This is a copy of the sleeve of Satchmo Play King Oliver . Hope you can read it OK . Good reading and no less true today . If you never heard this it is way beyond today's typical recordings . Note the pride in microphones used . I was exactly 3 years old as it was recorded ( 1959 Sept Oct ) .
Although a long shot . Might I recommend the Swan Inn Islip Oxon . UK . I had my birthday meal there and was delighted by service and price ( last night ) .
Not going there ( mono) stereo works fine with my imagination, those with Mono imagination better than mine can revert to such ...
EDIT: Are we finished wringing out D's amp ..? ...
Just a word on the Shure M 44-7 . I discovered by accident how good it can be . We built a mastering turntable based on a Garrard 401 and the modern SME 309 . The arm is much like the Series 5 except with an excellent interchangeable cartridge mount . The Shure is the cartridge we could most easily get 78 styli for . Even the ball type of Pathe acoustic recordings that play on a standard turntable from the middle outwards . The ball type suits similar disc of the period of the conventional type . With the Shure came a 0.7 thou ( inch ) stereo/mono compatible stylus ( Spherical 0.5 = stereo 1 = mono , 0.7 fits all ) . I had a quick listen and was astonishing . Never does anyone do this . As Douglas Self pointed out only a high output MM ( 10 mV ) has the full dynamic range of about 69 dB . Due to noise and cable losses an MC has slightly less ( 62 dB if an SPU into transistors direct , that is without a transformer ) . The Shure sets problems of overload . Personally I prefer a passive/active design ( 75 uS passive , 3180/318 uS active ) . The old Leak Varislope with OPA 2604 buffer was what we used ( buffer was all transistor to start with ) . I could well believe the Shure is a superior device today as the production is in smaller numbers . I don't have the same affection for the V15 after mk2 . The M44 was tried in a Series 5 and almost made me love the arm . Usually it sounds a bit mechanical , all transient and little trailing edge . I don't doubt it is not the fault of the arm before anyone says . To be honest I detest most pick up arms . Naim Aro is very good and the Hadcock it is derived from . Old SME 12 inch is nice as are the new M series . The Jelco is beautifully made , sounds a bit tame . Origin Live arms I like . Why other chunky arms sound brash I do not know when compared with O L that also is chunky . The Graham I can not make my mind up about . Fidelity Research I like .
The M44/7 if wired series mono is 19 mV output . 40 dB more than some MC's .
D's amp . I want him to take bits out until it stops working then put that one bit back . That's what I do Dvv jokes . No joke , I do and very proud to say it . Protection circuit first . It's fuses and capacitors for me . Mr D caused a storm on the D Self thread . He was told how wrong he was by the clones of Mr S . Not so Mr D . The high bias you run mitigates any loop gain problems and shifts doubt to where loud enough to mask any crossover artifacts . Swings and roundabouts you win . Personally I would like to see the bias move to the 5 watt region or down to near zero as Quad 303 ( Bryston I guess ) . The latter needs as much open loop gain as possible . Anything else I suspect will be wrong if bipolar . 5 watts is too much and I am sure not a big deal difference . Dvv stated 1 watt which seems a good figure . High loop feedback amps do not benefit from high bias I suspect . Gain doubling is the stated reason . Equally FET's suffer high on resistance if over biased . This seems to say similar problem when in fact the 100 mA level is purely coincidental . FET's being very fast seem to use feedback better . Thrive would be the word . Never more so than when class D . To restate my objection to class D . Like class AB with more problems deliberately added .
The M44/7 if wired series mono is 19 mV output . 40 dB more than some MC's .
D's amp . I want him to take bits out until it stops working then put that one bit back . That's what I do Dvv jokes . No joke , I do and very proud to say it . Protection circuit first . It's fuses and capacitors for me . Mr D caused a storm on the D Self thread . He was told how wrong he was by the clones of Mr S . Not so Mr D . The high bias you run mitigates any loop gain problems and shifts doubt to where loud enough to mask any crossover artifacts . Swings and roundabouts you win . Personally I would like to see the bias move to the 5 watt region or down to near zero as Quad 303 ( Bryston I guess ) . The latter needs as much open loop gain as possible . Anything else I suspect will be wrong if bipolar . 5 watts is too much and I am sure not a big deal difference . Dvv stated 1 watt which seems a good figure . High loop feedback amps do not benefit from high bias I suspect . Gain doubling is the stated reason . Equally FET's suffer high on resistance if over biased . This seems to say similar problem when in fact the 100 mA level is purely coincidental . FET's being very fast seem to use feedback better . Thrive would be the word . Never more so than when class D . To restate my objection to class D . Like class AB with more problems deliberately added .
Last edited:
TRANSMISSION-LINE SPEAKERS
My dad tried to teach me about radar transmission-lines when about 5 years old . The ideas stuck although the connection with transmission-line speakers is tenuous . At least it made me alert to the idea . Although only five years old slotted waveguides stay in the mind .
I have never found a transmission-line speaker that delivered the performance expected . Thinking about this it should not be so . I suggest the problem is the mid-range would be better as an open baffle ( OB ) . Then the openness would be equal from low bass to mid treble . A ribbon top to complete things . Experiments with my friends excellent Roksan speakers shows that slightly blocking the port destroys the magic . This was his idea as he thinks the bass is slightly one note . I asked him to buy some drinking straws to make a resistive port . Rule of thumb is it changes resonance by shifting it up circa 10% whilst damping the Q . I said we could consider very tight sounding sub woofers that double as speaker stands . I would then crossover to the Roksan K2 above the critical frequency ( 100 Hz ? ) . The subs 190 x 280 x 600 perhaps ? The speakers remind me of the Red Rose Rose Bud at a fraction of the price . I was told the Rose Bud has it's origins in a Chinese design that is slightly upgraded ( selected drive units , better box ) . The price of the originals about $200 a pair . The Roksans are nice enough for me to pay retail .
Superfi - ROKSAN KANDY K2 SPEAKERS (PAIR)
http://www.redrosemusic.com/speakers.shtml
http://www.google.co.uk/imgres?imgu...a=X&ei=BoJKUtn0DYal0AXcgoHwBA&ved=0CDwQ9QEwBA
Looking for the first time in years I see this is exactly what he said . Never too young to teach kids . Problems is they turn out like me if you do . I can see why speaker designers made the connection . Like my dad trained in radar I think ? I learned motorcycles also . I could rebuild a carburetor aged seven . Only because I insisted . Turntables also aged five . Making a new fibre needle on a little machine we had . 78's with me is not nostalgia , it is genuinely preferring them to CD . Perhaps because the time they were made was different and people worked harder ? Benny Goodman insisted on a single microphone . The singer stood on a box when time to sing . Benny thought it sounded better ! He was right . Mixing desks mostly destroy simple music . They make complex music possible . The art is in knowing when to use what .
http://www.radartutorial.eu/03.linetheory/tl10.en.html
My dad tried to teach me about radar transmission-lines when about 5 years old . The ideas stuck although the connection with transmission-line speakers is tenuous . At least it made me alert to the idea . Although only five years old slotted waveguides stay in the mind .
I have never found a transmission-line speaker that delivered the performance expected . Thinking about this it should not be so . I suggest the problem is the mid-range would be better as an open baffle ( OB ) . Then the openness would be equal from low bass to mid treble . A ribbon top to complete things . Experiments with my friends excellent Roksan speakers shows that slightly blocking the port destroys the magic . This was his idea as he thinks the bass is slightly one note . I asked him to buy some drinking straws to make a resistive port . Rule of thumb is it changes resonance by shifting it up circa 10% whilst damping the Q . I said we could consider very tight sounding sub woofers that double as speaker stands . I would then crossover to the Roksan K2 above the critical frequency ( 100 Hz ? ) . The subs 190 x 280 x 600 perhaps ? The speakers remind me of the Red Rose Rose Bud at a fraction of the price . I was told the Rose Bud has it's origins in a Chinese design that is slightly upgraded ( selected drive units , better box ) . The price of the originals about $200 a pair . The Roksans are nice enough for me to pay retail .
Superfi - ROKSAN KANDY K2 SPEAKERS (PAIR)
http://www.redrosemusic.com/speakers.shtml
http://www.google.co.uk/imgres?imgu...a=X&ei=BoJKUtn0DYal0AXcgoHwBA&ved=0CDwQ9QEwBA
Looking for the first time in years I see this is exactly what he said . Never too young to teach kids . Problems is they turn out like me if you do . I can see why speaker designers made the connection . Like my dad trained in radar I think ? I learned motorcycles also . I could rebuild a carburetor aged seven . Only because I insisted . Turntables also aged five . Making a new fibre needle on a little machine we had . 78's with me is not nostalgia , it is genuinely preferring them to CD . Perhaps because the time they were made was different and people worked harder ? Benny Goodman insisted on a single microphone . The singer stood on a box when time to sing . Benny thought it sounded better ! He was right . Mixing desks mostly destroy simple music . They make complex music possible . The art is in knowing when to use what .
http://www.radartutorial.eu/03.linetheory/tl10.en.html
Last edited:
Yes, I do say that about Nige, that's true. It's not that I completely disagree with him - all other things being equal, a low parts count ciruit is more reliable than a more complex one, no doubt.
However, there is also such a thing as oversimplification, perhaps not in the absolute sense of working or not working, rather working but how. For example, using a classic single ended diff pair input, I have frequently worked with classic current mirrors, but over the years I have come to the conclusion that a "short form" mirror, using one PNP device from the inverting input collector to the -Vcc and a simple 1N4148 diode works just as well as the "long form", with two trannies base to base and the third trannies base to ground. In effect, one trannie gives way to the said diode. but in my view and my hearing, that is the better version, and simpler to boot.
The question of what is "simple" and what is "complicated" is rather personal. For example, in my book, the one thing that seems to be almost a constant, although there are excpetions, is that I got the best sonic results from amps which used the simplest form of a differential input stage - no cascodes, not FETs, no this or that, just two plain bipolars, with a third acting as a CCS.
I nearly fainted the other day when I looked over a top of the line Sony integrated amp, from their ES series, a model from 1993 or so - holy cow, there was a FET input, cascoded, with current mirror and a few additional transistors which I haven't worked out yet regarding their function, a total of like 10 or 12 devices. A man could make a very decent complete power amp with all that material. I don't mind complex, but that's really stretching the very term to the limit.
This is why I like talking to Thorsten so much, aside from the personal moment. At first glance, his designs may also appear to be complex, but sit down and think about it and soon you find that each component has its job very exactly cut out for it, no fooling around, no waste, but no undue "savings" either (e.g. using a resistor as a CCS, etc).
He always reminds me of the practical aspects. For example, we were discussing how many output pairs should I use. "My" outpt devices are Motorola/ON Semi MJL 3281/1302, I am happy with them and well stocked, never with less than 20+20 on stock at home. Thankfully, they are rated at 200W, rather than 150W as were the Toshiba originals, a very handy aspect. I got to three pairs and proposed four pairs; he simply asked me what would the fourth pair do exactly, in a home amp, which is not expected to be working with very high power levels most of the time? Upon reflection, I realized that a fourth pair ws mereley for show.
This is where Wayne and I do not see eye to eye. This is partly because of Wayne's unique requirements, a speaker with an impedance of 1 Ohm is lamost a short circuit, so Wayne needs a welding machine, not an amp. But the remaining 99.9999% of us own speakers with nomin al impednaces of say 4 Ohms or so, many of 8 Ohms or so, and we do not listen at full blast, but rather at aroubnd 0.5-1 W continuous in our rooms (so as not to drown in music the heavenly sound of our dear wives' voices), with peaks going up. Fair enough, not to cut it too short, one should consider higher power levels as well, just in case, and lower impedances as well. 3 pairs, backed up by 500 VA toroids per side and enough capacitance should be able to deliver peaks into 2 Ohms effectively limited by the PSU voltage only.
So, when is enough really enough? Think about it and you'll see that a lot of work in the High End is spent on pure and simple show rather than necessity. American designers and manufacturers are especially prone to overdoing it, although others are as well. An European manufacturer (ACR, Germany) offers, if memory serves, like 16 pairs per side for a nominally 100W/8 Ohms amp - wonderful ad pictures, but in my view and as per my hearing, that amp sounds no better then several others with far less in their output stages IN A ROOM, not on the test bench.
I found JC's 1205 amp (although in just one audition I had with it, somebody else grabbed it before I could, Parasound is VERY rare in these parts, about as rare as an honest politician), which uses 3 pairs of, if memory serves, nominally 150W devices in such a way that it probably never even noticed it was driving my AR94s, which are no easy load, quite loudly (wife was at work). I did not headbang with it, I just went up to the level where my window panes started rattling, which I could not sustain without developing a massive headache fast.
So, while on a lab bench that amp might have less current capacity than another amp with more output devices, in real life and under regular home conditions, it actually had far more power than it needs for a very convincing result.
The point being, we should not be making anything for theoretical lab bench results, rather we should slweays also be practical and always ask ourselves - what's the point? What is enough? I understand that Wayne has a very special set of requirements, which quite logically influences his views, but we should also bear in mind that in my entire life, I have never come across anyone else with anything like similar requirements.
However, there is also such a thing as oversimplification, perhaps not in the absolute sense of working or not working, rather working but how. For example, using a classic single ended diff pair input, I have frequently worked with classic current mirrors, but over the years I have come to the conclusion that a "short form" mirror, using one PNP device from the inverting input collector to the -Vcc and a simple 1N4148 diode works just as well as the "long form", with two trannies base to base and the third trannies base to ground. In effect, one trannie gives way to the said diode. but in my view and my hearing, that is the better version, and simpler to boot.
The question of what is "simple" and what is "complicated" is rather personal. For example, in my book, the one thing that seems to be almost a constant, although there are excpetions, is that I got the best sonic results from amps which used the simplest form of a differential input stage - no cascodes, not FETs, no this or that, just two plain bipolars, with a third acting as a CCS.
I nearly fainted the other day when I looked over a top of the line Sony integrated amp, from their ES series, a model from 1993 or so - holy cow, there was a FET input, cascoded, with current mirror and a few additional transistors which I haven't worked out yet regarding their function, a total of like 10 or 12 devices. A man could make a very decent complete power amp with all that material. I don't mind complex, but that's really stretching the very term to the limit.
This is why I like talking to Thorsten so much, aside from the personal moment. At first glance, his designs may also appear to be complex, but sit down and think about it and soon you find that each component has its job very exactly cut out for it, no fooling around, no waste, but no undue "savings" either (e.g. using a resistor as a CCS, etc).
He always reminds me of the practical aspects. For example, we were discussing how many output pairs should I use. "My" outpt devices are Motorola/ON Semi MJL 3281/1302, I am happy with them and well stocked, never with less than 20+20 on stock at home. Thankfully, they are rated at 200W, rather than 150W as were the Toshiba originals, a very handy aspect. I got to three pairs and proposed four pairs; he simply asked me what would the fourth pair do exactly, in a home amp, which is not expected to be working with very high power levels most of the time? Upon reflection, I realized that a fourth pair ws mereley for show.
This is where Wayne and I do not see eye to eye. This is partly because of Wayne's unique requirements, a speaker with an impedance of 1 Ohm is lamost a short circuit, so Wayne needs a welding machine, not an amp. But the remaining 99.9999% of us own speakers with nomin al impednaces of say 4 Ohms or so, many of 8 Ohms or so, and we do not listen at full blast, but rather at aroubnd 0.5-1 W continuous in our rooms (so as not to drown in music the heavenly sound of our dear wives' voices), with peaks going up. Fair enough, not to cut it too short, one should consider higher power levels as well, just in case, and lower impedances as well. 3 pairs, backed up by 500 VA toroids per side and enough capacitance should be able to deliver peaks into 2 Ohms effectively limited by the PSU voltage only.
So, when is enough really enough? Think about it and you'll see that a lot of work in the High End is spent on pure and simple show rather than necessity. American designers and manufacturers are especially prone to overdoing it, although others are as well. An European manufacturer (ACR, Germany) offers, if memory serves, like 16 pairs per side for a nominally 100W/8 Ohms amp - wonderful ad pictures, but in my view and as per my hearing, that amp sounds no better then several others with far less in their output stages IN A ROOM, not on the test bench.
I found JC's 1205 amp (although in just one audition I had with it, somebody else grabbed it before I could, Parasound is VERY rare in these parts, about as rare as an honest politician), which uses 3 pairs of, if memory serves, nominally 150W devices in such a way that it probably never even noticed it was driving my AR94s, which are no easy load, quite loudly (wife was at work). I did not headbang with it, I just went up to the level where my window panes started rattling, which I could not sustain without developing a massive headache fast.
So, while on a lab bench that amp might have less current capacity than another amp with more output devices, in real life and under regular home conditions, it actually had far more power than it needs for a very convincing result.
The point being, we should not be making anything for theoretical lab bench results, rather we should slweays also be practical and always ask ourselves - what's the point? What is enough? I understand that Wayne has a very special set of requirements, which quite logically influences his views, but we should also bear in mind that in my entire life, I have never come across anyone else with anything like similar requirements.
Nige, if there's one thing I aam rock solid sure of it's that if you want a big sound, you have to have a big speaker. A small speaker may have many very desirable propertis, but it will always sound small no matter what you do.
If you solve the problem by adding two self-powered subwoofers, what you are really doing is patching up a wrong buy - you should have gone for a proper 3 way speaker straight off.
What might be a good idea is to regard your 3 way speaker as an ideally single sonic unit, but each driver as a separate entity, requiring its own separate enclosure, no matter what the type of enclosure is.
So, you optimise your bass enclosure for best results with the chosen bass driver, using whatever enclosure is best for it. The midrange has its onw enclosure, 100% separate from the bass and tweeter, only being mounted on top of the bass encosure. Ditto for tweeter, which will benefit from the absolute minimum of side refraction, as it should have a maximum of say 1 inch of wood for its enclosure on eack side.
Much like the KEF three chamber reference speaker from the early 80ies. Or the German made KS Linea like - that used a folded transmission line for the bass, ending with something of a large horn at the top of the case. The mid driver was placed just above it in a traditional way, but the tweeter was mounted on a vertical board of almost its net dimanesion in the middle of the horm space above, with nothing solid around it excpet for the board holding it above and below. 3D to die for, and that 8 inch bass has a kick of a mule.
If you solve the problem by adding two self-powered subwoofers, what you are really doing is patching up a wrong buy - you should have gone for a proper 3 way speaker straight off.
What might be a good idea is to regard your 3 way speaker as an ideally single sonic unit, but each driver as a separate entity, requiring its own separate enclosure, no matter what the type of enclosure is.
So, you optimise your bass enclosure for best results with the chosen bass driver, using whatever enclosure is best for it. The midrange has its onw enclosure, 100% separate from the bass and tweeter, only being mounted on top of the bass encosure. Ditto for tweeter, which will benefit from the absolute minimum of side refraction, as it should have a maximum of say 1 inch of wood for its enclosure on eack side.
Much like the KEF three chamber reference speaker from the early 80ies. Or the German made KS Linea like - that used a folded transmission line for the bass, ending with something of a large horn at the top of the case. The mid driver was placed just above it in a traditional way, but the tweeter was mounted on a vertical board of almost its net dimanesion in the middle of the horm space above, with nothing solid around it excpet for the board holding it above and below. 3D to die for, and that 8 inch bass has a kick of a mule.
One thing I learned a long time ago is what is house keeping and what is signal path in op amps . The humble MC33078/79 is as simple as an op amp can be . It has near state of the art performance when bearing that in mind . House keeping is resistors gone wrong . Sometimes it is better . A current source might be as benign as a resistor yet it reaches nearly to infinity . Now only a fool wouldn't use that ? This fool didn't because it was valves and his friend wouldn't like it . I had to accept double distortion as a result and that was by being crafty , five times if being religious to that school of thought .
I saw a circuit that like an idiot I didn't keep . It was a NE5534 with two BD139 connected to pins 1 and 8 ( comp , in LTP ) . The two op amp inputs ( 2, 3 ) taken to the power rail . I meant to look at the internals of a 5534 and didn't . If this wasn't a typo it might work with other chips that have comp . I can sort of see the idea . Knock out the inputs and feed in up the op amp . Is it too good to be true ? The 5534 is quiet , interesting BD139 should better it ? 2N4403 was said to work . CCS to the tail . +/- 15 V . It also said never reverse bias a transistor as it destroys the noise performance . Never use a meter diode test was said .
Dvv . The friend had a very nasty divorce and lives in a shoebox , like a fool he bashed his ex over the head with a solicitor. Thus the shoebox house . By being a double woofer we could match the Naim SBL . Also the imaging should be out of this world good if I do . These subs would give colour more than weight . Open would be nice . Big headphones .
When I divorced I said to my ex if you keep away from solicitors you can have half . I made a future deal to say at any time you choose . She still hasn't cashed it in four years on ? She said would I keep to that if I made some money ? She made $30 000 since then and I am happier to give it to her than them . In England we are allowed to do that . A $500 divorce . If I spent it all she would still get half . She said fine , she knows I won't . Crazy England hey . She got cancer and I could never put that right is my guess , always in my work must have seemed so bad ? I am taking her to hospital this week , alas a new lump . I met Colleen through this and sorry to say Colleen is who I would have fallen for regardless . I sort of knew her a longtime and whilst being very correct made it clear I liked her . Then she divorced as I after 30 years of marriage . Both of us free at the same time . I am sure this story is common in our industry ?
I saw a circuit that like an idiot I didn't keep . It was a NE5534 with two BD139 connected to pins 1 and 8 ( comp , in LTP ) . The two op amp inputs ( 2, 3 ) taken to the power rail . I meant to look at the internals of a 5534 and didn't . If this wasn't a typo it might work with other chips that have comp . I can sort of see the idea . Knock out the inputs and feed in up the op amp . Is it too good to be true ? The 5534 is quiet , interesting BD139 should better it ? 2N4403 was said to work . CCS to the tail . +/- 15 V . It also said never reverse bias a transistor as it destroys the noise performance . Never use a meter diode test was said .
Dvv . The friend had a very nasty divorce and lives in a shoebox , like a fool he bashed his ex over the head with a solicitor. Thus the shoebox house . By being a double woofer we could match the Naim SBL . Also the imaging should be out of this world good if I do . These subs would give colour more than weight . Open would be nice . Big headphones .
When I divorced I said to my ex if you keep away from solicitors you can have half . I made a future deal to say at any time you choose . She still hasn't cashed it in four years on ? She said would I keep to that if I made some money ? She made $30 000 since then and I am happier to give it to her than them . In England we are allowed to do that . A $500 divorce . If I spent it all she would still get half . She said fine , she knows I won't . Crazy England hey . She got cancer and I could never put that right is my guess , always in my work must have seemed so bad ? I am taking her to hospital this week , alas a new lump . I met Colleen through this and sorry to say Colleen is who I would have fallen for regardless . I sort of knew her a longtime and whilst being very correct made it clear I liked her . Then she divorced as I after 30 years of marriage . Both of us free at the same time . I am sure this story is common in our industry ?
Last edited:
D,
Forget about my special needs for the moment, i dont think of it when in discussion, if i did you would need 8 pr not 4 ...
Now , i have had many amplfiers over the years , i still count at least 5 different ones at this time , usually its in The teens , varying from vintage to current, in such travels do you know how many had less than 4 prs ...
Errr one ...
I still have it, it has 3 pr , its a PS Audio 200cx and yes it can play the Big Bembeh at 1 ohm. My other speakers are nominal 8 and 4, dips and phase angles on the 4 ohm version will represent 2 ohm. Now if you are building this DIY , then why compromise , why not 4 pr or 8 pr , why run such a small output Stage vs PSU ..? Is there a sonic advantage to doing so..?
Also take a look at how many speakers have impedance magnitude and phase representing 2 ohms , look at ML esl's or a pr of stacked quads , Wilson Audio, Et al .
Most large 3 or 4, 4.5 way system done correctly will have amp loads as low as 2 ohm, then Again if you want another me too mid amp , well , I'm past that , anything done now is for SOTA sound or close to it , that starts with large speakers for proper scale , then amplification ( multiple) to deliver said scale, with ease ...
So again nothing wrong with a 2 or 3 pr amplifier , i only ask why , if only for your application , then Ok, if flea power for high sensitivity speakers , then Ok too, my Questions were only to ascertain your choosen Direction , for me i like over built stuff , it has served me well over the years , never lost an amp in 35 years from doing so and I'm too old to be playing around with pam pam stuff , so if your 3 pr works then I'm fine with that...
Seen ....
Forget about my special needs for the moment, i dont think of it when in discussion, if i did you would need 8 pr not 4 ...
Now , i have had many amplfiers over the years , i still count at least 5 different ones at this time , usually its in The teens , varying from vintage to current, in such travels do you know how many had less than 4 prs ...
Errr one ...
I still have it, it has 3 pr , its a PS Audio 200cx and yes it can play the Big Bembeh at 1 ohm. My other speakers are nominal 8 and 4, dips and phase angles on the 4 ohm version will represent 2 ohm. Now if you are building this DIY , then why compromise , why not 4 pr or 8 pr , why run such a small output Stage vs PSU ..? Is there a sonic advantage to doing so..?
Also take a look at how many speakers have impedance magnitude and phase representing 2 ohms , look at ML esl's or a pr of stacked quads , Wilson Audio, Et al .
Most large 3 or 4, 4.5 way system done correctly will have amp loads as low as 2 ohm, then Again if you want another me too mid amp , well , I'm past that , anything done now is for SOTA sound or close to it , that starts with large speakers for proper scale , then amplification ( multiple) to deliver said scale, with ease ...
So again nothing wrong with a 2 or 3 pr amplifier , i only ask why , if only for your application , then Ok, if flea power for high sensitivity speakers , then Ok too, my Questions were only to ascertain your choosen Direction , for me i like over built stuff , it has served me well over the years , never lost an amp in 35 years from doing so and I'm too old to be playing around with pam pam stuff , so if your 3 pr works then I'm fine with that...
Seen ....
Last edited:
D,
Forget about my special needs for the moment, i dont think of it when in discussion, if i did you would need 8 pr not 4 ...
Now , i have had many amplfiers over the years , i still count at least 5 different ones at this time , usually its in The teens , varying from vintage to current, in such travels do you know how many had less than 4 prs ...
Errr one ...
I still have it, it has 3 pr , its a PS Audio 200cx and yes it can play the Big Bembeh at 1 ohm. My other speakers are nominal 8 and 4, dips and phase angles on the 4 ohm version will represent 2 ohm. Now if you are building this DIY , then why compromise , why not 4 pr or 8 pr , why run such a small output Stage vs PSU ..? Is there a sonic advantage to doing so..?
Also take a look at how many speakers have impedance magnitude and phase representing 2 ohms , look at ML esl's or a pr of stacked quads , Wilson Audio, Et al .
Most large 3 or 4, 4.5 way system done correctly will have amp loads as low as 2 ohm, then Again if you want another me too mid amp , well , I'm past that , anything done now is for SOTA sound or close to it , that starts with large speakers for proper scale , then amplification ( multiple) to deliver said scale, with ease ...
So again nothing wrong with a 2 or 3 pr amplifier , i only ask why , if only for your application , then Ok, if flea power for high sensitivity speakers , then Ok too, my Questions were only to ascertain your choosen Direction , for me i like over built stuff , it has served me well over the years , never lost an amp in 35 years from doing so and I'm too old to be playing around with pam pam stuff , so if your 3 pr works then I'm fine with that...
Seen ....
So you're a "person with special needs"? Sounds omnious, easy to misconstrue ...
Just yanking your chain, Wayne, ol' buddy.
First of all, I see nothing flea sized about my output stage. It will work steady state full power into 3 Ohms. However, into 2 Ohms, it will do a full power output in short term only. As I see it, for real world music listening rather than full power steady state, that is more than enough, assuming you listen to music rather than steady state sine waves. If so, you can literally NEVER runt it at full power because you would have very little headroom before clipping, reducing your music to incomprehensible hash and your speakers, NO MATTER WHICH, to cinders, because no speaker can take clipped waveforms for longer than a few seconds.
Let me try it like this - in the real world, you buy a 100W amp so you can enjoy 25W as they should be emnjoyed.
And it never ceases to amaze me how few people have any idea of how loud just 10W of continuous power is in a room.
Secondly, rather than brutally keep piling up output transistor pairs, let's use our heads a bit. My H/K 680 integrated amp has two pairs of 150W devices at its outputs. It is a dual mono design, using one 500 VA (judging from the voltages and assciated fuse values) classic transformer, but with separate bridge rectifiers and 2 x 8,200 uF caps per channel. Yet, it has no problem whatsoever to pump out 540 Watts into 2 Ohms in IEC standards defined 20 mS intervals. My H/K Citation 24 power amp also uses 2 pairs of 150W output devices per channel and will deliver 570 Watts into 2 Ohms in peaks. Some Denon designs can do that as well, not to even mention Sony, or even low cost Nikko, whose Alpha amps were pulling that stunt off 20+ years ago.
This begs the rather obvious question - why use more to much more for the same effect? What's the point? Unless you want it to be able to work steady state into a lab resistor.
But, being cautious, I opted not for 150W devices, but rather for 200W devices, and not 2 but 3 pairs, 50% more. Just to be on the safe side. My H/K Citation 24 amp originally had 2 x 6,800uF caps per each channel and pulled it off; I use a total of 2 x 26,600 uF, alsmost 4 TIMES (!) more, again, just to be on the safe side.
Thus, I see absolutely no reason whatsoever to pile up more output devices, rather my point is to make the ones I do have as well uased as possible, obviously without putting them in danger.
Do not forget that adding more and more output devices has its downsides as well. For one, your per transistor capacitances start adding up, for another, your heat sink requirements start to soar, creating a problem with space and, let's be honest, price as well. The mere problem of pairing them well increases exponentially, unless you can get hold of factory pairid to say 5% NPN and PNP devices. Most of us are way too small to have access to that service.
And yes, price is most definitely a factor in my reckoning. Like everybody else, I want the most bang for my buck/pound/euro, however, I do try to mix quality as well into that potpurri. I know that's a slippery slope concept and choice, others will have different ideas, but that's just fine with me - if they feel they can do better, by all means, roll up your sleeves and go for it. The more, the merrier and the better the DIYer's choice.
All I can say has spolit my day is the fact that as I understand it, Thorsten has abandoned his Sir Pottymouth project. This is, I feel, a loss for us all, and certainly for me, I would have made one just because I want to have some of Thorsten's work, which I believe to be way above the norm. Let's hope I got it wrong. If not, I will give him hell for it.
Lastly, I will refer you to what already wrote here on several occasions - the way to have a great 100W amp is to make a good 150W amp and call it a 100W amp. My take is a nominally 100 W/8 Ohms amp which will not clip before 170+ Watts into 8 Ohms, so, as a matter of fact, I expect peak power into 2 Ohms to be over 600 Watts.
If that's "flea power", then I would refer you to my friend Milan Karan, his top amp, costing top dollars, will deliver over 3,500 Watts into 2 Ohms and weight is at around 180 lbs. If memory serves, it has 16 pairs of top Sanken power transistors - per channel. It does 1,200 W/8 Ohms. Don't know the current price, but I doubt it's below US$ 50,000. Delivered in a lovely wooden crate, tailor made for it. He's my friend of old, I'm sure I could get you a discount.
Yes , Karan .....
Now That's what I'm talking about you could diy it at 5% of that cost , c'mon D, go big or go home ...
Anyway 89db/w/m speakers listening distance 4m , how much power to reproduce 103 db peaks, and remember or in case you were not aware best sonics from sand amplifier is at 33% of rated full RMS..
Now go figure .....
Now That's what I'm talking about you could diy it at 5% of that cost , c'mon D, go big or go home ...
Anyway 89db/w/m speakers listening distance 4m , how much power to reproduce 103 db peaks, and remember or in case you were not aware best sonics from sand amplifier is at 33% of rated full RMS..
Now go figure .....
Actually, the Karan IA-180 I have, now 10 years old, is an outstanding device in terms of sound quality. I wouldn't part with it at any price, all the more so as Milan put it in my hands himself. All 6'6", 280 lbs of smiling himself.
BUT, his amp is not indifferent to what it is driving. It's a fully balanced design, and has only 2 pairs of Sanken 200W devices per side, so it's not too happy with low impedance loads.
With specifically my speakers, I got the same comment from two tube affictionado friends who don't know each other, that I shafted them because I was listening to a typical SET sound but minus the SET shortcomings, primarily referring to output power (IA-180 is rated at 180/250W into 8/4 Ohms, kinda more than your typical SET amp). It does sound sublime in my room, and I do love it dearly.
You see Wayne, this is why I understand so well why you require what you do, because I know full well how much the speakers influence the overall musical result. As opposed to yours, mine seem to be the exact opposite, such an easy load that real SET amps drive them with aplomb. That was one of our requirements when making it, we both (my coworker and friend and myself) have several tube lover friends, and two are professionally involved with tube audio. You don't do things so as to leave your friends out in the cold.
Theoretically, my speakers do 92 dB SPL at 1m with 2.83V. The Karan can do 22 dBW, so theoretically, I should be able to get (92+22) 114 dB SPL at 1 m at max output. Actually about 3 dB more, since I have two speakers working in parallel. If we assume that SPL decreases by 3 dB for every additional 1m distance, the actual maximum would be around 111 dB in my room at 3m (10 ft). As I see it, I definitely do not have any dynamic range issues in my room, not with my speakers.
Now, if I were you, I'd definitely think of biamping. For the 1 Ohm bass driver, I'd use like 4 or 5 pairs of Motorola/ON Semi MJ 21195/21196 TO-3 250W devices. They are well neigh indestructible, generally you need a hammer to do them any harm, unless you keep them in short circuit for like 2 or 3 minutes continuously; but aside from sheer and obvious stupidity, they - like diamonds- are forever.
For mid and treble, I'd use something like the Centurion, trust me, that's quite enough.
BUT, his amp is not indifferent to what it is driving. It's a fully balanced design, and has only 2 pairs of Sanken 200W devices per side, so it's not too happy with low impedance loads.
With specifically my speakers, I got the same comment from two tube affictionado friends who don't know each other, that I shafted them because I was listening to a typical SET sound but minus the SET shortcomings, primarily referring to output power (IA-180 is rated at 180/250W into 8/4 Ohms, kinda more than your typical SET amp). It does sound sublime in my room, and I do love it dearly.
You see Wayne, this is why I understand so well why you require what you do, because I know full well how much the speakers influence the overall musical result. As opposed to yours, mine seem to be the exact opposite, such an easy load that real SET amps drive them with aplomb. That was one of our requirements when making it, we both (my coworker and friend and myself) have several tube lover friends, and two are professionally involved with tube audio. You don't do things so as to leave your friends out in the cold.
Theoretically, my speakers do 92 dB SPL at 1m with 2.83V. The Karan can do 22 dBW, so theoretically, I should be able to get (92+22) 114 dB SPL at 1 m at max output. Actually about 3 dB more, since I have two speakers working in parallel. If we assume that SPL decreases by 3 dB for every additional 1m distance, the actual maximum would be around 111 dB in my room at 3m (10 ft). As I see it, I definitely do not have any dynamic range issues in my room, not with my speakers.
Now, if I were you, I'd definitely think of biamping. For the 1 Ohm bass driver, I'd use like 4 or 5 pairs of Motorola/ON Semi MJ 21195/21196 TO-3 250W devices. They are well neigh indestructible, generally you need a hammer to do them any harm, unless you keep them in short circuit for like 2 or 3 minutes continuously; but aside from sheer and obvious stupidity, they - like diamonds- are forever.
For mid and treble, I'd use something like the Centurion, trust me, that's quite enough.
OK, now take 12 db away for 4 M listening distance you end up with 102, and that is at max RMS power. So if we factor i n power output at 33% , you are way below , so technically you would need 3 times your amp power to achieve best sonics and dynamics, at 4 M listening distance.
This is the very reason why big amps do better than small, even when running high sensitivity speakers, power speaks to you ....
For the linesource, the low impedance (1.5) is from 200-22K , bass is 4 ohm (dynamic speaker), the centurion would ask for a drink after 5 mins .......
* Anyway , hurry up on this beast , i wanna taste it ...
This is the very reason why big amps do better than small, even when running high sensitivity speakers, power speaks to you ....
For the linesource, the low impedance (1.5) is from 200-22K , bass is 4 ohm (dynamic speaker), the centurion would ask for a drink after 5 mins .......
* Anyway , hurry up on this beast , i wanna taste it ...
Last edited:
OK, now take 12 db away for 4 M listening distance you end up with 102, and that is at max RMS power. So if we factor i n power output at 33% , you are way below , so technically you would need 3 times your amp power to achieve best sonics and dynamics, at 4 M listening distance.
This is the very reason why big amps do better than small, even when running high sensitivity speakers, power speaks to you ....
For the linesource, the low impedance (1.5) is from 200-22K , bass is 4 ohm (dynamic speaker), the centurion would ask for a drink after 5 mins .......
* Anyway , hurry up on this beast , i wanna taste it ...
I neither understand nor see how you worked that out - please elaborate. The part about listening distance, I mean.
My speakers are almost exactly 3m away from me, so why should I factor 4m?
Have you taken into account both the losses and the reverberation of the room?
My 92 dB/2.83V/1m applies to free air, in room that's more like 94 dB.
So, my math could look something like this: 94+22=114 dB, + 3dB because two speakers are working, but minus whatever the room soaks up, and -9 dB for the 3 m distance, which leaves me with 117-9-?= app. 108 dB SPL.
BUT, using professional peak meters, I discovered that my actual normal power use is 0.4-0.6 Watts only, with peaks hitting 12-15 Watts on exceptionally dynamic material. Under normal conditions, I very much doubt you would ever use 92 dB SPL as your usual normal listening reference point. It's simply too loud, although this depends on many other factors as well (such as ambient noise, possibly time of day as nights will be more quiet, etc.
Obviously, these are my personal needs, not any general needs by any means. I have had a pair of Studer badged pro peak meters in a console for some monzhs with me, and NEVER ONCE did it hit even the 50W mark. Not once.
A guy I know called Guy showed me the effect of room sealing to get high SPL . It was surprisingly difficult to get above 110 dB . He thought cars to be better as they have door seals . He even showed me how the room works as a compressor in all the senses . What works is to have the greatest dynamic range . That is low distortion at any levels and low noise ( which is also a distortion ) . Then the mechanism of relative difference in the brain says goodness that is loud . MP3 robs us of any chance of that . I now see layers and niceness as priorities as many sources are loosely speaking MP3 . Guy was the Klipshe distributor . He felt regardless of the efficiency of the speakers it never really goes as loud as one expects . Guy even said the air gets warm above 110 db more than anything useful . I ran 350 watts into La Scala's once in a small room . Was awful and not especially loud . Knocked the protection on in the process .
As best as I can draw it this is the adapted NE5534 I wrote about previously . Better than the usual way of adding a microphone stage to the old 741 . That in itself was neat as the new input pair biased of of the internal pair . Having done this with a MC 33079 that is like 5534 the noise improvement is not worth the sound trade off . Here it is different . A substantial chunk of circuitry removed and a similar signal path . The difference is a transistor better suited to the task can be used . Even BD139 suits a 600R source better , 5534 suits 5 K from memory . I dare say there are many more to look at . BD 139 is about 0.7 nV root Hz ( 2n4401 perhaps ) . 5534 about 4 nV ( total , guess 3 times ) . Some op amps have a similar compensation arrangement I think ? Any thoughts or other op amps ? Good Rbb- NPN ? I am a little dubious about taking the inputs to the rail . LM324 would be fine . I can see why as any noise is unwelcome . I will read up and see if 5534 is safe that way . I will use a very high gain on this so will avoid and need for comp ( 120 ) . Doubtless it would still work anyway .
Last edited:
This very cheap stereo DAC ( $2 ) comes with more useful data compered with the one I have been asked about . Anyone have an idea of - 60dB distortion of a state of the art DAC ?
http://docs-europe.electrocomponents.com/webdocs/0edd/0900766b80eddd46.pdf
The DAC to be studied .
http://docs-europe.electrocomponents.com/webdocs/0c97/0900766b80c97a93.pdf
http://docs-europe.electrocomponents.com/webdocs/0edd/0900766b80eddd46.pdf
The DAC to be studied .
http://docs-europe.electrocomponents.com/webdocs/0c97/0900766b80c97a93.pdf
This DAC chip, sells for less than $1.00 in large (100K+) quantities is substantially better in both numbers and sound http://www.akm.com/akm/en/file/datasheet/AK4430ET.pdf
Usually the -60 dB THD+N test on a decent DAC chip is a measure of the noise. The attached plot of the AK4399 will illustrate what I mean. The premium parts are very good today. I think real sound differences with the best parts (AKM ESS TI Wolfson) don't come from the chips as much as the implementations.
Usually the -60 dB THD+N test on a decent DAC chip is a measure of the noise. The attached plot of the AK4399 will illustrate what I mean. The premium parts are very good today. I think real sound differences with the best parts (AKM ESS TI Wolfson) don't come from the chips as much as the implementations.
Attachments
- Status
- Not open for further replies.