Jfet folded cascode with diamond buffer opamp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi all.

Well i thought it was time to do some new stuff. :)

So my first attempt is to make an opamp replacement with jfet inputs.

I will try to hit the 700x700mil size for this one.

This will be fun.

The jfet has an input noise level of 6nV so it will be able to be used for riaa as well as linelevel.

I have attached the diagram. I expect to have the board ready by tomorrow.
 

Attachments

  • jfetopampv10.pdf
    29.3 KB · Views: 1,402
Hi Sonnya,

Nice circuit.

Question . Does cascoding the folded cascode help much? I would think it's use in this application would be minimal but I could be wrong.
I like the use of the Wilson Mirror which I believe is far better than the Widlar that is often used.

Keep up the good work.

Regards,

Jam
 
Last edited:
Hi Jam.

Thanks for the comment.

The widlar currentsource is made for the silicon process as it is not possible to make high ohm resistors in the silicon. By adding the resistor in "mirror" transistor it is possible to compensate for it. Correct me if i am wrong.

I can see that i maybee has been "blinded" by the use of folded cascode as i have used them before.

I would be easy to replace it by two wilson mirror and it would require the same or less transistor. It would actually raise the impedance in the tranconductance stage, like i wanted it to.

So i will change that part. thanks for the guiding question.
 
Sonnya,

Actually I like the four transistor Wilson Mirror (Still might be a bit of overkill).
I will not disagree with you that the folded cascode can sound very good and only requires matching of the diff and current mirror offers equal pull-up and pull-down. All in all a very elegant circuit.
 

Attachments

  • Imp_wilson_cs.gif
    Imp_wilson_cs.gif
    2.5 KB · Views: 1,444
Last edited:
jfet opamp module simulation data

Hi all.

I have attached the simulation data.
 

Attachments

  • 101xgain.jpg
    101xgain.jpg
    247.3 KB · Views: 626
  • 11xgain.jpg
    11xgain.jpg
    238.3 KB · Views: 680
  • 2xgain.jpg
    2xgain.jpg
    245.8 KB · Views: 706
  • CLGx1+100pFComp.jpg
    CLGx1+100pFComp.jpg
    237.2 KB · Views: 464
  • CLGx1+100pFComp+220pF.jpg
    CLGx1+100pFComp+220pF.jpg
    247.5 KB · Views: 422
  • CLGx11+1nF.jpg
    CLGx11+1nF.jpg
    251 KB · Views: 82
  • CLGx11+47pFComp+220pF.jpg
    CLGx11+47pFComp+220pF.jpg
    253.4 KB · Views: 88
  • OLG+1nF.jpg
    OLG+1nF.jpg
    257.6 KB · Views: 105
  • OLG+470pF.jpg
    OLG+470pF.jpg
    257.3 KB · Views: 90
  • OLG+220pF.jpg
    OLG+220pF.jpg
    253.1 KB · Views: 92
The sim data should speech for them self.

The supply voltage is +/-15V

When compensation is added it is added from "COMP" to V-.

Source:50Ohm.

Feedback for x11:10K+1K
Feedback for x101:10K+100R
Feedback for x1:1K

The picture "CLGx11-47pfcomp-22pF-1vppstep" is a 220pF load and not 22pF
 

Attachments

  • openloopgain.jpg
    openloopgain.jpg
    254.7 KB · Views: 129
  • CLGx11+1vppstep.jpg
    CLGx11+1vppstep.jpg
    194.8 KB · Views: 102
  • CLGx11+220pF+1vppstep.jpg
    CLGx11+220pF+1vppstep.jpg
    198.3 KB · Views: 74
  • CLGx11+47pFComp+22pF+1Vppstep - sim B.jpg
    CLGx11+47pFComp+22pF+1Vppstep - sim B.jpg
    183.3 KB · Views: 76
  • CLGx11+47pFComp+220pF+1Vppstep - sim A.jpg
    CLGx11+47pFComp+220pF+1Vppstep - sim A.jpg
    198.9 KB · Views: 77
Last edited:
There are two versions of it.

The first one, which you refer too with folded cascode and the one with wilson mirror in post 7 (Easier to implement on the small pcb).

I have tried with a LED as D2 as weel for simulation. And there is no difference. But we also have .65V CE drop across U5-A and U5-B. This also works good for current mirrors and hawksford cascode which i use in my CFB amp.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.