LMH6722: About Output offset

Morning Tsukui-san

It looks like you are treating the LMH6722 like a voltage feedback op amp, it is not, it is current feedback - it is not well suited to what looks like a transimpedance design. 

Even if you get the DC output right, this circuit will almost certainly oscillate, Need to probably go to decomp VFA if your source is really a current source. 

Well here is an early CFA dangers app note I wrote, to recommend a part, would need exact requirements, 

http://www.ti.com/lit/an/snoa367c/snoa367c.pdf?keyMatch=OA-15&tisearch=Search-EN-everything

Hello Yusuke,

1. The THS4012 is a voltage feedback amplifier not a CFA. Also CFAs can be used in an inverting configuration, they aren't recommended for trans-impedance designs due to low input impedance on the inverting node. For a current source input, it would be better to switch to a VFB amplifier. 

2. Another issue that is probably affecting your TINA circuit is your use of a 200nF feedback capacitor. Stability for a CFB amplifier works differently and is mostly dominated by your feedback resistor. 

I would recommend going through the current feedback precision lab series. This covers how CFAs work, their differences to a regular VFB amplifier and how to check the stability of your circuit on TINA-TI.

Also if you could clarify your design specs and what you are trying to achieve, I could help with recommending another part/design.

Best,

Hasan Babiker

Well Yusuke, 

You did not reply with customer requirements? So , your question of why the THS4012 "works" and the LMH6722 does not is misplaced - there might be 2 to 3 different levels to that question but at essence what you are trying to do with the LMH6722 CFA is not suitable to CFA, but is to VFA like the THS4012. 

Too open ended to answer, what is the target requirements?. 

'Also, Hasan's reply that the CFA does not work because of low inverting input impedance is a long running misconception - must be in textbooks that were written too long ago to bother correcting. 

Open loop the inverting V- node is low impedance but once you close the CFA loop the loop gain drives the error current to zero so it effectively draws no signal current. 

CFA can be used in transimpedance under very constrained conditions - it has stability problems with a source C and cannot directly use a simple cap feedback to stabilize like a VFA - not worth explaining in detail as it is the wrong way to go usually. 

Micheal,

Thank you for the correction, that is a good point.

Yusuke,

In summary we need more details on your design in order to move forward.

Best,

Hasan Babiker