Home |
Search |
Today's Posts |
#41
|
|||
|
|||
|
#42
|
|||
|
|||
John Byrns wrote: In article .com, wrote: Thanks for taking the time to unconfuse me. I think I have it now. Question to be answered when you have built the KISSASS: Is it dollar for dollar better than the 38 year-old design? I hope the transformer on your KISSASS meets your expectations in overcoming Pinkerton's failures to meet the KISS guidelines which caused the near-universal rejection of Pinkerton's so-called KISASS Travesty. Remember that the power requirement is only 3W. I look forward to your report in the fullness of time when you have built the KISSASS. Hi Andre, I don't think I will be building my current "KISSASS" design because I don't believe it would have the desired tube sound. The problem is that it clips much too sharply and suddenly. Just below the clipping point the distortion is all second harmonic, but when pushed even slightly into clipping, with say 10% overdrive, the second harmonic disappears completely, and is replaced by a comb like spectrum of all the harmonics beyond the second. I believe this is caused by two factors, the first being that transistors naturally clip very suddenly, unlike tubes which have a soft clipping effect, and in addition my amplifier uses negative feedback which magnifies the harmonics above the second once clipping starts. Tube amps clip sharply even without NFB, and even triode amps. With music program, occasional clipping is not noticed by most people, but there is a point where it is all to easily noticed, and it makes little difference whether there is NFB or not. Operation of any audio amp should always be well short of clipping. Stewart's amplifier should be better in this regard as the negative feedback used in his amplifier is not as egregious as in my amplifier, but his amplifier probably still would suffer from the sharp clipping effect of transistors, the effect just wouldn't be magnified by the negative feedback as in my initial design. But Pinky's amp has an enormous amount of series voltage NFB, and it will definately clip sharply, like a Williamson, even with only 12 dB of NFB. They all damn well clip, and a triode amp can only make a couple of dB above onset of clipping before the sound turns to mud. To have a true solid state equivalent to the 300B amplifier I will have to first find a way to remove the negative feedback, which is not easily accomplished in my current design which depends on the negative feedback loop to establish the DC operating point for the transistors, and also to provide the required input resistance with only two transistors. People have been trying to get rid of NFB when using output transistors and they have never succeeded. A power bjt has a collector resistance to load resistance ratio which is as poor as a pentode anode Ra to RL. It is imperative to use NFB. Pinky has chosen around 70 dB of series voltage NFB in the form of emitter follower configuration. Nelson Pass uses the barest minimum of 12 dB of shunt NFB around one lone gain mosfet. He approximates a tube's outcome behaviour fairly well, but Pinky does not. It would probably be easy enough to retain negative feedback at DC to stabilize the operating point while eliminating the feedback in the audio band, but that would cause the input resistance to be too low. A third transistor would probably be needed to restore the input resistance to the desired value. The simplest and easiest poweramp uses a source follower pair of mosfets as a simple buffer with no gain. Then you can have a preamp which makes 10 vrms, and you get 10 watts of class A into 8 ohms. But the original idea was to make something usable from a CD player with no pre-amp. OK, use a single gain transistor ahead of the two comp pair class A mosfets. Pinky said the transistors he chose would be ideal. MJL4281, MJL4302. http://www.ebv.com/en/products/mip_l...conductor.html The hfe stated for these devices is 80 to 240. Therfore, with EF config, and assuming hfe might be 100, if you have 5 vrms into 8 ohms, collector current is 0.625 amps and base current is 0.00625 amps Rin = 5 / 0.00625 = 800 ohms. I recall Pinky said Rin would be higher, but lets say it was 800 ohms. You will need another transistor to buffer this 800 ohms, hence Pinky's gain transistor, and his input emitter follower. Its a 3 stage amp, not a two stage one, like the 300B amp. It might sound OK, and you must build it to find out. maybe its a nice amp in its own way; it just don't weigh much. Eliminating the feedback should bring my design into parity with Stewart's, but at that point both would still suffer from the sudden onset of clipping and the rich harmonic spectrum that produces. Forget what amps do when they clip. Its all bad news. What about designing for what they can do well before they clip? To fix that problem would require the development of a simple circuit to produce a soft clipping effect similar to tubes, which is a trickier problem, or I am sure Stewart would have already incorporated such a circuit in his design. For so called smooth clipping, you need to build a compressor, with a diode in a feedback path, and then you rely on the non linear turn on character of the diode, and voila, a logarithmic amp, which makes a nice compressor. If you set up such an amp with speakers at the bottom of a swimming pool, it will sound like it should. Patrick Turner. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#43
|
|||
|
|||
|
#44
|
|||
|
|||
"Stimpy" wrote in message ... Corton rouge rather than Corton Rouge surely? Correct. with a lower-case "r" Does that make it any less drinkable, I wonder? :-)) Iain |
#45
|
|||
|
|||
On Mon, 21 Mar 2005 20:41:48 +1100, Patrick Turner
wrote: John Byrns wrote: In article .com, wrote: Thanks for taking the time to unconfuse me. I think I have it now. Question to be answered when you have built the KISSASS: Is it dollar for dollar better than the 38 year-old design? I hope the transformer on your KISSASS meets your expectations in overcoming Pinkerton's failures to meet the KISS guidelines which caused the near-universal rejection of Pinkerton's so-called KISASS Travesty. Remember that the power requirement is only 3W. I look forward to your report in the fullness of time when you have built the KISSASS. Hi Andre, I don't think I will be building my current "KISSASS" design because I don't believe it would have the desired tube sound. The problem is that it clips much too sharply and suddenly. Just below the clipping point the distortion is all second harmonic, but when pushed even slightly into clipping, with say 10% overdrive, the second harmonic disappears completely, and is replaced by a comb like spectrum of all the harmonics beyond the second. I believe this is caused by two factors, the first being that transistors naturally clip very suddenly, unlike tubes which have a soft clipping effect, and in addition my amplifier uses negative feedback which magnifies the harmonics above the second once clipping starts. Tube amps clip sharply even without NFB, and even triode amps. With music program, occasional clipping is not noticed by most people, but there is a point where it is all to easily noticed, and it makes little difference whether there is NFB or not. Operation of any audio amp should always be well short of clipping. Stewart's amplifier should be better in this regard as the negative feedback used in his amplifier is not as egregious as in my amplifier, but his amplifier probably still would suffer from the sharp clipping effect of transistors, the effect just wouldn't be magnified by the negative feedback as in my initial design. But Pinky's amp has an enormous amount of series voltage NFB, and it will definately clip sharply, like a Williamson, even with only 12 dB of NFB. No it won't, because 'soft clipping' takes place in the voltage gain stage. Indeed, if you increase R12,13 and 14 to drop the supply voltage to Tr2 to the same as that of the output stage, there's no clipping at all from the emitter followers, it's *all* coming from the voltage stage. People have been trying to get rid of NFB when using output transistors and they have never succeeded. Depends how you define NFB, doesn't it Pat? :-) A power bjt has a collector resistance to load resistance ratio which is as poor as a pentode anode Ra to RL. It is imperative to use NFB. Pinky has chosen around 70 dB of series voltage NFB in the form of emitter follower configuration. Nelson Pass uses the barest minimum of 12 dB of shunt NFB around one lone gain mosfet. He approximates a tube's outcome behaviour fairly well, but Pinky does not. It would probably be easy enough to retain negative feedback at DC to stabilize the operating point while eliminating the feedback in the audio band, but that would cause the input resistance to be too low. A third transistor would probably be needed to restore the input resistance to the desired value. The simplest and easiest poweramp uses a source follower pair of mosfets as a simple buffer with no gain. Then you can have a preamp which makes 10 vrms, and you get 10 watts of class A into 8 ohms. But the original idea was to make something usable from a CD player with no pre-amp. OK, use a single gain transistor ahead of the two comp pair class A mosfets. Pinky said the transistors he chose would be ideal. MJL4281, MJL4302. http://www.ebv.com/en/products/mip_l...conductor.html The hfe stated for these devices is 80 to 240. Therfore, with EF config, and assuming hfe might be 100, if you have 5 vrms into 8 ohms, collector current is 0.625 amps and base current is 0.00625 amps Rin = 5 / 0.00625 = 800 ohms. I recall Pinky said Rin would be higher, but lets say it was 800 ohms. I said it would be 1600, because at a Tj of 100C, hfe is typically 200 for these devices. You will need another transistor to buffer this 800 ohms, hence Pinky's gain transistor, and his input emitter follower. Depends on the CD player, but yes, that's why the input EF is there. Its a 3 stage amp, not a two stage one, like the 300B amp. Depends how you look at it - my output EFs are an impedance transformer, just like the OPT on the 300B amp. It might sound OK, and you must build it to find out. maybe its a nice amp in its own way; it just don't weigh much. Depends - you could cast the heatsinks from good thick copper! :-) Eliminating the feedback should bring my design into parity with Stewart's, but at that point both would still suffer from the sudden onset of clipping and the rich harmonic spectrum that produces. Forget what amps do when they clip. Its all bad news. Agreed, and I understood the flea-power KISS design rationale to be that 'the first watt is the most important'. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#46
|
|||
|
|||
John:
The premise of the KISS 300B is explicitly that it will never clip. It is a c4W amp designed for a horn speaker that will never in normal or abnormal service demand as much as a whole watt. The question of overload behaviour therefore does not arise. It seems to me that you are cutting a rod for your own back by assuming your transistor design must be good all the way to clipping, never mind behaving gracefully beyond. The question is what happens at 1W, at which point the determining factor in the comparison with the KISS 300B should be the quality of the harmonic spectrum. Comparison at 1W is a fair competition. It is not serendipidous that my amp is good right up to the design max of 3.8W but for the purposes of the competition it is irrelevant. Andre Jute John Byrns wrote: In article .com, wrote: Thanks for taking the time to unconfuse me. I think I have it now. Question to be answered when you have built the KISSASS: Is it dollar for dollar better than the 38 year-old design? I hope the transformer on your KISSASS meets your expectations in overcoming Pinkerton's failures to meet the KISS guidelines which caused the near-universal rejection of Pinkerton's so-called KISASS Travesty. Remember that the power requirement is only 3W. I look forward to your report in the fullness of time when you have built the KISSASS. Hi Andre, I don't think I will be building my current "KISSASS" design because I don't believe it would have the desired tube sound. The problem is that it clips much too sharply and suddenly. Just below the clipping point the distortion is all second harmonic, but when pushed even slightly into clipping, with say 10% overdrive, the second harmonic disappears completely, and is replaced by a comb like spectrum of all the harmonics beyond the second. I believe this is caused by two factors, the first being that transistors naturally clip very suddenly, unlike tubes which have a soft clipping effect, and in addition my amplifier uses negative feedback which magnifies the harmonics above the second once clipping starts. Stewart's amplifier should be better in this regard as the negative feedback used in his amplifier is not as egregious as in my amplifier, but his amplifier probably still would suffer from the sharp clipping effect of transistors, the effect just wouldn't be magnified by the negative feedback as in my initial design. To have a true solid state equivalent to the 300B amplifier I will have to first find a way to remove the negative feedback, which is not easily accomplished in my current design which depends on the negative feedback loop to establish the DC operating point for the transistors, and also to provide the required input resistance with only two transistors. It would probably be easy enough to retain negative feedback at DC to stabilize the operating point while eliminating the feedback in the audio band, but that would cause the input resistance to be too low. A third transistor would probably be needed to restore the input resistance to the desired value. Eliminating the feedback should bring my design into parity with Stewart's, but at that point both would still suffer from the sudden onset of clipping and the rich harmonic spectrum that produces. To fix that problem would require the development of a simple circuit to produce a soft clipping effect similar to tubes, which is a trickier problem, or I am sure Stewart would have already incorporated such a circuit in his design. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#47
|
|||
|
|||
Stewart Pinkerton wrote: On Mon, 21 Mar 2005 20:41:48 +1100, Patrick Turner wrote: John Byrns wrote: In article .com, wrote: Thanks for taking the time to unconfuse me. I think I have it now. Question to be answered when you have built the KISSASS: Is it dollar for dollar better than the 38 year-old design? I hope the transformer on your KISSASS meets your expectations in overcoming Pinkerton's failures to meet the KISS guidelines which caused the near-universal rejection of Pinkerton's so-called KISASS Travesty. Remember that the power requirement is only 3W. I look forward to your report in the fullness of time when you have built the KISSASS. Hi Andre, I don't think I will be building my current "KISSASS" design because I don't believe it would have the desired tube sound. The problem is that it clips much too sharply and suddenly. Just below the clipping point the distortion is all second harmonic, but when pushed even slightly into clipping, with say 10% overdrive, the second harmonic disappears completely, and is replaced by a comb like spectrum of all the harmonics beyond the second. I believe this is caused by two factors, the first being that transistors naturally clip very suddenly, unlike tubes which have a soft clipping effect, and in addition my amplifier uses negative feedback which magnifies the harmonics above the second once clipping starts. Tube amps clip sharply even without NFB, and even triode amps. With music program, occasional clipping is not noticed by most people, but there is a point where it is all to easily noticed, and it makes little difference whether there is NFB or not. Operation of any audio amp should always be well short of clipping. Stewart's amplifier should be better in this regard as the negative feedback used in his amplifier is not as egregious as in my amplifier, but his amplifier probably still would suffer from the sharp clipping effect of transistors, the effect just wouldn't be magnified by the negative feedback as in my initial design. But Pinky's amp has an enormous amount of series voltage NFB, and it will definately clip sharply, like a Williamson, even with only 12 dB of NFB. No it won't, because 'soft clipping' takes place in the voltage gain stage. Indeed, if you increase R12,13 and 14 to drop the supply voltage to Tr2 to the same as that of the output stage, there's no clipping at all from the emitter followers, it's *all* coming from the voltage stage. Nad did "soft clip" amps about 20 years ago, and lord knows why. People have been trying to get rid of NFB when using output transistors and they have never succeeded. Depends how you define NFB, doesn't it Pat? :-) Oh, you wanting to define so you can say you ain't using any because its only emitter follower FB? Series voltage NFB is series voltage NFB, and your schematic shows you have a shirtload of it. It ain't global, but its local, and just there. And you have another shirtload in the input buffer follower, And you have third shirtful of series current NFB in the driver gain stage, without which your amp would sound abominable. You have 3 NFB stooges in your amp, all big burly fellas, there to whip the transistors into linearity, and without these heavy dudes, your transistors are quite useless. A power bjt has a collector resistance to load resistance ratio which is as poor as a pentode anode Ra to RL. It is imperative to use NFB. Pinky has chosen around 70 dB of series voltage NFB in the form of emitter follower configuration. Nelson Pass uses the barest minimum of 12 dB of shunt NFB around one lone gain mosfet. He approximates a tube's outcome behaviour fairly well, but Pinky does not. It would probably be easy enough to retain negative feedback at DC to stabilize the operating point while eliminating the feedback in the audio band, but that would cause the input resistance to be too low. A third transistor would probably be needed to restore the input resistance to the desired value. The simplest and easiest poweramp uses a source follower pair of mosfets as a simple buffer with no gain. Then you can have a preamp which makes 10 vrms, and you get 10 watts of class A into 8 ohms. But the original idea was to make something usable from a CD player with no pre-amp. OK, use a single gain transistor ahead of the two comp pair class A mosfets. Pinky said the transistors he chose would be ideal. MJL4281, MJL4302. http://www.ebv.com/en/products/mip_l...conductor.html The hfe stated for these devices is 80 to 240. Therfore, with EF config, and assuming hfe might be 100, if you have 5 vrms into 8 ohms, collector current is 0.625 amps and base current is 0.00625 amps Rin = 5 / 0.00625 = 800 ohms. I recall Pinky said Rin would be higher, but lets say it was 800 ohms. I said it would be 1600, because at a Tj of 100C, hfe is typically 200 for these devices. Let's not split hairs. Allow for device variations. I'll believe hfe = 200 when i see it. Goorrd, 100C? I'd be worried..... You will need another transistor to buffer this 800 ohms, hence Pinky's gain transistor, and his input emitter follower. Depends on the CD player, but yes, that's why the input EF is there. Its a 3 stage amp, not a two stage one, like the 300B amp. Depends how you look at it - my output EFs are an impedance transformer, just like the OPT on the 300B amp. It might sound OK, and you must build it to find out. maybe its a nice amp in its own way; it just don't weigh much. Depends - you could cast the heatsinks from good thick copper! :-) I have zero intention of building your amp. Eliminating the feedback should bring my design into parity with Stewart's, but at that point both would still suffer from the sudden onset of clipping and the rich harmonic spectrum that produces. Forget what amps do when they clip. Its all bad news. Agreed, and I understood the flea-power KISS design rationale to be that 'the first watt is the most important'. -- Stewart Pinkerton | Music is Art - Audio is Engineering Good luck with your project. I don't see too many wanting to build the thing...... Even you don't seem keen to. Crazy, really. Patrick Turner. |
#48
|
|||
|
|||
On Wed, 23 Mar 2005 00:23:57 +1100, Patrick Turner
wrote: Series voltage NFB is series voltage NFB, and your schematic shows you have a shirtload of it. It ain't global, but its local, and just there. And you have another shirtload in the input buffer follower, And you have third shirtful of series current NFB in the driver gain stage, without which your amp would sound abominable. No more than you'll find in a triode valve with the same stage gain. You have 3 NFB stooges in your amp, all big burly fellas, there to whip the transistors into linearity, and without these heavy dudes, your transistors are quite useless. Similarly, without a big burly air-gapped OPT, a single-ended 300B is no darn use for audio. Different techiques for different technologies. Pinky said the transistors he chose would be ideal. MJL4281, MJL4302. http://www.ebv.com/en/products/mip_l...conductor.html The hfe stated for these devices is 80 to 240. Therfore, with EF config, and assuming hfe might be 100, if you have 5 vrms into 8 ohms, collector current is 0.625 amps and base current is 0.00625 amps Rin = 5 / 0.00625 = 800 ohms. I recall Pinky said Rin would be higher, but lets say it was 800 ohms. I said it would be 1600, because at a Tj of 100C, hfe is typically 200 for these devices. Let's not split hairs. Allow for device variations. I'll believe hfe = 200 when i see it. Check the data sheets for these devices, that's the typical figure. Goorrd, 100C? I'd be worried..... Tj of 100C is not an issue for decent silicon, and is a pretty common operating point for power transistors used in class A amps. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#49
|
|||
|
|||
On Wed, 23 Mar 2005 00:23:57 +1100, Patrick Turner
wrote: Series voltage NFB is series voltage NFB, and your schematic shows you have a shirtload of it. It ain't global, but its local, and just there. And you have another shirtload in the input buffer follower, And you have third shirtful of series current NFB in the driver gain stage, without which your amp would sound abominable. No more than you'll find in a triode valve with the same stage gain. You have 3 NFB stooges in your amp, all big burly fellas, there to whip the transistors into linearity, and without these heavy dudes, your transistors are quite useless. Similarly, without a big burly air-gapped OPT, a single-ended 300B is no darn use for audio. Different techiques for different technologies. Pinky said the transistors he chose would be ideal. MJL4281, MJL4302. http://www.ebv.com/en/products/mip_l...conductor.html The hfe stated for these devices is 80 to 240. Therfore, with EF config, and assuming hfe might be 100, if you have 5 vrms into 8 ohms, collector current is 0.625 amps and base current is 0.00625 amps Rin = 5 / 0.00625 = 800 ohms. I recall Pinky said Rin would be higher, but lets say it was 800 ohms. I said it would be 1600, because at a Tj of 100C, hfe is typically 200 for these devices. Let's not split hairs. Allow for device variations. I'll believe hfe = 200 when i see it. Check the data sheets for these devices, that's the typical figure. Goorrd, 100C? I'd be worried..... Tj of 100C is not an issue for decent silicon, and is a pretty common operating point for power transistors used in class A amps. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#50
|
|||
|
|||
In article , Stewart Pinkerton
wrote: On Mon, 21 Mar 2005 20:41:48 +1100, Patrick Turner wrote: Its a 3 stage amp, not a two stage one, like the 300B amp. Depends how you look at it - my output EFs are an impedance transformer, just like the OPT on the 300B amp. "Just like"? You have a mighty strange way of "looking at it"! An emitter follower might change the impedance level in a circuit, but the comparison with a transformer ends there. A transformer changes impedance, but does not add any power, in fact it subtracts a small amount of power. Your emitter follower on the other hand not only changes the impedance level, but unlike a transformer, it also adds considerable power to the output, in the case of your amplifier it provides the vast majority of the output power from the amplifier. Your comparison of the EF to a Transformer is completely bogus, they are in no way similar. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#51
|
|||
|
|||
|
#52
|
|||
|
|||
Stewart Pinkerton wrote: On Wed, 23 Mar 2005 00:23:57 +1100, Patrick Turner wrote: Series voltage NFB is series voltage NFB, and your schematic shows you have a shirtload of it. It ain't global, but its local, and just there. And you have another shirtload in the input buffer follower, And you have third shirtful of series current NFB in the driver gain stage, without which your amp would sound abominable. No more than you'll find in a triode valve with the same stage gain. You don't know what I would find re the sound. But we note your reluctance to build, test, and supply tech data, thd/imd, etc. You have 3 NFB stooges in your amp, all big burly fellas, there to whip the transistors into linearity, and without these heavy dudes, your transistors are quite useless. Similarly, without a big burly air-gapped OPT, a single-ended 300B is no darn use for audio. Different techiques for different technologies. Utter garbage. Pinky said the transistors he chose would be ideal. MJL4281, MJL4302. http://www.ebv.com/en/products/mip_l...conductor.html The hfe stated for these devices is 80 to 240. Therfore, with EF config, and assuming hfe might be 100, if you have 5 vrms into 8 ohms, collector current is 0.625 amps and base current is 0.00625 amps Rin = 5 / 0.00625 = 800 ohms. I recall Pinky said Rin would be higher, but lets say it was 800 ohms. I said it would be 1600, because at a Tj of 100C, hfe is typically 200 for these devices. Let's not split hairs. Allow for device variations. I'll believe hfe = 200 when i see it. Check the data sheets for these devices, that's the typical figure. Goorrd, 100C? I'd be worried..... Tj of 100C is not an issue for decent silicon, and is a pretty common operating point for power transistors used in class A amps. I worry if anything goes over 50C, unless its designed to run happily at 200C, external temp, like a tube, which runs happily at 900 at the cathode. 100C for a transistor is too hot for me. If you only want 5 watts, You should have no more than 15 watts Pd per device. two devices allow up to 30 watts; Anyway, no need to really have the chips running at 100C at all. And why try to make the transistors run hot just so the hfe can be high? And if you had the buffer where it should be, between the input gain stage and output bases, your design wouldn't need to have an input buffer because the input stage will be high enough Rin. A gain pot and CD source is quite well naturally buffered by such an input stage. The buffer input stage is utterly superflous, and erroneous, if you want to keep it simple. The input gain stage can then use higher R values for Rc and Re, and the applied mountain of local current NFB will be all the more effective since the open loop gain rises with collector RL. The soft clipping in the output stage is no great option. The whole drive amp is bit "spongy", ie, the output current tries to modulate the input current to a small but rotten amount. Get real, build the SS amp they way it should be done, and forget all about trying to be a hero and outsmart a simple 300B design. But be a man, and like us, solder something together you can be proud of, and stop bitching about crummy 300B amps. We already know 300B is OK, and any crap you spout won't change anyone. What we dunno is whether your concoction is going to sound any good. Nor do you because you ain't made the damn thing. So until you build it, and test it, and gain some peer reviews at least, quit crowing like a silly rooster about the performance of some bloomin thing that has not actually been built. If you want respect around these parts, where most people think you just don't belong, do some bloomin work. Patrick Turner -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#53
|
|||
|
|||
"Patrick Turner" wrote in message ... (Abvout the KISASS amp) But we note your reluctance to build, test, and supply tech data, thd/imd, etc. I was quite interested to build it, until I noticed that the designer had not even bothered to do so. That doesn't show much faith in his design:-) Come on Stewart, get the soldering iron going. Build this thing and post some test results. Get real, build the SS amp they way it should be done, and forget all about trying to be a hero and outsmart a simple 300B design. But be a man, and like us, solder something together you can be proud of, and stop bitching about crummy 300B amps. We already know 300B is OK, and any crap you spout won't change anyone. What we dunno is whether your concoction is going to sound any good. Nor do you because you ain't made the damn thing. So until you build it, and test it, and gain some peer reviews at least, quit crowing like a silly rooster about the performance of some bloomin thing that has not actually been built. If you want respect around these parts, where most people think you just don't belong, do some bloomin work. I took a look at Stewart's pics of his listening room. I was impressed by the equipment. Lots of solid engineering there, but all other people's engineering. Nothing (with the possible exception of a RIAA stage) of his own. Compare with most people on RAT. The difference speaks for itself. Iain |
#54
|
|||
|
|||
On Wed, 23 Mar 2005 13:35:35 +1100, Patrick Turner
wrote: Tj of 100C is not an issue for decent silicon, and is a pretty common operating point for power transistors used in class A amps. I worry if anything goes over 50C, unless its designed to run happily at 200C, external temp, like a tube, which runs happily at 900 at the cathode. 100C for a transistor is too hot for me. A *junction* temperature of 50C is virtually unachievable with any class A amp, and is utterly pointless. Note that these devices have a junction/case resistance of 2.5C/watt, so that they'll be running above 50C with a heatsink the size of Greenland! If you only want 5 watts, You should have no more than 15 watts Pd per device. two devices allow up to 30 watts; Anyway, no need to really have the chips running at 100C at all. And why try to make the transistors run hot just so the hfe can be high? Just do the math, Patrick................. And if you had the buffer where it should be, between the input gain stage and output bases, your design wouldn't need to have an input buffer because the input stage will be high enough Rin. The output buffer *is* where it should be, for the amp to work as intended. Get real, build the SS amp they way it should be done, and forget all about trying to be a hero and outsmart a simple 300B design. You just don't get it, do you? If you want to build an *optimum* 'KISASS', just use the Linsley Hood design from 1969. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#55
|
|||
|
|||
On Wed, 23 Mar 2005 08:32:10 +0200, "Iain M Churches"
wrote: "Patrick Turner" wrote in message ... (Abvout the KISASS amp) But we note your reluctance to build, test, and supply tech data, thd/imd, etc. I was quite interested to build it, until I noticed that the designer had not even bothered to do so. That doesn't show much faith in his design:-) Come on Stewart, get the soldering iron going. Build this thing and post some test results. I've already stipulated that this isn't a *good* amplifier, just similar to KISS in its fundamental simplicity and avoidance of loop NFB. If you want to build a *good* small SS amp, use the Linsley Hood design from 1969. I took a look at Stewart's pics of his listening room. I was impressed by the equipment. Lots of solid engineering there, but all other people's engineering. Nothing (with the possible exception of a RIAA stage) of his own. Untrue, the passive controller and cabling are also homebrewed. However, the Krell cost me about the same as I would have paid for the parts, so why bother? The existence of a commercial SS amplifier is the only difference between my gear and that of the RATers - note that the UKRA website also contains illustrations of the gear of several RATers - and you won't find *any* of their own designs. Building a WAD kit doesn't count. Compare with most people on RAT. The difference speaks for itself. Oh, you guys build large planar speakers, FM tuners, CD players and record decks, do you? I've paid my dues as a hobbyist in the past, even built my own speakers, but with age comes wisdom........ You love to sneer, Churches, but you are a whited sepulchre. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#56
|
|||
|
|||
Stewart Pinkerton wrote: On Wed, 23 Mar 2005 08:32:10 +0200, "Iain M Churches" wrote: "Patrick Turner" wrote in message ... (Abvout the KISASS amp) But we note your reluctance to build, test, and supply tech data, thd/imd, etc. I was quite interested to build it, until I noticed that the designer had not even bothered to do so. That doesn't show much faith in his design:-) Come on Stewart, get the soldering iron going. Build this thing and post some test results. I've already stipulated that this isn't a *good* amplifier, just similar to KISS in its fundamental simplicity and avoidance of loop NFB. If you want to build a *good* small SS amp, use the Linsley Hood design from 1969. Well if your design is not a *good* amplifier, why the **** do you promote it here on a group devoted to tubecraft???? You are very immature about this whole simple SS amp BS. BTW, Linsley Hood's SS class A design is no big deal. I took a look at Stewart's pics of his listening room. I was impressed by the equipment. Lots of solid engineering there, but all other people's engineering. Nothing (with the possible exception of a RIAA stage) of his own. Untrue, the passive controller and cabling are also homebrewed. However, the Krell cost me about the same as I would have paid for the parts, so why bother? The existence of a commercial SS amplifier is the only difference between my gear and that of the RATers - note that the UKRA website also contains illustrations of the gear of several RATers - and you won't find *any* of their own designs. Building a WAD kit doesn't count. Compare with most people on RAT. The difference speaks for itself. Oh, you guys build large planar speakers, FM tuners, CD players and record decks, do you? I've paid my dues as a hobbyist in the past, even built my own speakers, but with age comes wisdom........ Wisdom? from Pinky? shish........................ You love to sneer, Churches, but you are a whited sepulchre. -- Yes but Pinky's whole attitude is one that sneers at ppl using tubes. Patrick Turner. Stewart Pinkerton | Music is Art - Audio is Engineering |
#57
|
|||
|
|||
"Stewart Pinkerton" wrote in message ... You love to sneer, Churches, but you are a whited sepulchre. You promised to *entertain* us (your words) Can't you do better than that Oh Master of Wit and Repartee? Don't get me wrong, I am trying so hard to take you seriously:-)) But I just can't fathom out why such a gifted precision analogue engineer doesn't design or build, and works 9-5 in the post room of a bank. But, hey, who cares? Just as long as I don't have to do it:-)) Iain |
#58
|
|||
|
|||
On Thu, 24 Mar 2005 00:14:55 +1100, Patrick Turner
wrote: Stewart Pinkerton wrote: On Wed, 23 Mar 2005 08:32:10 +0200, "Iain M Churches" wrote: "Patrick Turner" wrote in message ... About KISASS I've already stipulated that this isn't a *good* amplifier, just similar to KISS in its fundamental simplicity and avoidance of loop NFB. If you want to build a *good* small SS amp, use the Linsley Hood design from 1969. Well if your design is not a *good* amplifier, why the **** do you promote it here on a group devoted to tubecraft???? Because it's simulating *another* bad amplifier - a single-ended 300B. Besides which, ****wit, I was *invited* to submit such a design. You are very immature about this whole simple SS amp BS. You are very immature and tubecentric about the appropriate application of technology. BTW, Linsley Hood's SS class A design is no big deal. It is the best-known example of a simple but high-quality low-powered SS amp, and it is particularly relevant here because it was *specifically* designed to be a sonic match for the legendary Willaimson valve amplifier. I took a look at Stewart's pics of his listening room. I was impressed by the equipment. Lots of solid engineering there, but all other people's engineering. Nothing (with the possible exception of a RIAA stage) of his own. Untrue, the passive controller and cabling are also homebrewed. However, the Krell cost me about the same as I would have paid for the parts, so why bother? The existence of a commercial SS amplifier is the only difference between my gear and that of the RATers - note that the UKRA website also contains illustrations of the gear of several RATers - and you won't find *any* of their own designs. Building a WAD kit doesn't count. Compare with most people on RAT. The difference speaks for itself. Oh, you guys build large planar speakers, FM tuners, CD players and record decks, do you? I've paid my dues as a hobbyist in the past, even built my own speakers, but with age comes wisdom........ Wisdom? from Pinky? shish........................ Pearls before swine, it would seem..... :-) You love to sneer, Churches, but you are a whited sepulchre. Yes but Pinky's whole attitude is one that sneers at ppl using tubes. Only if they make bull**** claims about tubes. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#59
|
|||
|
|||
"Stewart Pinkerton" wrote in message ... : On Wed, 23 Mar 2005 13:35:35 +1100, Patrick Turner : wrote: : : : Tj of 100C is not an issue for decent silicon, and is a pretty common : operating point for power transistors used in class A amps. : : I worry if anything goes over 50C, unless its designed to run happily : at 200C, external temp, like a tube, which runs happily at 900 : at the cathode. : 100C for a transistor is too hot for me. : : A *junction* temperature of 50C is virtually unachievable with any : class A amp, and is utterly pointless. Note that these devices have a : junction/case resistance of 2.5C/watt, so that they'll be running : above 50C with a heatsink the size of Greenland! : : : Get real, build the SS amp they way it should be done, and forget all about : trying to be a hero and outsmart a simple 300B design. : : You just don't get it, do you? If you want to build an *optimum* : 'KISASS', just use the Linsley Hood design from 1969. : : -- : : Stewart Pinkerton | Music is Art - Audio is Engineering You just don't get it, do you? The argument being originally that a minimal parts count amplifier with a 300B could be *matched or surpassed* by a SS design of similar design is by your own statement above a hopeless task for the SS design: "If you want to build an *optimum* 'KISASS', just use the Linsley Hood design from 1969" iow no match, eh ? Rudy |
#60
|
|||
|
|||
Iain M Churches wrote: "Stewart Pinkerton" wrote in message ... You love to sneer, Churches, but you are a whited sepulchre. You promised to *entertain* us (your words) Can't you do better than that Oh Master of Wit and Repartee? Don't get me wrong, I am trying so hard to take you seriously:-)) But I just can't fathom out why such a gifted precision analogue engineer doesn't design or build, and works 9-5 in the post room of a bank. Maybe he likes getting his digits near the money. Its easier than stuffing transistors up his analog. Patrick Turner. But, hey, who cares? Just as long as I don't have to do it:-)) Iain |
#61
|
|||
|
|||
Stewart Pinkerton wrote: On Thu, 24 Mar 2005 00:14:55 +1100, Patrick Turner wrote: Stewart Pinkerton wrote: On Wed, 23 Mar 2005 08:32:10 +0200, "Iain M Churches" wrote: "Patrick Turner" wrote in message ... About KISASS I've already stipulated that this isn't a *good* amplifier, just similar to KISS in its fundamental simplicity and avoidance of loop NFB. If you want to build a *good* small SS amp, use the Linsley Hood design from 1969. Well if your design is not a *good* amplifier, why the **** do you promote it here on a group devoted to tubecraft???? Because it's simulating *another* bad amplifier - a single-ended 300B. Besides which, ****wit, I was *invited* to submit such a design. PPL thought you would submit a simple SS design and build and test a sample, and earn recognition for it. But all you've done is create a solid state parody of what is possible, a mocking gesture of a tube amp, and your achievement is entirely negligible so far. And the design is rather fruity, not an optimal use of transistors, as any engineer would see. You are very immature about this whole simple SS amp BS. You are very immature and tubecentric about the appropriate application of technology. I know how to use both appropriately, and constantly build fine working examples using SS devices or tubes. One project this week is to rebuild a James B Lansing amp from around 1960, complete with Ge transistors and input transformers. The Ge transistors and emitter follower class B output stage without loop NFB will be completely revised along the lines of what Douglas Self promotes. Its a really crummy old amp, but has a nice box and layout, so the cosmetics will be kept in honour of Mr JBL, who gave the amp to my customer. The amp will use modern silicon SS devices, and a decent amount of loop FB, and there will be no more random failures of old bits and peices in the old amp. Absolutely no ideas invented by one Mr S Pinkerton will be incorporated. You build nothing, and your game is mere cheap jest. IMHO, its a silly waste of time trying to design a BJT amp electronically as simple as any tube amp, and to mimic the highish Ro of a tube amp, although you are free to do it. It will *never* be as simple as a Pass mosfet design, nor as simple as a two stage SE tube amp. And its doubtful if it will sound as good, but I leave *you* to do the hard yards and find out. Most SS amps sound their best with loop NFB, since without NFB, SS amps are so often such poor performers. Patrick Turner BTW, Linsley Hood's SS class A design is no big deal. It is the best-known example of a simple but high-quality low-powered SS amp, and it is particularly relevant here because it was *specifically* designed to be a sonic match for the legendary Willaimson valve amplifier. Its not the best SS class A design. And only such low power. Sure the designer is well known, and sure he wanted a match for the Williamson, but so what? The Williamson is a good tube design, but not necessarily the best. Patrick Turner. I took a look at Stewart's pics of his listening room. I was impressed by the equipment. Lots of solid engineering there, but all other people's engineering. Nothing (with the possible exception of a RIAA stage) of his own. Untrue, the passive controller and cabling are also homebrewed. However, the Krell cost me about the same as I would have paid for the parts, so why bother? The existence of a commercial SS amplifier is the only difference between my gear and that of the RATers - note that the UKRA website also contains illustrations of the gear of several RATers - and you won't find *any* of their own designs. Building a WAD kit doesn't count. Compare with most people on RAT. The difference speaks for itself. Oh, you guys build large planar speakers, FM tuners, CD players and record decks, do you? I've paid my dues as a hobbyist in the past, even built my own speakers, but with age comes wisdom........ Wisdom? from Pinky? shish........................ Pearls before swine, it would seem..... :-) You love to sneer, Churches, but you are a whited sepulchre. Yes but Pinky's whole attitude is one that sneers at ppl using tubes. Only if they make bull**** claims about tubes. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#62
|
|||
|
|||
On Wed, 23 Mar 2005 23:46:56 +0100, "Ruud Broens"
wrote: "Stewart Pinkerton" wrote in message .. . : On Wed, 23 Mar 2005 13:35:35 +1100, Patrick Turner : wrote: : : : Tj of 100C is not an issue for decent silicon, and is a pretty common : operating point for power transistors used in class A amps. : : I worry if anything goes over 50C, unless its designed to run happily : at 200C, external temp, like a tube, which runs happily at 900 : at the cathode. : 100C for a transistor is too hot for me. : : A *junction* temperature of 50C is virtually unachievable with any : class A amp, and is utterly pointless. Note that these devices have a : junction/case resistance of 2.5C/watt, so that they'll be running : above 50C with a heatsink the size of Greenland! : : : Get real, build the SS amp they way it should be done, and forget all about : trying to be a hero and outsmart a simple 300B design. : : You just don't get it, do you? If you want to build an *optimum* : 'KISASS', just use the Linsley Hood design from 1969. You just don't get it, do you? The argument being originally that a minimal parts count amplifier with a 300B could be *matched or surpassed* by a SS design of similar design is by your own statement above a hopeless task for the SS design: "If you want to build an *optimum* 'KISASS', just use the Linsley Hood design from 1969" iow no match, eh ? You just don't get it, do you? While KISASS may certainly meet or exceed (in fact it's intended to simulate, not 'exceed') the performance of KISS, it is *not* an optimum simple SS design, being crippled IMHO by the pre-requisites of no loop feedback, and a design targetted at having a 'single-ended' transfer function. IOW, it's easily possible to *vastly* exceed the performance of KISS, but only by doing things very differently, which was *not* the point of the exercise. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#63
|
|||
|
|||
On Thu, 24 Mar 2005 10:09:17 +1100, Patrick Turner
wrote: Iain M Churches wrote: "Stewart Pinkerton" wrote in message ... You love to sneer, Churches, but you are a whited sepulchre. You promised to *entertain* us (your words) Can't you do better than that Oh Master of Wit and Repartee? Don't get me wrong, I am trying so hard to take you seriously:-)) But I just can't fathom out why such a gifted precision analogue engineer doesn't design or build, and works 9-5 in the post room of a bank. Maybe he likes getting his digits near the money. Its easier than stuffing transistors up his analog. Got it in one. I enjoy my work, which pays well in comparison with hardware engineering, and I enjoy my hobbies. IME, best not to mix the two. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#64
|
|||
|
|||
On Thu, 24 Mar 2005 10:35:15 +1100, Patrick Turner
wrote: Stewart Pinkerton wrote: On Thu, 24 Mar 2005 00:14:55 +1100, Patrick Turner wrote: Stewart Pinkerton wrote: On Wed, 23 Mar 2005 08:32:10 +0200, "Iain M Churches" wrote: "Patrick Turner" wrote in message ... About KISASS I've already stipulated that this isn't a *good* amplifier, just similar to KISS in its fundamental simplicity and avoidance of loop NFB. If you want to build a *good* small SS amp, use the Linsley Hood design from 1969. Well if your design is not a *good* amplifier, why the **** do you promote it here on a group devoted to tubecraft???? Because it's simulating *another* bad amplifier - a single-ended 300B. Besides which, ****wit, I was *invited* to submit such a design. PPL thought you would submit a simple SS design and build and test a sample, and earn recognition for it. Did they indeed? Then why did they insist that it be single-ended, and not have loop feedback? But all you've done is create a solid state parody of what is possible, a mocking gesture of a tube amp, Thanks - that was exactly what was intended. and your achievement is entirely negligible so far. In your humble opinionj, of course........... And the design is rather fruity, not an optimal use of transistors, as any engineer would see. Not argued at all, as it was never *meant* to be an optimal use of transistors, merely a simulation of a single-ended 300B. You are very immature about this whole simple SS amp BS. You are very immature and tubecentric about the appropriate application of technology. I know how to use both appropriately, and constantly build fine working examples using SS devices or tubes. Yes, of course you do, dearie. One project this week is to rebuild a James B Lansing amp from around 1960, complete with Ge transistors and input transformers. The Ge transistors and emitter follower class B output stage without loop NFB will be completely revised along the lines of what Douglas Self promotes. Its a really crummy old amp, but has a nice box and layout, so the cosmetics will be kept in honour of Mr JBL, who gave the amp to my customer. The amp will use modern silicon SS devices, and a decent amount of loop FB, and there will be no more random failures of old bits and peices in the old amp. OK, you're going to put a modern amp into an old box. Whoop-de-doo. Absolutely no ideas invented by one Mr S Pinkerton will be incorporated. Well, if you wanted a decent phono amp in there................ :-) You build nothing, and your game is mere cheap jest. The lady doth protest too much, methinks. IMHO, its a silly waste of time trying to design a BJT amp electronically as simple as any tube amp, and to mimic the highish Ro of a tube amp, although you are free to do it. It will *never* be as simple as a Pass mosfet design, nor as simple as a two stage SE tube amp. Agreed, but it meets the requirements, despite your typical aussie whining. And its doubtful if it will sound as good, but I leave *you* to do the hard yards and find out. Already stated that I won't be building it. Most SS amps sound their best with loop NFB, since without NFB, SS amps are so often such poor performers. Never argued - appropriate techiniques should be used for each technology, and the 'single-ended zero feedback' thing that some tubies have going, is utter ********, as I'm sure you agree. BTW, Linsley Hood's SS class A design is no big deal. It is the best-known example of a simple but high-quality low-powered SS amp, and it is particularly relevant here because it was *specifically* designed to be a sonic match for the legendary Willaimson valve amplifier. Its not the best SS class A design. Sez who? And only such low power. Everything is designed within certain parameters, and the JLH is a 10-watter, produced at a time when most speakers had 90dB/w/m sensitivity. Did you have a point, or are you just whining again, because no one outside RAT ever heard of *you*? :-) Sure the designer is well known, and sure he wanted a match for the Williamson, but so what? The designer is well-known *because* of that ampl;ifier, which was a landmark design. The Williamson is a good tube design, but not necessarily the best. Sez who? -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#65
|
|||
|
|||
Stewart Pinkerton wrote: On Wed, 23 Mar 2005 23:46:56 +0100, "Ruud Broens" wrote: "Stewart Pinkerton" wrote in message .. . : On Wed, 23 Mar 2005 13:35:35 +1100, Patrick Turner : wrote: : : : Tj of 100C is not an issue for decent silicon, and is a pretty common : operating point for power transistors used in class A amps. : : I worry if anything goes over 50C, unless its designed to run happily : at 200C, external temp, like a tube, which runs happily at 900 : at the cathode. : 100C for a transistor is too hot for me. : : A *junction* temperature of 50C is virtually unachievable with any : class A amp, and is utterly pointless. Note that these devices have a : junction/case resistance of 2.5C/watt, so that they'll be running : above 50C with a heatsink the size of Greenland! : : : Get real, build the SS amp they way it should be done, and forget all about : trying to be a hero and outsmart a simple 300B design. : : You just don't get it, do you? If you want to build an *optimum* : 'KISASS', just use the Linsley Hood design from 1969. You just don't get it, do you? The argument being originally that a minimal parts count amplifier with a 300B could be *matched or surpassed* by a SS design of similar design is by your own statement above a hopeless task for the SS design: "If you want to build an *optimum* 'KISASS', just use the Linsley Hood design from 1969" iow no match, eh ? You just don't get it, do you? While KISASS may certainly meet or exceed (in fact it's intended to simulate, not 'exceed') the performance of KISS, it is *not* an optimum simple SS design, being crippled IMHO by the pre-requisites of no loop feedback, and a design targetted at having a 'single-ended' transfer function. IOW, it's easily possible to *vastly* exceed the performance of KISS, but only by doing things very differently, which was *not* the point of the exercise. I don't know about you, but exactly what is your point of posting a paper design to ppl at r.a.t while rubbishing any creativity with just a couple of tubes? Your design uses 70 dB of local follower NFB. I for one won't say that is an evil thing in itself, and I would also say SS devices need all the NFB that can be mustered, because without NFB at all they are unusable, unlistenable, and there *must* be a lotta NFB. There is no need to have emitter follower topology; and remember that your paper wonder is an inverting amp, something that would seriously offend a lot of ppl. If the load is in the collector circuit of the outputs, loop FB can be brought back easily to the emitter of an input device, thus setting the gain, amount of FB and the input impedance all in the one loop of NFB, rather than having 3 cascaded stages with 3 loops of FB as you propose. Its useless talking to you though, you already told us your'e stubborn. I saw 3 transistor class A amp which beats your idea for simplicity in a recent Audio Express article. You might think your idea is a little you beaut wonder, but any discriminating SS cognescenti would never agree. Patrick Turner. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#66
|
|||
|
|||
Stewart Pinkerton wrote: On Thu, 24 Mar 2005 10:35:15 +1100, Patrick Turner wrote: Stewart Pinkerton wrote: On Thu, 24 Mar 2005 00:14:55 +1100, Patrick Turner wrote: Stewart Pinkerton wrote: On Wed, 23 Mar 2005 08:32:10 +0200, "Iain M Churches" wrote: "Patrick Turner" wrote in message ... About KISASS I've already stipulated that this isn't a *good* amplifier, just similar to KISS in its fundamental simplicity and avoidance of loop NFB. If you want to build a *good* small SS amp, use the Linsley Hood design from 1969. Well if your design is not a *good* amplifier, why the **** do you promote it here on a group devoted to tubecraft???? Because it's simulating *another* bad amplifier - a single-ended 300B. Besides which, ****wit, I was *invited* to submit such a design. PPL thought you would submit a simple SS design and build and test a sample, and earn recognition for it. Did they indeed? Then why did they insist that it be single-ended, and not have loop feedback? You proposed at your outset upon your quest that NFB is verboten. We all laughed, because we know its impossible to build anything with SS, and not have shirtfulls and bucket fulls of NFB. But all you've done is create a solid state parody of what is possible, a mocking gesture of a tube amp, Thanks - that was exactly what was intended. Well if any one of us designed a tube amp to mimic the ways and whiles of SS, we'd be laughed at. And they are all laughing at you. and your achievement is entirely negligible so far. In your humble opinionj, of course........... You have nothing soldered together, just a paper proposition. And the design is rather fruity, not an optimal use of transistors, as any engineer would see. Not argued at all, as it was never *meant* to be an optimal use of transistors, merely a simulation of a single-ended 300B. I must build a motor launch one day to mimic/simulate a sail boat. But only if I was desperate to be the fool of the harbour. You are very immature about this whole simple SS amp BS. You are very immature and tubecentric about the appropriate application of technology. I know how to use both appropriately, and constantly build fine working examples using SS devices or tubes. Yes, of course you do, dearie. One project this week is to rebuild a James B Lansing amp from around 1960, complete with Ge transistors and input transformers. The Ge transistors and emitter follower class B output stage without loop NFB will be completely revised along the lines of what Douglas Self promotes. Its a really crummy old amp, but has a nice box and layout, so the cosmetics will be kept in honour of Mr JBL, who gave the amp to my customer. The amp will use modern silicon SS devices, and a decent amount of loop FB, and there will be no more random failures of old bits and peices in the old amp. OK, you're going to put a modern amp into an old box. Whoop-de-doo. But it will work very well technically, and I hope sound a lot better than a transformer driven quad of class B Ge transistors with no loop FB. Anything would be better than this JBL chamber of sonic horrors. And JBL himself would approve the upgrade; he moved with the times, and addopted SS with a passion asap. I don't think the 30 watt amps he made with SS that I have here were anywhere near as good as most 30 watt tube amps.. Absolutely no ideas invented by one Mr S Pinkerton will be incorporated. Well, if you wanted a decent phono amp in there................ :-) Well I wouldn't. You build nothing, and your game is mere cheap jest. The lady doth protest too much, methinks. IMHO, its a silly waste of time trying to design a BJT amp electronically as simple as any tube amp, and to mimic the highish Ro of a tube amp, although you are free to do it. It will *never* be as simple as a Pass mosfet design, nor as simple as a two stage SE tube amp. Agreed, but it meets the requirements, despite your typical aussie whining. But you are the man on a mission, and its a pointless one. And its doubtful if it will sound as good, but I leave *you* to do the hard yards and find out. Already stated that I won't be building it. And I doubt anyone else will either. Most SS amps sound their best with loop NFB, since without NFB, SS amps are so often such poor performers. Never argued - appropriate techiniques should be used for each technology, and the 'single-ended zero feedback' thing that some tubies have going, is utter ********, as I'm sure you agree. You see, there you go again, de-legitimizing a fine way to listen to audio. Basically, you are calling us *******s, and none are impressed by your rudeness. It just makes you look a fool. BTW, Linsley Hood's SS class A design is no big deal. It is the best-known example of a simple but high-quality low-powered SS amp, and it is particularly relevant here because it was *specifically* designed to be a sonic match for the legendary Willaimson valve amplifier. Its not the best SS class A design. Sez who? Plenty. And only such low power. Everything is designed within certain parameters, and the JLH is a 10-watter, produced at a time when most speakers had 90dB/w/m sensitivity. Did you have a point, or are you just whining again, because no one outside RAT ever heard of *you*? :-) I earn a living as an audio engineer, and none of my many clients bother with r.a.t ; most don't know r.a.t exists. The point is that you sould broaden your mind, and realize that your judgements of SET are seen as all BS, because we have all heard fine music using SET based amp systems. Its an ironic situation. Sure the designer is well known, and sure he wanted a match for the Williamson, but so what? The designer is well-known *because* of that ampl;ifier, which was a landmark design. But JLH already had a big profile at Wireless World, so it wasn't difficult for him to gild it further with a class A SS design. After 1960, all the electronics in the world was being reformed and tubes all replaced with transistors, so the new heroes of this reformation glowed with pride at their tube replacements. Many would say they threw out the baby with the bathwater. The Williamson is a good tube design, but not necessarily the best. Sez who? Plenty. Do your own homework. Patrick Turner. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#67
|
|||
|
|||
In article , Stewart Pinkerton
wrote: On Wed, 23 Mar 2005 13:35:35 +1100, Patrick Turner wrote: And if you had the buffer where it should be, between the input gain stage and output bases, your design wouldn't need to have an input buffer because the input stage will be high enough Rin. The output buffer *is* where it should be, for the amp to work as intended. Get real, build the SS amp they way it should be done, and forget all about trying to be a hero and outsmart a simple 300B design. You just don't get it, do you? If you want to build an *optimum* 'KISASS', just use the Linsley Hood design from 1969. How do you figure the 1969 JLH design is "optimum"? It looks to me like your "KISASS" design is at least as "optimum" as the JLH design. I would think a more optimum design would be a combination of the two designs, taking the best features of each. The JLH design is flawed because the basic circuit topology produces a high level of distortion, and it is only through the application of negative feedback that the distortion is reduced to tolerable levels. It would be better to take a design similar to yours, with lower open loop distortion, and then put loop feedback around that to achieve a really good SS design. I have been surfing the web these last couple of days trying to locate information on the 1969 JLH design, when I finally found the schematic I was amazed to see that the basic circuit of JLH's amplifier is essentially identical to the current source version of my "KISSASS" design. The difference being that my "KISSASS" design is a true single ended design, with Tr2 connected as a current source, or replaced by an output transformer, rather than being driven by the collector of Tr3. This leaves Tr3 & Tr1 connected as a simple Darlington pair, making it possible to delete Tr3 if the beta of the remaining transistors is sufficiently high. However I don't particularly care for the JLH/"KISSASS" approach because it depends on negative feedback to make the high inherent distortion more tolerable. A better idea would be to take a design like your "KISASS", with its lower inherent distortion, and then apply NFB to reduce the distortion to even lower levels. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#68
|
|||
|
|||
|
#69
|
|||
|
|||
On Thu, 24 Mar 2005 22:04:50 +1100, Patrick Turner
wrote: Sure the designer is well known, and sure he wanted a match for the Williamson, but so what? The designer is well-known *because* of that amplifier, which was a landmark design. But JLH already had a big profile at Wireless World, so it wasn't difficult for him to gild it further with a class A SS design. Utter ********. Show evidence of a 'big profile' prior to 1969. You're just a typical whining aussie who's jealous of a *real* innovator. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#70
|
|||
|
|||
John Byrns wrote: In article , Stewart Pinkerton wrote: On Wed, 23 Mar 2005 13:35:35 +1100, Patrick Turner wrote: And if you had the buffer where it should be, between the input gain stage and output bases, your design wouldn't need to have an input buffer because the input stage will be high enough Rin. The output buffer *is* where it should be, for the amp to work as intended. Get real, build the SS amp they way it should be done, and forget all about trying to be a hero and outsmart a simple 300B design. You just don't get it, do you? If you want to build an *optimum* 'KISASS', just use the Linsley Hood design from 1969. How do you figure the 1969 JLH design is "optimum"? It looks to me like your "KISASS" design is at least as "optimum" as the JLH design. I would think a more optimum design would be a combination of the two designs, taking the best features of each. The JLH design is flawed because the basic circuit topology produces a high level of distortion, and it is only through the application of negative feedback that the distortion is reduced to tolerable levels. It would be better to take a design similar to yours, with lower open loop distortion, and then put loop feedback around that to achieve a really good SS design. The open loop thd of pinky's amp is very high. But he uses 3 loops to reduce it, EF connection on the buffer input, current FB in the gain stage, and EF connected outputs, so a total of perhaps 150 dB of NFB is applied. Its no use tellin pinky that a single loop of NFB could have been used which would also render his amp being non-inverting, like all good amps are. Its impossible to use bjt with low amounts of NFB, so one may as well use high levels, and be done with it. The idea that a low amount of loop or other NFB will be OK with bjts isn't quite right, because bjt voltage transfer curves are so horrid to begin with. I have been surfing the web these last couple of days trying to locate information on the 1969 JLH design, when I finally found the schematic I was amazed to see that the basic circuit of JLH's amplifier is essentially identical to the current source version of my "KISSASS" design. The difference being that my "KISSASS" design is a true single ended design, with Tr2 connected as a current source, or replaced by an output transformer, rather than being driven by the collector of Tr3. This leaves Tr3 & Tr1 connected as a simple Darlington pair, making it possible to delete Tr3 if the beta of the remaining transistors is sufficiently high. However I don't particularly care for the JLH/"KISSASS" approach because it depends on negative feedback to make the high inherent distortion more tolerable. Don't be too perturbed by the NFB used by JLH. The alternative is an unlisteable and useless amplifier, since collector resistance is far higher than the load value, like a pentode's plate resistance. And distortion from collector circuits without the emitter follower NFB connection is truly appalling, although with class A this is limited to mild odd order at very low levels. A better idea would be to take a design like your "KISASS", with its lower inherent distortion, and then apply NFB to reduce the distortion to even lower levels. This is the practice now in place for nearly all of today's SS amps. They use pinky's idea of 70 dB of local EF NFB in the output stage, and then apply another 50 dB of global NFB. thd levels of 0.002% are routine at 200 watts. Its very hard to measure any thd at 3 watts, even with a low bias current output stage operating in nearly class B conditions. But with class A, the total open loop gain used in the circuit can be much reduced, since the use of a total of a typical 120 dB of applied NFB is totally unnecessary, since there is no switching distortions, and one really only needs a maximum total of applied NFB = 20 dB, just as one would with a pentode class A amp. But because the voltage gain with bjt is so extraordinary, and greater than tubes, since the gm of the bjt is so high, it is so easy to apply 120 dB of sereis voltage NFB rather than just 20 dB. Would a bjt amp with a total of 12 dB of NFB sound well? 12dB is the equivalent of the approximate 12 dB within a 300B. It is a question not immediately answerable. The objectivists would say that to make sure it won't sound as bad as a tube we'll use 120 dB of NFB, just to be sure. This opinion is as bad as a rotten fart because it negates the subjective experience of music, and negates all listening tests. The objectivists assume that measured thd between 0.05% and 0.5% is totally unacceptable. They don't listen. It seems pinky doesn't even want to build his own wretched design, and proove what he says is true, nor does he want to do the hard yards to dispassionately compare the measured and sonic results without rancour, and without stooping to a ****e smearing campaign. All roads built by pinky lead to a pointless smear campaign against tubecraft. Patrick Turner. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
#71
|
|||
|
|||
On Fri, 25 Mar 2005 17:08:54 +1100, Patrick Turner
wrote: The open loop thd of pinky's amp is very high. But he uses 3 loops to reduce it, EF connection on the buffer input, current FB in the gain stage, and EF connected outputs, so a total of perhaps 150 dB of NFB is applied. By what definition are these "loops"? I feel like I've stepped through a mirror on r.a.t. lately. Any word can mean anything one wishes it to mean. And the Red Queen said "Off with her head!" Is this in the same category of word useage as "internal feedback in triodes"? Chris Hornbeck |
#72
|
|||
|
|||
On Fri, 25 Mar 2005 17:08:54 +1100, Patrick Turner
wrote: pinky And, by the way, this is incredibly childish. Chris Hornbeck |
#73
|
|||
|
|||
"Patrick Turner" Its no use tellin pinky that a single loop of NFB could have been used which would also render his amp being non-inverting, like all good amps are. ** There is no fool like a self taught fool. The positive feedback is never ending. .............. Phil |
#74
|
|||
|
|||
Chris Hornbeck wrote: On Fri, 25 Mar 2005 17:08:54 +1100, Patrick Turner wrote: The open loop thd of pinky's amp is very high. But he uses 3 loops to reduce it, EF connection on the buffer input, current FB in the gain stage, and EF connected outputs, so a total of perhaps 150 dB of NFB is applied. By what definition are these "loops"? I feel like I've stepped through a mirror on r.a.t. lately. Loops of NFB exist even when the loops are local loops, such as using a follower connection, or unbypassed cathode or emitter resistor. Your text books explain it further. Any word can mean anything one wishes it to mean. And the Red Queen said "Off with her head!" I always invite a rehead home for dinner, and be very polite about it all. Is this in the same category of word useage as "internal feedback in triodes"? All feedback is in some type of loop, including the loop in a triode. You need to read the right books. Books were invented hundreds of years ago, and have been replaced by dazzling screens, with lots of lettered and numberd buttons in front. Both books and screens convey BS very well, as well as the truth. But much basic truth was all written in old books, before screens replaced them. Then people made mistakes when they tried to copy the wisdom from books to screens, mainly because they suddenly became terribly impatient...... Patrick Turner. Chris Hornbeck |
#75
|
|||
|
|||
Chris Hornbeck wrote: On Fri, 25 Mar 2005 17:08:54 +1100, Patrick Turner wrote: pinky And, by the way, this is incredibly childish. Chris Hornbeck Oh, should I have used the word "pinko" to nick name pinkerton? "Pinko" would imply pinky was a communist and I don't see any evidence just yet that he believes in communism, which to me means that what's your's is mine, and what's mine is also mine, if we were to examine the more brutal communist regimes. Maybe I should call him "red". He should be red in the face for the embarrassing stance he takes here at r.a.t. Patrick Turner. |
#76
|
|||
|
|||
Patrick Turner wrote: Chris Hornbeck wrote: On Fri, 25 Mar 2005 17:08:54 +1100, Patrick Turner wrote: pinky Oh, I don't think Patrick is being disrespectful. It's just a typo, not hitting the caps key. And, by the way, this is incredibly childish. What the devil are you going on about? Pinkerton is widely known as Pinky. I looked into it when he first stormed into RAT last year, not wanting to use a derogatory nickname without good reason. We've since been given more than sufficient reason. Chris Hornbeck Oh, should I have used the word "pinko" to nick name pinkerton? "Pinko" would imply pinky was a communist and I don't see any evidence just yet that he believes in communism, I named him Pinko for his entryist method, devised by Lenin, who surely was a communist, and widely practised by communists for over 80 years since the 1890s. They would smarm their way into some group, often saying they came only to learn. Once they were on the inside, their true nature would soon out. By then it would be too late for the original members of the group. The commies would have taken over and the natives would be in the gulag or kneeling on the rubber mat for the bullet in the back of the neck. which to me means that what's your's is mine, and what's mine is also mine, if we were to examine the more brutal communist regimes. Maybe I should call him "red". He should be red in the face for the embarrassing stance he takes here at r.a.t. Hear, hear. And not to forget social climbing in a pink coat, trying to buy his way into the mounted classes. Pinko Presumptuous indeed. Patrick Turner. Andre Jute |
#77
|
|||
|
|||
On Fri, 25 Mar 2005 17:08:54 +1100, Patrick Turner
wrote: The open loop thd of pinky's amp is very high. But he uses 3 loops to reduce it, EF connection on the buffer input, current FB in the gain stage, and EF connected outputs, so a total of perhaps 150 dB of NFB is applied. None of these are loops, as you well know, so your argument is bogus. Its no use tellin pinky that a single loop of NFB could have been used which would also render his amp being non-inverting, like all good amps are. It's no use tellin Turner that many classic hi-fi amps are inverting, the Quad 405 being but one example. And it's no use tellin Turner that zero loop feedback is fundamental to KISASS. -- Stewart Pinkerton | Music is Art - Audio is Engineering |
#78
|
|||
|
|||
"Stewart Pinkerton" It's no use tellin Turner that many classic hi-fi amps are inverting, the Quad 405 being but one example. ** The Quad 405 uses an op-amp gain stage that inverts the signal - while the "current dumping" power stage is non inverting with a voltage gain of 3.8 times. The early Bose 1800 also used an op-amp inverting buffer . The Quad 303 is an inverting amp. ............. Phil " Stewart Pinkerton | Massive Fart - All the rest is Bull**** " |
#79
|
|||
|
|||
Stewart Pinkerton wrote: On Fri, 25 Mar 2005 17:08:54 +1100, Patrick Turner wrote: The open loop thd of pinky's amp is very high. But he uses 3 loops to reduce it, EF connection on the buffer input, current FB in the gain stage, and EF connected outputs, so a total of perhaps 150 dB of NFB is applied. None of these are loops, as you well know, so your argument is bogus. The shallow state of your bogus understanding of electronic engineering is becoming all too apparent. All applications of FB, current, voltage, series, shunt, positive, negative; all involve a loop. The simplest and most widely understood is the follower connection where *all* of the output signal is fed back in series with the input voltage. The device might be an opamp, with a tiny multi bjt amplifier with a chip arranged like a simple single transistor emitter follower. Regardless of the device number, the series voltage feedback is still applied around a loop, albeit a short on in the case of a single bjt, or mosfet, or tube. Let us consider a power mosfet with a 3 ohm load operating in SE mode, but in source follower mode. Let us suppose the open loop gm = 1 amp per volt. This means that the 1 volt applied between gate and source produces 1 amp of load current so we would get 3volts in the load. The open loop gain, regardless of how the NFB is applied is 3 / 1 = 3. When we "close the feedback loop" when we apply FB, and in this case by choosing the source follower connection, we still need 1v applied between gate and source to give a 3v load change, so we must apply 4v to the gate to keep this "closed lood" condition. The closed loop gain reduction amount = open loop gain / closed loop gain. Open loop gain = 3, closed loop gain = 4 / 3 = 1.333 In this case gain reduction = 4 / 1.33 = 3.00. Amount of applied series voltage NFB = 20 log 3 = 9.54 dB of applied NFB. The example I give is also described in the text books. I hope the conventions of engineering terminology become crystal clear to all, for they allow us to all focus on the same level of understading, although I am not sure if pinky could. Now had one set up the mosfet with the load powered from the drain instead of the source, the load voltage is inverted with respect to the gate voltage, and since the source is grounded, there is no NFB able to be applied unless we have either a source resistance to get series current NFB, or we use a shunt resistance network as Nelson Pass uses in his Zen amp. But at least with the load in the drain circuit we can examine the single fet's performance without any NFB present, ie, examine the open loop gain, and that will be -3, but numerically the same as the source follower case above. That is because for -1v input, we get +3v output at the 3 ohm load. But we also will get perhaps 0.15v of distortion, which is 5%. Now when we connect the load in the source follower mode, that 0.15% thd still tries to appear at the load output, but it also exists between gate and source, but it is amplified in such a way to reduce its own production. It can't be completely cancelled away by NFB, but from out calculation of gain reduction above, we can also deduce that the reduction of "closed loop thd" will be the same as the gain reduction, ie, 3 times, or by 9.54 dB, so with source follower, the thd is 0.15v/3 = 0.05v = 1.66%, which is better than no reduction of thd at all. But of course if the thd was all 2H to start with in the open loop voltage, we'd see a small but significant amount of 3H in the output with the loop closed where there was none in the open loop to begin with. This is because the amount of NFB applied is low, only 9.5 dB, and amount of thd high, 5%. and the 3H appears due to the intermodulation process for which an explantion now is not part of my brief. Now it matters not what sort of amp you have, but if you did have open loop 2H = 5%, and you applied 9.5 dB of NFB, the same predictable and calculatable 1.66% of 2H would appear, regardless of whether you have an amp with an open loop gain of 6, like many tube amps, complete with 3 stages and an output tranny, and you reduced that 3 times to 2, or whether you have a mosfet in source follower mode where the open loop is also reduced 3 times. Pinky said he wanted to avoid the magic of the 300B. We will of course let him. But there is no technical magic in the follower connection. The follower connection obeys the same rules of physics as any other application of series voltage NFB application. In the case of his rather illconcieved SS amp, he has the output transistors in PP class A and in emitter follower with an open loop gain claimed to be around 3,000 which is reduce to just under 1.0 since it only takes 0.00033 volts applied between the base and emitter to produce 1v of rated value load change. 70 dB of series voltage NFB is being applied. Pinky says this somehow differs from loop FB, which in his mind means NFB applied around *more* than one device in cascade. But an engineer would not make such a distinction at all, although he would say that the stable bandwidth of a multi stage amp with 70 dB of series voltage NFB will be less because of the miller effects in the multistage amp. An engineer working with amps operating at GHz would have more to say about L and C effects. I hope it becomes clear to all that my argument is not bogus, as pinky suggests. If one uses a single tube for a cathode follower, and a gain reduction due to the FB was sat 10 times, or 20 dB, then it is the same as having a multi stage amp and applying 20 dB of NFB, as in a Williamson. However, the multi stage amp will have open loop distortions of all the stages, and the distortion of the early stages will be distorted by the following stages, and these artifacts are called second order artifacts, and are not present in a simple one device when connected as a follower. NFB applied around several stages is usually called global NFB, and usually it is series voltage NFB, because it allows an amp to be non inverting, ie, the output phase is the same as the input phase. With tube amps, there it is good practice to make the driver amp with as litle thd as possible, so that there are astonishingly low amounts of second order products produced when global NFB is applied around the several stages and including the OPT. Therefore such an amp acts very closely to a perfect voltage amp applied to a cathode follower output stage. CF tube output stages are rarely done because it means applying say 240 vrms to each output tube grid to get say 220 vrms at each end of the PP output tranny. With a bjt amp, we may only want 20 vrms at the output, and its not impossible to make just over 20v at low thd to apply to to an emitter follower base circuit. But with most bjt amps, this drive voltage is not produced very linearly unless another lot of NFB is applied globally, or in the case of pink's amp, locally in the form of a boot full of series current NFB. Its a short loop of NFB, but is still loop NFB. Its no use tellin pinky that a single loop of NFB could have been used which would also render his amp being non-inverting, like all good amps are. It's no use tellin Turner that many classic hi-fi amps are inverting, the Quad 405 being but one example. And it's no use tellin Turner that zero loop feedback is fundamental to KISASS. Most amps I have ever worked on are non inverting. Its not a problem to me, but to many folks an inverting power amp is just not right. I routinely use a single triode for a preamp, and of course the signal is inverted. The detectability of inverted signals isn't all that great. But if a source signal does have significant even order harmonics, then the phase of those harmonics may either cancel or reinforce those in the amp chain or the speakers. As I explained above, loop FB includes short loops. Global loop FB usually means a loop around several stages, such as in Williamson's amp, and the not so wonderful JLH replacement. As I recall, JLH didn't want to have to build another identical tube amp to make two channels for stereo, and took the lazy man's avenue, and went all solid state. It explains why there are so many old single tube monoblock amps around. Patrick Turner. |
#80
|
|||
|
|||
In message , Patrick Turner
writes The detectability of inverted signals isn't all that great. But if a source signal does have significant even order harmonics, then the phase of those harmonics may either cancel or reinforce those in the amp chain or the speakers. Dear God, how much more nonsense do we have to put up with from this man? I do not believe that even PT really believes this! -- Chris Morriss |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Easter approaches, whether Stewart Pinkerton? | Vacuum Tubes | |||
Lionel's Demonstration of His Insanity = His Delusional Attack Threads | Audio Opinions | |||
For John, definitely not the thread Once more into the breach, dear friends | Vacuum Tubes |