I am playing around with the Idea of making a descrete component op amp of modest complexity yet still retain all the Virtues of an IC monolithic Op Amp.
I would like some ideas on this type of op amp I was thinking about a SMT Board that would plug into a DIP 8 OP Amp socket.
I would like some ideas on this type of op amp I was thinking about a SMT Board that would plug into a DIP 8 OP Amp socket.
Attachments
Discrete Opamp
Hi, This has already be done by Sonny and me.:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=1794&highlight=
Hi, This has already be done by Sonny and me.:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=1794&highlight=
Yes indeed, but there is one thing you wont have with modern opamp - simplicity and the ability to adjust exactly the one or two parameters that you consider important to your application. I think it is proven enough in audio that the rule "keep it as simple as possible, but no simpler" applies.
What I aim at is a discrete stage with dif pair input to be used right after a balanced out voltage output DAC chip.
I would especially like to try a solution where there is no global feedback for this discrete stage. I have an assumption that it might sound better than a stage with feedback and LP filter put together that is usually used.
Who knows if I suceed but at least I'll try
I'm smarter than before doing it no matter what the physical outcome is - and thats already one of the two goals achieved.Ergo
Thanks for all your comments the Elso Kwak Schematic referred to was not shown in my search however another one showing a Clone of an AD825 however rescaled to be used with Loudspeakers. From the Thread it appears this is available in a Drop in DIP 8 Board? IN addition this is slightly more complex than I was wanting and I don’t need an output stage in my application, However if the Board has provisions for it I world just leave out the components for the output stage. Like Sonny I als like jfet current sources however I am having great luck with using the Wilson Three transistor current Mirror as a current generator, I will have to do an in-depth comparison between the two
Peranders I understand that I am not going to get 200 MHz Bandwidth out of this however there are Plenty of all discrete component op amps around that are sounding quite good Like The Jensen JE-990 that dose 10 MHz and is extremely low noise. I nave nothing against op amps I have used just about every type that has come down the Pike since the LM-709 My first preamp (1970) Used a uA741 in a day when discrete were the norm and while not the most stellar in upper midrange and Treble performance was a lot better than some other reasonably priced Preamps of the Day. So why am I doing this cause I am board with op amps kind of been there done that way too many times.
Ergo I surly would like to see your Design when its completed. Also you are correct in that one can optimize a discrete component circuit to the application something that is not readily possible with Op amps intended for a verity of uses. Due to the compromises of Monolithic fabrication the best circuits get seriously degraded when implemented in monolithic form IMHO. Leakage currents between devices and the sub in addition to the difficulty of obtaining close tolerance resistors without Laser trimming so the circuit is designed from the onset as a compromise. Op amps are also all not built on the best processes while some of the old Classics like the OPA627 and some new devices are upon the Dielectric isolation process dramatically reducing leakage currents and capacitances allowing better performance that if the circuit were built with say Junction isolation. AD744, AD712 and a more refined JI process sometimes called CB. AD-829, AD845, AD843 and I suspect the AD825 is also a CB process.
I just completed a far more complex discrete Headphone amplifier that is rather complex and Power hungry and it sure works good so I was wishing to get most of that in a lower power simpler topology and yes the feedback free idea is attractive is the circuit is optimized for linearity Distortion open loop should be around 0.1-0.3%THD Plus noise
Peranders I understand that I am not going to get 200 MHz Bandwidth out of this however there are Plenty of all discrete component op amps around that are sounding quite good Like The Jensen JE-990 that dose 10 MHz and is extremely low noise. I nave nothing against op amps I have used just about every type that has come down the Pike since the LM-709 My first preamp (1970) Used a uA741 in a day when discrete were the norm and while not the most stellar in upper midrange and Treble performance was a lot better than some other reasonably priced Preamps of the Day. So why am I doing this cause I am board with op amps kind of been there done that way too many times.
Ergo I surly would like to see your Design when its completed. Also you are correct in that one can optimize a discrete component circuit to the application something that is not readily possible with Op amps intended for a verity of uses. Due to the compromises of Monolithic fabrication the best circuits get seriously degraded when implemented in monolithic form IMHO. Leakage currents between devices and the sub in addition to the difficulty of obtaining close tolerance resistors without Laser trimming so the circuit is designed from the onset as a compromise. Op amps are also all not built on the best processes while some of the old Classics like the OPA627 and some new devices are upon the Dielectric isolation process dramatically reducing leakage currents and capacitances allowing better performance that if the circuit were built with say Junction isolation. AD744, AD712 and a more refined JI process sometimes called CB. AD-829, AD845, AD843 and I suspect the AD825 is also a CB process.
I just completed a far more complex discrete Headphone amplifier that is rather complex and Power hungry and it sure works good so I was wishing to get most of that in a lower power simpler topology and yes the feedback free idea is attractive is the circuit is optimized for linearity Distortion open loop should be around 0.1-0.3%THD Plus noise
ppl said:
Op Amps are great for high gain but most don't work to well a high freq. mostly due to the fact that there maybe 50-100 transistors in the circuit. This makes for a lot of miller capacitance and will reduce the gain at high freq. For a high gain differential amp, this will limit the harmonics created to cancel out the ones made by saturation(if this is a - feedbak circuit) If you limit the 'signal' transistors to a minimum, your overall slew rate will thank you.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.