Opamp Having Oscillation

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

I am running off a single rail 12V and playing around with this type of differential output design, but the the opamp got oscillated badly. a SAW oscillation wave at 250kHz +/-. The opamp is 4558. What's actually goes wrong:confused:

P.S. I tried to ground both of the input, but it still oscillate.
 

Attachments

  • osc.JPG
    osc.JPG
    38.6 KB · Views: 193
Last edited:
Err, uh, I think I just got the negative and postive feedback stuff excatly the wrong way round... My bad.

Anyway, a feedback setup like this, as neat as it looks in simulation, is potentially prone to parasitic oscillation. What do you use for rail decoupling in the practical circuit? The simulation gives you perfect, stiff supplies and ground after all, which is not necessarily the case in practice...
 
If you want made fully differential opamp, then give up your schematic and try to develop something like this:
8yttl.png


Later Edit: @Mooly, U2 wants to be a virtual ground generator but can not perform this function if is powered unbalanced. So, you need a supply voltage like + /-15Vdc, not only +12Vdc.
 
Last edited:
If you want made fully differential opamp, then give up your schematic and try to develop something like this:
8yttl.png

Ya, i read through this in the earlier time. But I don't like it with a reason (perhaps it's my personal reason) :

1. I am running on single rail. I need to bias the +in into midrail. And with it, my Source need to seen into the midrail's resistors during the path into the +in point.

2. And the simulated result doesn't seem like it's in similar amplitude for both output when I am running off from a SE source. :confused:
 

Attachments

  • BAL BAL.JPG
    BAL BAL.JPG
    50.4 KB · Views: 100
  • BAL BAL LTSPICE.asc
    2.2 KB · Views: 25
Last edited:
Then it will not work as it is.

I think I didn't stated up my goal properly in the early stage. Sorry for it.
I would want it to be a fully differential opamp. This is a design that I got it from the google image, and it looks simply super brilliant(to me).

To make you further clear about what I want, I change it into dual rail, and you can easier to read what's U2 actually doing in the circuit. It's not just simply providing a vref/Vmid into the non-inverting pin.

U2 is not under open loop condition, it's under unity gain by taking the output from U1 and U3 (either one act as input and output. ) and feedback into the inverting pin. (As what I can read)

Under this design, no matter which input is having larger or smaller signal, the output differential signal will still be 'compensate' by the U2 and giving the same output amplitude from U1 and U3.

what's making me suspect now is maybe the 4558 is not stable under unity gain and causing it oscillate badly, I guess.
 

Attachments

  • FULLY DIFFERENTIAL.JPG
    FULLY DIFFERENTIAL.JPG
    25.1 KB · Views: 96
Last edited:
Hi,

I´d say the reason that Your circuit of #1 doesn´t work is due to U2 working virtually open loop.
U2 is intended to be just a lowimpedance voltage source (voltage follower) suppling the potential it finds at its positive input 1:1 to the inputs of U1 and U3.
As such U2 needs a direct feedback path from its output to its negative input with a feedback resistor of value (R7||R8).

This structure is just a simple 2-channel gain stage with a common Bias offset generator.
If You gnd the Input at C2, U3´s Output will be simply the DC Offset supplied by U2.
It certainly is not a fully differential amplifier like the cited OPA1632 where You may ground any of its inputs and still create a differential output.

jauu
Calvin
 
Hi,

I´d say the reason that Your circuit of #1 doesn´t work is due to U2 working virtually open loop.
U2 is intended to be just a lowimpedance voltage source (voltage follower) suppling the potential it finds at its positive input 1:1 to the inputs of U1 and U3.
As such U2 needs a direct feedback path from its output to its negative input with a feedback resistor of value (R7||R8).

This structure is just a simple 2-channel gain stage with a common Bias offset generator.
If You gnd the Input at C2, U3´s Output will be simply the DC Offset supplied by U2.
It certainly is not a fully differential amplifier like the cited OPA1632 where You may ground any of its inputs and still create a differential output.

jauu
Calvin



Hi Calvin.

The circuit actually work, and tested to work. the problem is , it's not stable and oscillating.

The U2 is not completely use for creating the potential for DC level of the Vout+ and Vout- ONLY. The R1 and R2 is there for it's reason in this design.

"U2 needs a direct feedback path"
I understood what you and mooly trying to say, you are creating a low impedance 'virtual ground buffer' there by doing so. But it's not the purpose of doing it here, in this design.

This structure is just a simple 2-channel gain stage with a common Bias offset generator. - No, not that simple in this design.

"It certainly is not a fully differential amplifier like the cited OPA1632 where You may ground any of its inputs and still create a differential output."
Yes this circuits works that way, completely and very fine. Just the stability issues. I am figuring, what's and which stage actually make it not stable and oscillate.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.