john curl said:
I see your associates are up at the whistle
I showed, among other, exactly this: feedback, current servo (making JFET offsets and handmatching a moot point), very low distortion, no caps in the signal path.
Let me understand correctly: now that complementary low noise JFETs are no longer available, nobody is going to build low noise MC preamps? Wanna learn something about bipolar base current cancelling?
scott wurcer said:
Hello Scott
With your AD797 to reduce input offset currents you feed current from a bipolar current source to the + and - terminals of the opamp and achieve a low noise design . Why cant this approach be used with discrete bipolar designs which use non complementary inputs. Is there something I am missing.
Are these current source elaborate in the AD797 , circuit details would be nice.
Regards
Arthur
WOW! .132nV/rt Hz. Does anyone know what it takes to get this over a wide frequency range? Certainly no feedback loop can be conveniently used.
Let's go back to this truly exceptional design and look at it more carefully. Mind you, anyone could have done it for the last 40 years, but let us see your version.
Let's go back to this truly exceptional design and look at it more carefully. Mind you, anyone could have done it for the last 40 years, but let us see your version.
Input bias current compensation requires a really good knowledge of the sources of the input current and how it changes over temperature voltages etc. Tracking is also a challenge. Its done on precision voltmeters but through some circuitry that would not work well in an audio product (biasing the input to ground resistor). I'm not sure it works that well in terms of error in a meter done that way. (I checked the input circuit for the Fluke 8810.)
And its a noise source that needs to be mitigated.
Why is a bias current through a cartridge a great sin? Will it cause distortion? And does anyone believe that a cartridge demagnitizer really demagnitizes a cartridge? Would you really want the platinum magnet in your obscenely expensive Olympos demagnitized?
And is bias current the real issue between a bipolar and a Jfet?
And its a noise source that needs to be mitigated.
Why is a bias current through a cartridge a great sin? Will it cause distortion? And does anyone believe that a cartridge demagnitizer really demagnitizes a cartridge? Would you really want the platinum magnet in your obscenely expensive Olympos demagnitized?
And is bias current the real issue between a bipolar and a Jfet?
When it comes to autos, let me inform our European and Asian audience that 80mph is a very high speed here. This means that autos do not have to effortlessly do 100 mph or more on the Autobahn, because we don't have any. American cars and their motors are designed around this. They usually have high torque, pretty good HP, and are large displacement. However, if you modified a Jag and went on the autobahn, you would find its compromises.
Because my friend, Jack Bybee, has lived all over the world, he has a modified British car, but he did not replace the engine. He just modified it to get 651 HP, with the original engine. It works well enough here, but it would work well, anywhere in the world. That is what is called real engineering, rather than hot rodding.
Because my friend, Jack Bybee, has lived all over the world, he has a modified British car, but he did not replace the engine. He just modified it to get 651 HP, with the original engine. It works well enough here, but it would work well, anywhere in the world. That is what is called real engineering, rather than hot rodding.
The simplest current mirroring circuits work pretty well, there is nothing special here. Beta matches well enough to count on a 50X reduction in Ib. It is thermodynamicly impossible to replicate the current without increasing the noise.
Modern bipolar transistors work fine even at nano-amp collector currents.
Modern bipolar transistors work fine even at nano-amp collector currents.
A 105mph trip on the far right lane of the M20 from London to Dover last friday was pretty boring.
Doing 40 average on twit curvy roads through a chicken village every 5 miles in West Sussex is rollercoaster grade however.
The Silver Shadow MKI is for tasteless pimps, why bother with a nitrous oxide injection blown Hemi in a Sjeique mobile if you can have 1200HP from a racy truck engine.
A 750 horsepower Camaro HPE750 for $110k is likely cheaper than just the supercharged engine behind the naked Ho'od chick, and i wouldn't expect any luggage room left in the trunk of the Roller.
Doing 40 average on twit curvy roads through a chicken village every 5 miles in West Sussex is rollercoaster grade however.
The Silver Shadow MKI is for tasteless pimps, why bother with a nitrous oxide injection blown Hemi in a Sjeique mobile if you can have 1200HP from a racy truck engine.
A 750 horsepower Camaro HPE750 for $110k is likely cheaper than just the supercharged engine behind the naked Ho'od chick, and i wouldn't expect any luggage room left in the trunk of the Roller.
Unfortunately, the point is missed completely. If you can make something that performs extremely well, yet maintain a practical, and esthetic look and feel, then you have made something special.
I'm reminded of what the priest in Samurai Trilogy' tells Musashi (Toshiro Mifune), that there is a difference between a Samurai Warrior and a tough man.
You might say that putting an American motor in a European car, that will most probably upset its balance and even physical aesthetics, is more the tough guy approach.
Paralleling fets endlessly can produce a very low noise input stage, but it lacks everything else. For example, how do you couple to the next stage? Cap, transformer, bipolar transistor, etc?
There is no inherent DC stability or even order distortion reduction. It would appear that negative feedback is nearly impossible, without effectively shorting the input,(not always the best effective load), so watch out for CCIR IM distortion due to the RIAA network amplifying the inherent distortion of this stage out of line from typical expectations. 19.9KHz and 20KHz should do nicely.
I'm reminded of what the priest in Samurai Trilogy' tells Musashi (Toshiro Mifune), that there is a difference between a Samurai Warrior and a tough man.
You might say that putting an American motor in a European car, that will most probably upset its balance and even physical aesthetics, is more the tough guy approach.
Paralleling fets endlessly can produce a very low noise input stage, but it lacks everything else. For example, how do you couple to the next stage? Cap, transformer, bipolar transistor, etc?
There is no inherent DC stability or even order distortion reduction. It would appear that negative feedback is nearly impossible, without effectively shorting the input,(not always the best effective load), so watch out for CCIR IM distortion due to the RIAA network amplifying the inherent distortion of this stage out of line from typical expectations. 19.9KHz and 20KHz should do nicely.
- Status
- Not open for further replies.