Re: focus pocus
You're getting sillier by the post. Please post a schematic of both your chip amp and your "hybrid" without cutting and pasting someone else's work.
DrG said:
You're getting sillier by the post. Please post a schematic of both your chip amp and your "hybrid" without cutting and pasting someone else's work.
Nothing, but clearly we love to argue about nothing! Maybe we should get real jobs 🙄
DrG doesn't even know if he doesn't like all chips, just the one (or 2?) that he has tried, in god knows what circuit, yet he lumps all chips in all configurations in all circuits as all bad, while admitting that topography and circuit design are paramount
Variac said:
Hehehe. Well there are technical aspects that can be argued. And I started to myself, but upon further thought, I don't see that technical aspects are particularly germane to this thread.
DrG doesn't even know if he doesn't like all chips, just the one (or 2?) that he has tried, in god knows what circuit, yet he lumps all chips in all configurations in all circuits as all bad, while admitting that topography and circuit design are paramount
That's fine. He can be as prejudicial as he likes. Doesn't matter to me if he likes them or not, even those he's never tried. And don't see why it would matter to anyone else.
The only thing arguable here as I see it is whether one individual's subjective tastes and preferences have any inherently greater validity than the subjective tastes and preferences of others.
se
sheet!
Dos vidanya, comrade pedroskova, sitting on your cornflake... snap, crackle and pop! Do you take milk?
Scanning my drawn hybrid circuit is such a schlepp... I have a whole file of goodies to digitize and so little time... But it's simple: EF86, pentode-connected, dc into EL86(triode-connected, resistor-loaded), caps into 2SJ50/2SK135 Hitachis simply biased by a resistor network, about 6dB NFB.
If you have the relevant tube curves, figuring out suitable resistances and cap values should be easy. And lotsa fun!. If you don't have the data I can e-mail it to you. But don't expect miracles - any more than I buy the chip miracles claimed here... did you guys hear they took the word 'gullible' out of the dictionary... hehehe
Just do it right and get a cheap, simple, solid amp with a sweet, musical character. Or better still, sit down and do something all your own from scratch! You may just surprize yourself. And everyone else here. Better than another chip off the ol' block I think.
See what I can do about posting the circuit sometime this weekend. Oh, and I absolutely never plagiarise the circuits of others so I guarantee no "cutting and pasting". After all, I'm a balanced guy... I have a chip on both shoulders... but none in my amps
Dos vidanya, comrade pedroskova, sitting on your cornflake... snap, crackle and pop! Do you take milk?
Scanning my drawn hybrid circuit is such a schlepp... I have a whole file of goodies to digitize and so little time... But it's simple: EF86, pentode-connected, dc into EL86(triode-connected, resistor-loaded), caps into 2SJ50/2SK135 Hitachis simply biased by a resistor network, about 6dB NFB.
If you have the relevant tube curves, figuring out suitable resistances and cap values should be easy. And lotsa fun!. If you don't have the data I can e-mail it to you. But don't expect miracles - any more than I buy the chip miracles claimed here... did you guys hear they took the word 'gullible' out of the dictionary... hehehe
Just do it right and get a cheap, simple, solid amp with a sweet, musical character. Or better still, sit down and do something all your own from scratch! You may just surprize yourself. And everyone else here. Better than another chip off the ol' block I think.
See what I can do about posting the circuit sometime this weekend. Oh, and I absolutely never plagiarise the circuits of others so I guarantee no "cutting and pasting". After all, I'm a balanced guy... I have a chip on both shoulders... but none in my amps
Hi,
Somebody please look at the equivalent circuit of LM3875 from the spec-sheet. I have read good things about such circuit (in electronics book/theory), but why have I never found such circuit (the driver) in normal amps?
Why cannot anyone of you try to immitate the circuit with discrete transistors? I'm no EE and never been successful in designing my amps in the past (Well, I don't have necessary tools, only a DMM and a calculator! 😀). But even so, unless any of you think it is unnecessary to do so, I will try again using NPN darlingtons to replace the output transistors. I will try MJ11016, BD139, BD140, and any small signal transistors I have on hand. Hmmm, may be I need to buy TIP darlingtons, otherwise I will burn all the Motorolas... Wait!
DrG,
I'm looking forward to plagiarising your hybrid amp!! 😀 😀 😀 I think I better cancel the Discrete LM3875 project and wait for the hybrid. Anyway I need to find a good house for my J50/K135 (They are too expensive for a Mini-Crescendo!)
___________________________________________________
"Am no expert, take my advice at yer own risk." - Sigmund Freud
Somebody please look at the equivalent circuit of LM3875 from the spec-sheet. I have read good things about such circuit (in electronics book/theory), but why have I never found such circuit (the driver) in normal amps?
Why cannot anyone of you try to immitate the circuit with discrete transistors? I'm no EE and never been successful in designing my amps in the past (Well, I don't have necessary tools, only a DMM and a calculator! 😀). But even so, unless any of you think it is unnecessary to do so, I will try again using NPN darlingtons to replace the output transistors. I will try MJ11016, BD139, BD140, and any small signal transistors I have on hand. Hmmm, may be I need to buy TIP darlingtons, otherwise I will burn all the Motorolas... Wait!
DrG,
I'm looking forward to plagiarising your hybrid amp!! 😀 😀 😀 I think I better cancel the Discrete LM3875 project and wait for the hybrid. Anyway I need to find a good house for my J50/K135 (They are too expensive for a Mini-Crescendo!)
___________________________________________________
"Am no expert, take my advice at yer own risk." - Sigmund Freud
Variac: I take the point about watches... something similar happens here, but I think we haven't reached that point yet. And the comparison limps a little since quartz technology killed the traditional watch and not some comparable, but integrated mechanical technology - as such completely different from the mechanical approaches.
The point remains the same: discrete configurations can be adapted to the application. Chips get designed for allround qualities. If you use a discrete technology indiscriminately, use chips instead indeed. If you want fail safe approaches or save money - again, use chips. If you want to learn about the underlying technology and build a coherent system adapted to your needs, I believe you should use the discrete approach.
I started with chips to get a cheap, reliable system quick. I was happy up to some point. Then I wanted better, and I wanted to learn. So now I start building discretes. I still have to plagiarize since I don't know enough yet. But I really enjoy the possibilities by discrete - to play with the circuit, the values, to adapt it to the system. How much more fun than just swapping resistor types.
Besides: If chips really reac so sensitively to components, I'd call them not "revealing" but "bitchy".
Besides (bis): chip amps have one drawback you can see by data sheet: distortion usually rises with *falling* output power, and reaches best numbers around 100 Hz (!) - ex. see LM3886 data sheet. That can not help low level resolution or midrange clarity. If you look closely, they perfom OK but at low power and in the midrange, even simple single ended class A circuits a la Pass Labs reach better numbers that the chips! Don't go by the average specs, look more closely ;-)
This can be used in a systems approach such as in KYW 's comment: bi- or tri-amp, build amps for each band who display strengths where you need it. So - for instance - use chip amps in mid bass or bass: high power at low f for lowest THD. Class A SE for midrange - lowest THD in low to medium power midrange. And either a chip or again a class A for the tweeter: low THD initially, and high THD with higher order harmonics in a band where the harmonics fall outof the audible range.
The point remains the same: discrete configurations can be adapted to the application. Chips get designed for allround qualities. If you use a discrete technology indiscriminately, use chips instead indeed. If you want fail safe approaches or save money - again, use chips. If you want to learn about the underlying technology and build a coherent system adapted to your needs, I believe you should use the discrete approach.
I started with chips to get a cheap, reliable system quick. I was happy up to some point. Then I wanted better, and I wanted to learn. So now I start building discretes. I still have to plagiarize since I don't know enough yet. But I really enjoy the possibilities by discrete - to play with the circuit, the values, to adapt it to the system. How much more fun than just swapping resistor types.
Besides: If chips really reac so sensitively to components, I'd call them not "revealing" but "bitchy".
Besides (bis): chip amps have one drawback you can see by data sheet: distortion usually rises with *falling* output power, and reaches best numbers around 100 Hz (!) - ex. see LM3886 data sheet. That can not help low level resolution or midrange clarity. If you look closely, they perfom OK but at low power and in the midrange, even simple single ended class A circuits a la Pass Labs reach better numbers that the chips! Don't go by the average specs, look more closely ;-)
This can be used in a systems approach such as in KYW 's comment: bi- or tri-amp, build amps for each band who display strengths where you need it. So - for instance - use chip amps in mid bass or bass: high power at low f for lowest THD. Class A SE for midrange - lowest THD in low to medium power midrange. And either a chip or again a class A for the tweeter: low THD initially, and high THD with higher order harmonics in a band where the harmonics fall outof the audible range.
MBK said:
Better numbers? What happened to "trust your ears"? Or are we heading back to the "trust the numbers" THD wars of the 70s?
se
Ok chip amps may still be a bit controversial but what about chip preamps based on something like the widely appreciated opa627 for example. More and more designers and diy ers are using these instead of discrete designs and with good results.So the argument is that a chip has the capability to sound very very good in line stages so why shouldn't it be possible to have a power chip that sounds as good as discrete designs. I think expressing a blanket view that "op-amps suck" just doesn't convince anybody anymore. It reminds one of the old and tired SS vs tube debate or the more recent digital vs vinyl debate.
It can only go around in circles for ever . At the end of the day I agree that personal choices will matter for the individual but a way to prove that chip amps are good sounding is to see how many people are replacing their good tube or ss amps out there for them. You can't really beat statistics in the end.This forum gives an indication of how many experienced DIY audiophiles have positively appraised a well built chip amp with respect to other good designs.
It can only go around in circles for ever . At the end of the day I agree that personal choices will matter for the individual but a way to prove that chip amps are good sounding is to see how many people are replacing their good tube or ss amps out there for them. You can't really beat statistics in the end.This forum gives an indication of how many experienced DIY audiophiles have positively appraised a well built chip amp with respect to other good designs.
Very interesting. I kinda like my GC but still have reservations about opa627s and similar. Of course my 'real' system is valves and in that context i seem to like the compromises offered by GC but the 637/627 seemply don't dig deep enough in a phono preamp. The sound is pleasant but mid-fi. I've tried a variety of other chips like INA103, LM394 and consistently they manage to dig less music out of the grooves than good valve circuits. With most CD based systems this will not be so noticeable as the dacs commonly use opamps. Maybe chips are simply not suited for very low levels. Even simple FET phonos like Pacific seem better than even the best opamps.
protos said:
But why does anything such as this need to be proved at all? Whether one likes or dislikes chip amps, why would anyone care whether anyone else liked or disliked them?
se
I have always hated the sound of solid state amp's including top model's from the likes of Krell and Chord, I decided to build a GC for the pc, I was so impressed with the sound it now sit's proudly in my main system next to my tube amp.
My GC circuit is same as NUUK's, for the price these little amps cost to make they sound fantstic
My GC circuit is same as NUUK's, for the price these little amps cost to make they sound fantstic
leo could be right...
For me the key phrase is "...for the price these little amps cost..." - which is no prob, really. But if this is so, then what might be possible for a little more... effort/money/innovation/whatever... Or does everyone here truly believe that audio is so defined by the gainclown, so as to make a pointless exercise of all attempts to look farther...
I made a really good phono stage using SSM2015 chip from AD. It's a low-noise microphone pre-amp running in class A. Now I know it's been discontinued but there's probably a suitable replacement. It sounded excellent on an Oracle with Ortofon MC cartridge into an AR VT100 and electroststics. Such a device would be an interesting alternative to conventional OP/OPA/LM devices in a regular pre-amp. Although I favour tubes for pre-amp duty.
For me the key phrase is "...for the price these little amps cost..." - which is no prob, really. But if this is so, then what might be possible for a little more... effort/money/innovation/whatever... Or does everyone here truly believe that audio is so defined by the gainclown, so as to make a pointless exercise of all attempts to look farther...
I made a really good phono stage using SSM2015 chip from AD. It's a low-noise microphone pre-amp running in class A. Now I know it's been discontinued but there's probably a suitable replacement. It sounded excellent on an Oracle with Ortofon MC cartridge into an AR VT100 and electroststics. Such a device would be an interesting alternative to conventional OP/OPA/LM devices in a regular pre-amp. Although I favour tubes for pre-amp duty.
Hi,
Just as you bash on the GC I can do the same for this phono gear.
Oracle Delphi ? : definetely passe, coloured by its mat and lacking in detail and dynamics because of that and its' suspension. ( the latter two going hand in hand)
Ortofon? Too low an output from most to be useable in most preamps, there are much better carts than this for less money unless you're talking about their entry level range.
ARC VT100? Never impressed me bar the ease with which they breakdown, again there's much better sounding and much, much more reliable gear around.
Exactly.
Happy now?🙄
Common DrG, nobody said the IGC, buffered or not, is beating all high-end gear hands down.
It is I believe a godsend to people trying to explore the realms of high-end at a very reasonable cost.
I've never heard any but from looking at the innards and from the feedback I get from others it wouldn't surprise me it can beat most twenty years old SS amps with the same power rating hands down.
Hey, unless you doubt me, I am one of those tube guys NOBODY is ever going to convince that SS amps are better...
Voila, I said it...
Cheers guys and have fun, 😉
Just as you bash on the GC I can do the same for this phono gear.
Oracle Delphi ? : definetely passe, coloured by its mat and lacking in detail and dynamics because of that and its' suspension. ( the latter two going hand in hand)
Ortofon? Too low an output from most to be useable in most preamps, there are much better carts than this for less money unless you're talking about their entry level range.
ARC VT100? Never impressed me bar the ease with which they breakdown, again there's much better sounding and much, much more reliable gear around.
Exactly.
Happy now?🙄
Common DrG, nobody said the IGC, buffered or not, is beating all high-end gear hands down.
It is I believe a godsend to people trying to explore the realms of high-end at a very reasonable cost.
I've never heard any but from looking at the innards and from the feedback I get from others it wouldn't surprise me it can beat most twenty years old SS amps with the same power rating hands down.
Hey, unless you doubt me, I am one of those tube guys NOBODY is ever going to convince that SS amps are better...
Voila, I said it...
Cheers guys and have fun, 😉
I also agree that my tube amp sounds better than my GC (to my ears) but I'm still yet to hear a solid state amp that easily beats a well made GC. I have heard alot of these so called highend amps,some of them are very detailed and powerful but to me there is no involvement, at least with the GC I can listen to it for longer than 10 minutes
Steve Eddy said:
My comment on the numbers aimed at the argument that the laser trimmed accuracy of the integrated approach in chips beats accuracy of discrete designs achievable by the DIY'er. I don't see this advantage in accuracy in the data sheets.
Also look at the production standard deviations. You can get around that by selecting the best out of a large number, but as with the PSU and resistor network tweaking: by the time you succeed, you might have built and fine tuned three discretes ;-)
Interestingly the LM3886 for instance also exhibits phase shift below 800 Hz, p.12 of data sheet.
In general: good numbers may not suffice to make it sound good, but bad numbers likely don't help...
😉
About resistor accuracy...
The only area where I see resistor accuracy/trimming practically having an advantage is DC offset. If the _tolerance_ of a resistor had been an area of any significant improvement, I would have hoped to have heard about this "tweak" by now. This is certainly do-able in a discrete design, where 1% metal films (in a respectable design) seem to be the order of the day.
The only area where I see resistor accuracy/trimming practically having an advantage is DC offset. If the _tolerance_ of a resistor had been an area of any significant improvement, I would have hoped to have heard about this "tweak" by now. This is certainly do-able in a discrete design, where 1% metal films (in a respectable design) seem to be the order of the day.
- Status
- Not open for further replies.