuing the switching powersupply route. While current off the shelf switchers might be as expensive, if not moreso, than the equivalent linear supply, the availability of switcher & PFC ICs, as well as the material costs of ferrite transformers should reverse that in the near term. Those very same constraints as to powerline harmonics and noise will ultimately benefit the DIY crowd. It is just a question as to when, and who with the requisite expertise, will make this knowledge available to our motley crew.
A switchmode PSU can also be made very simply and cheap. The ferrite needed for 1000 VA cost around a $ or two. And i think any DIY'er can wind the 20 - 30 windings required. I think maybe 20$ will build a complete 500 VA supply for audio use.
However there is the safety issue. Who would dare post designs at the risk of some unexperienced DIY'er getting in harms way of high voltages... ?
However there is the safety issue. Who would dare post designs at the risk of some unexperienced DIY'er getting in harms way of high voltages... ?
John: I think you make a good point, but it opens up to another issue, why have DIY audio at all?
You can buy audio stuff cheaper than you can buy parts and build it anyway. so why not just close DIYAUDIO.com alltogether?
Even though i know it is highly controversial to challenge this pragmatic sense of logic, i will nevertheless do exactly that.
I consider and have always considered audio design not to be an engineering discipline, but rather an artform requireing some engineering skills. Much like any other artform.
So what is the best possible amplifier? Not the one with the lowest THD and the highest Power and bandwidth!
Well it must be the one that brings you most pleasure all in all. Not only performance, but also the story behind it, if it has any interesting technical features, and if the topology is somehow enjoyable. Maybe other things too. Your amplifier project is interesting not because you need 50 cheap W for a new home cinema speaker, but because of it's intriguing beauty. And all you DIY'ers out there you are not really geeks with soldering irons, anymore than a painter is a geek with a paintbrush.
You are creative minds with an interest to explore the art of constructing sound gear. And that is the intriguing beauty of the whole DIY audio community....
You can buy audio stuff cheaper than you can buy parts and build it anyway. so why not just close DIYAUDIO.com alltogether?
Even though i know it is highly controversial to challenge this pragmatic sense of logic, i will nevertheless do exactly that.
I consider and have always considered audio design not to be an engineering discipline, but rather an artform requireing some engineering skills. Much like any other artform.
So what is the best possible amplifier? Not the one with the lowest THD and the highest Power and bandwidth!
Well it must be the one that brings you most pleasure all in all. Not only performance, but also the story behind it, if it has any interesting technical features, and if the topology is somehow enjoyable. Maybe other things too. Your amplifier project is interesting not because you need 50 cheap W for a new home cinema speaker, but because of it's intriguing beauty. And all you DIY'ers out there you are not really geeks with soldering irons, anymore than a painter is a geek with a paintbrush.
You are creative minds with an interest to explore the art of constructing sound gear. And that is the intriguing beauty of the whole DIY audio community....
Lars,
This philosophy discussion is great. It shows that there is much more to good sonics than merely a schematic, and that moreover the primary cost lies with the transformer and power supply.
If one goes for two power supplies, one for each channel, to achieve the finest possible separation and hence imaging performance, then the cost skyrockets.
Now, back to T8:
Yes, but, a couple of questions:
1. Does it mean we can eliminate the Vbe multiplier and replace with resistors?
2. Will the extra stage introduce more phase shift, bringing the pole frequency lower?
3. Can we reduce lag compensation with T8?
I agree running T8 at constant current would be a grand idea, almost eliminating the intrinsic emitter follower distortions.
Finally, Lars, how much feedback does this circuit use? Any idea?
Cheers,
Hugh
This philosophy discussion is great. It shows that there is much more to good sonics than merely a schematic, and that moreover the primary cost lies with the transformer and power supply.
If one goes for two power supplies, one for each channel, to achieve the finest possible separation and hence imaging performance, then the cost skyrockets.
Now, back to T8:
Yes, but, a couple of questions:
1. Does it mean we can eliminate the Vbe multiplier and replace with resistors?
2. Will the extra stage introduce more phase shift, bringing the pole frequency lower?
3. Can we reduce lag compensation with T8?
I agree running T8 at constant current would be a grand idea, almost eliminating the intrinsic emitter follower distortions.
Finally, Lars, how much feedback does this circuit use? Any idea?
Cheers,
Hugh
Lars Clausen said:
Thanks for that equivalent of " I'd tell you, but... I'd have to kill you". Your selfless concern is very reassuring, and I do hope the tube discussion board takes this point to heart. I had assumed that anyone undertaking such an effort would follow standard operating procedures via the use of those ferrite transformers, optoisolators and startup powersupplies.... I'm very impressed that you can build a 500 VA audio use supply for 20$, but if so, why do your amplifier products use linear supplies? I had assumed that a 'proper' switching supply for audio use would require a quasi-resonant implementation, and hadn't realized it was so trivial.
Really? While I can purchase the 'bones' of an amp cheaper than I can purchase the parts and assemble, this, at best, provides a 'minimum level of mediocraty' that may be amenable to upgrades. And for those objectivists, I see no reason why they would desire anything beyond this. And while the buy 'pro' and mod might be the sweet spot for bang for the buck, it can be far from optimal as one attempts to climb that steepening cost/performance curve, ultimately leaving one polishing a turd. As someone who sells mods, in addition to your components, I had assumed you understood the rational (and economics) behind incremental mods.Lars Clausen said:
Lars Clausen said:
Golly, do I feel embarassed... I had cynically thought that your proposal for that ULCA might be an attempt to 'disadvantage' your competitors that offered lower cost alternatives to your own products, and that your tangential 'its cheaper to buy, why DIY?' pronouncement might be along those same lines, but now I realize that you are a pure soul, driven by Aesthetics.
PS - and what exactly is wrong about being a geek with a soldering iron? And certainly no really great artist (I actually know a number of them) could be considered a geek with a paintbrush, but far more appropriately, an insufferable, self indulgent jerk with a paintbrush. It does seem to come with the territory.
Re: Well stated, John........
Your last remark, is that nice? 
I think Lars named the thread wrongly. It should be called Ultra Easy Built Amp or something. With easy I mean no trimming and selecting. Just solder it together and everything work.john curl said:
Jocko, be positiveJocko Homo said:
Your last remark, is that nice? pmkap: 
i get your irony .. fun .. I was reading a book on the history of nuclear physics, and i got inspired by some discussions quoted between Niels Bohr and David Bohm, and found the analogy of artists to be applicable for the DIY audio community as well. (Sorry it was 4 a.m. here in Denmark ..
AKSA:
1..Of course not.
2..Yes, but compensated by the split frequency feedback network.
3..I don't understand this question.
Just see T8 as a almost distorsion free temperature and capacitance compensation for the VAS transistor.
i get your irony .. fun .. I was reading a book on the history of nuclear physics, and i got inspired by some discussions quoted between Niels Bohr and David Bohm, and found the analogy of artists to be applicable for the DIY audio community as well. (Sorry it was 4 a.m. here in Denmark .. AKSA:
1..Of course not.
2..Yes, but compensated by the split frequency feedback network.
3..I don't understand this question.
Just see T8 as a almost distorsion free temperature and capacitance compensation for the VAS transistor.
I don't understand the negativity appearing on this thread
. Isn't this thead just an attempt to put some fun back into this hobby?
ULC means it's possible to be creative, experiment, take chances and learn from real experience without the cost burden or the risk of destroying expensive components. This is DIY audio isn't it?
Lighten up
. Isn't this thead just an attempt to put some fun back into this hobby? ULC means it's possible to be creative, experiment, take chances and learn from real experience without the cost burden or the risk of destroying expensive components. This is DIY audio isn't it?
Lighten up
1. Then this is probably not a good reason. Temp compensation is readily achievable with a Vbe multiplier, suitably located, with a collector resistor.
2. OK, both caps are identified as C6 on the schemat; one is 1nF, phase lead, and the other is 2.2pF, pulling back gain at HF. 1nF seems quite high; I use an 18pF in this precise position, and it has considerable influence, making the amp tolerant of capacitive loads but greatly influencing sonics. I would have thought the use of T8 would enable considerable reduction in value, but the only way to assess this would be to build it and listen. What is your feeling about the sonics? Good? Excellent?
3. Fair call. You don't use lag compensation between collector and base of the VAS because of the split feedback scheme. But it could be tried, of course, and applied sparingly, might enable reduction of the 1nF, and improvement in sonics. Just an idea.
Lars, you have 100K supplying the input device base bias, and 22K in parallel with 1M for the feedback device base. Technically the resistance to ground for both bases should be the same for best offset trimming. Any reason for your choices?
I am not criticising; merely asking questions. I sense some resentment; if so, it's not my intention.
Cheers,
Hugh
Dear Hugh
No resentment mate!
1..It is not intended to temp. compensate the output stage, merely the VAS, so the whole circuit stays at the same DC point no matter the temp.
2..It also works as a fedforward compensation, keeping the amp stable in any capacitive load. You may be right about the value of the cap.
3..If you keep the feedback loop short, it should be unnessescary with a lag compensation.
4.. 100k on the input is chosen to keep the DC blocking cap at a low value, so anybody can buy a really good cap for no money.
22k is chosen as a compromise not to have a 4M7 in the feedback loop.
Australia is my favorite country! The vast widths of nature are stunning, i went there a couple of years ago with a friend, and still we talk about Australia almost every day .. Hugh you are a lucky guy! Cheers
No resentment mate!
1..It is not intended to temp. compensate the output stage, merely the VAS, so the whole circuit stays at the same DC point no matter the temp.
2..It also works as a fedforward compensation, keeping the amp stable in any capacitive load. You may be right about the value of the cap.
3..If you keep the feedback loop short, it should be unnessescary with a lag compensation.
4.. 100k on the input is chosen to keep the DC blocking cap at a low value, so anybody can buy a really good cap for no money.
22k is chosen as a compromise not to have a 4M7 in the feedback loop.
Australia is my favorite country! The vast widths of nature are stunning, i went there a couple of years ago with a friend, and still we talk about Australia almost every day .. Hugh you are a lucky guy! Cheers
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.