OPA858: Gain Switching for OPA858

Hello Christian,

  Would you be able to post a schematic?

  Usually, high bandwidth parts in combination with mux/switches may be difficult to tune stability-wise due to these components additional additional parasitic to the feedback of the amplifier. 

  You mentioned that you have read other threads, but I will post some below for reference which contain information on how to build these models in simulation and discussion on avoiding instability in these types of applications: 

1. https://e2e.ti.com/support/amplifiers-group/amplifiers/f/amplifiers-forum/1074041/opa859-tia-with-gain-switching
2. https://e2e.ti.com/support/amplifiers-group/amplifiers/f/amplifiers-forum/754286/opa657-multi-gain-stage-transimpedance-amplifier-using-analog-switches
3. https://e2e.ti.com/support/amplifiers-group/amplifiers/f/amplifiers-forum/1074464/the-resistor-at-the-input-of-the-trans-impedance-amplifier/3977993?tisearch=e2e-sitesearch&keymatch=%20user%3A466066#3977993

Thank you,

Sima 

Hi Sima,

thanks for the quick answer. Okay these are some threads I haven't seen yet. But it seems as if we already tried some of the mentioned hints without success. Will have a look through the others. Anyhow it seems as if you do not recommend a combination OPA858 or another derivat out of that series and analog switches.

Could one option be to turn the parallel gain switching into a series one with shorting out some resistors to minimize feedback loop?

I have attached an excerpt of our schematic showing the siginificant parts. Hope it helps to get an idea of what we want to do.

We have also played around with OPA857 which as an integrated gain switch but unluckliy this parts is not that fast and there is no chance to get gain over 20k.

If there is any further idea worth to try please let me know.

Thanks in advance.

Christian

OTDR_REC_EAVL.pdf

Hallo Christian,

the issue is that you have installed the switches at the -input of OPAmp. By this the switch capacitances destabilize the OPAmp. A better approach is to mount the feedback resistors directly to the -input and the switches to the output. This will isolate the switch capacitances from the -input of OPAmp. Eventually, an additional small isolation resistor directly at the output of OPAmp could be used to further increase stability. You could also add individual phase lead caps in parallel to some of the feedback resistors.

Another issue is the TMUX1112 itself. It shows very high switch capacitances and doesn't seem to be well suited in your application. Take a special HF MUX or special HF analog switches with ultra low capacitances.

Kai

Hi Kai,

thanks for your input. Is there any MUX you could recommend for that issue in TI's portfolio? I thought this one was also a good choice regarding low leakage and other parasitic parameters.

Up to know I have manged to bring HF- Ringing to a minium level but still there as a peak- peak value auf 100mV. For me it seems that there is no chance to get rid of that just by adding components to tune in the current (layout-)setup. Maybe it is worth trying to change Mux to see its overall influence.

Anyhow I also read about a new integrated part OPA3S2859. On the technical point of view this part seems to be well suited for what we want to do. Would you agree with me in that case?

But unluckily the part is unavaible at the moment and I am quite sure that based on part's pricing there will be big discussions with our purchase departemt :-) Are there any plans to release a part with one channel only so that the price could drop?

Regards

Christian

Hi Christian,

I would look at the TMUX1511.

Also, I would think about removing the ground plane underneath the signal routing at the -input of OPAmp.

Kai

Hello Christian,

  Did the TMUX1511 help with eliminating the HF ringing?

Thank you,

Sima

Hi Sima,

thanks for aking. I had to make a redesigned board to test it so I have to wait unti i can use that for testing.

But what helped to stop hf ringing was to replaced OPA858 by its similar type OPA859. I know it is quite slow in comparsion to OPA858 and we still have to check if this setup will fulfill our requirments but it maybe one possible solution. But i will also test to use OPA858 in the optimized setup.

Anyhow I still need some time to try but because of other priorities testing will have to wait for a while.

Will let you know the final results.

Regards Christian

Hallo Christian,

when you send us the schematic we can help to perform a phase stability analysis.

Viel Glück

Kai

Hello Christian,

  That is a good debug method. The reason it could be stopping the hf ringing with the OPA859 is because like you said due to its lower bandwidth. The loop-gain crossover (AOL and noise gain intersection) occurs sooner (lower in frequency) which could be enough to increase stability of the circuit. I hope the board redesign fixes the issue!

Thank you for the updates, 
Sima