Fully balanced Opamp

[rkn]

Hello,

I'm not sure where to post this, but this project is inspired by Nelson Pass' article about Opamps so I guess it is the right place here. 😉
My idea is to build a discrete opamp with balanced inputs and outputs. My design follows mainly Nelson Pass' article but without the feedback loop. This can be attached externally if wanted. I'm going to make a seperate PCB for the discrete Opamp. Below you can see my design. I used JFETs inputs and a bipolar stage on every side.
Maybe the values I've selected are not good so comments regarding the design are welcome!
one point where I'm not sure is wether I should use two resistors in series with the jeft inputs or not. In Nelson Pass article you have these resistors but if you look on other discrete Opamp designs you don't have them.
I connected the three constant current sources to one reference. i hope it is allowed to do that. I've then that in some designs of power amplifiers but only for two current sources not for three like in his design.

An externally hosted image should be here but it was not working when we last tested it.

Greetings,

Raphael

 

[rkn]

Sorry, I forgot to refer to GRollins and his reverse engineering of the XVR1! As you can see the design shows some similarities to his design. I apologize for forgetting him! 🙂

Raphael

 

[rkn]

Ok, added some further stuff and changed some values. The "biggest" change is, that there is now a current mirror inserted into the dif pair.

Greetings,

Raphael

 

[jam]

Schematic?

 

[rkn]

Schematic? You shoul dbe able to see the first version in my first post.
For the new version with the added stuff I#ll post the schematic later this day when I'm back at home.

Raphael

 

[GRollins]

Before worrying about parts values, you're going to need to make at least one change. Each output will need a resistor to ground, feedback to the front end, or both. As it stands now, you'll have uncontrollable DC offset and horrendous amounts of gain (on the order of 60-80 dB).

Grey

 

[rkn]

Well, the idea was to place these feedback resisistors on the mainboard where I want to plug in my OPamp. But yesterday I found out that I have some space left on the OPamp PCB so that there is no big problem to place two additional feedback resistors.

Raphael

 

[Tim__x]

You need a common-mode feedback loop.

 

[GRollins]

It wouldn't be any problem to run the circuit with zero feedback. There are other ways to attend to the DC offset. But if feedback was the original intention, then that's easily enough arranged.

Grey

 

[rkn]

Ok, here is the new version. I added feedback resistors like Nelson Pass demonstrates it in his paper and I removed the 1M00 to gound at the inputs because I think they are not needed when I add feedback like now.
Comments?

Raphael

An externally hosted image should be here but it was not working when we last tested it.

 

[hifimaker]

Hi Raphael,

It looks like you are using a MAC. What schematic capture software are you using? I'm a recent convert to the iMac. What an awesome machine with OSX 15.5!

I cannot wait to get completely off Windows, and now need a good Mac based schematic caputre program.

Thank you,

-David

 

[uli]

IMO it wont work with both sides of a current mirror 😉

Uli

 

[jam]

Hi Uli,

Long time no see................you are right the current mirror has to be replaved with active loads or resistors. John Lindsey Hood has a design for a symetrical currernt mirror that might work in this application. It was publisrd in Wireless World years ago.

Regards,

Jam

 

[uli]

Jam, old fella 🙂

did they pardon you, didnt know that! Happy about that!

Regards Uli

 

[Nelson Pass]

The bigger the fall, the greater the redemption.

😎

 

[rkn]

OK, bad idea with current mirror. With a little bit thinking about it I night find it out myself... I'll return bac to the first verison with the resistors.

David, I'm using gEDA: www.geda.seul.org Originally it was written for Linux but after MacOSX arrived it is also possible to run it on a Mac. You'll find a MacOSX-bundle on the website.
I stopped working with Windows and changed to Linux two or three years ago. And I don't want Windows back. But after some bad experiences with HP I switched to Apple hoping that they sell a more reliable hardware. I like Apple. I've a Mac mini and a MacBook. Now I'm using both in parallel: Linux and MacOSX and cannot say which system is better, both have pros and cons. Ok that's offtopic.

Raphael

 

[rkn]

Here is now the next version of the schematic. I removed the current mirror and added feedback.

Raphael

An externally hosted image should be here but it was not working when we last tested it.

 

[forr]

Hi Uli,

Long time no see................you are right the current mirror has to be replaved with active loads or resistors. John Lindsey Hood has a design for a symetrical currernt mirror that might work in this application. It was publisrd in Wireless World years ago.

Regards,

Jam

Hi Jam and Rkn,

See
http://www.diyaudio.com/forums/showthread.php?postid=760090#post760090
schematics :
http://www.diyaudio.com/forums/showthread.php?postid=760486#post760486
Also see following comments.

Regards.

 

[Nelson Pass]

You will require some mechanism for adjusting and/or controlling the
absolute DC value of the outputs.

 

[rkn]

For what? For my version or if I want to use the John Lindsey Hood current mirror?
What aboout an DC-Servo?

Raphael