transimpedence amplifier

Hello Valmik,

Yes, the OPA656 is a good choice for your photodiode application.  I can help you with the design.

Your system may also require an additional amp stage before the input of the ADS931.  The reason for this is that the input to the ADC acts as a changing capacitive load, and a changing capacitive load could cause stability issues in the transimpedance stage.  Inserting an OPA698 in between the transimpedance stage and the ADC will guarantee consistent stability for the circuit.

I can put together an example simulation for you.  Can you tell me any more application details for the OPA656 stage (targeted bandwidth, ac vs. dc coupling, etc)?

Regards,

Nick

Applications Engineer - High Speed Amps

Hi Nick, thank you for your intrest.

I have been working on this problem for some time now. Actually the bandwidth  we are targeting is 8 MHz. One of my professors told me that OPA656 may not have sufficient bandwidth when used as a transimpedence amplifier with some gain. so i was considering using OPA659. For the photodiode we are using HR-040 from OSIoptoelectronics which have a sensitivity of 0.45A/W. I am not very sure about the kind of signal being fed to the photodiode. It will probably be from a 650nm laser. Also, 4x4 system was too complex so right now we are using only 4 individual photodiodes. Later if this works fine we may scale it up.

Thank you.

Valmik.

Valmik,

The following application note should guide you in the selection of a transimpedance amplifier:  http://focus.ti.com/general/docs/litabsmultiplefilelist.tsp?literatureNumber=sboa122.

Depending on the transimpedance gain you want to achieve and the photodiode capacitance, the OPA656 may be sufficient for your application.  Alternate device you may want to look at are: OPA657, OPA846, OPA847.  Although the OPA659 has higher bandwidth than the OPA656, it does not have that large an advantage on gain bandwidth product.  The OPA659 also has a much higher input voltage noise that would result in lower dynamic range.

From the photodiode, you will need its capacitance.  With that at hand, the application note can guide you on how to select the best device for your application.  For spice model evaluation, I would recommend using the OPA657 rev E model for simulation, as it is elaborate enough to provide realistic simulation results, in particular, compensation.

Let me know if you have any problem using the part you select.

Thank you for your help Xavier

The photodiodes we are using have a capacitance of 4.9pF and a dark current of about 2-2.5nA. The maximum output we get from the photodiode is about 350nA. I looked at your suggestions and i think the OPA846 should work well for us. I was thinking about using ADS931 for the ADC. Let me know what you think about it? Also, I am using Cadence Design Entry CIS for pspice simulation. But i am not able to run it. could you suggest something for pspice simulations.

Valmik,

For the ADS931, what resolution do you want to achieve at a system level.  The ADS931 is an 8bit ADC, its maximum resolution with a 1Vpp input is going to be 3.92mV.   Since the maximum clock frequency is 33MHz and to achieve decent time step in an oversample system you will need 10x the number of points, the fastest signal you will be able to see is going to be ~3ns.  If both the resolution and the time step are acceptable is acceptable, the ADS931 is a good choice.

For the transimpedance design, before looking at stability, you need to know the maximum peak amplitude that you are trying to achieve.  Using the maximum amplitude of the ADS831 of 1Vpp, the transimpedance gain would need to be 2.86Mohm.  If the 2Vpp input range of the ADS931 is used both the maximum resolution of the ADC will double and the transimpedance gain requirement.  Looking at figure 6 of the app note, the OPA846 will not be acceptable for transimpedance gain above 200kohm.  Note that figure 6 applies as you total input capacitance, photodiode + amplifier parasitic, is close to 10pF.  Only the OPA657 will be usable for these type of gains.  the bandwidth requirement of 8MHz will have to be verified.

Also the 8MHz bandwidth requirement does not seem to match the 3.3MHz maximum bandwidth signal you can resolve (10x samples) with the ADS931.  An 80MHz+ ADC may be better.

Could you describe the nature of the signal you are trying to measure?  What time step resolution do you need to achieve?

Hi Xavier,

The problem is that my professor, for some odd reason, does not give me the exact specs for the input signal. So assuming that the output from the photodiode will have a maximum amplitude of 350nA and the maximum frequency component of the input signal will be 2MHz , can you suggest some combination of transimpedence amplifier and ADC. The ADC has to run at atleast 10MSPS and preferably have INL or DNL less than or equal to 1.

thank you

valmik

Hi Valmik,

The ADS931 will work for the ADC.

For the transimpedance amplifier, consider using the OPA657 since your gain has the potential of being so large.

I tried to use the pspice model for the OPA657 available on the website, but it did not work. I am using Cadence design CIS for pspice simulation. Could you suggest a solution.

thank you

The pspice model for the OPA657 from the web site (Rev. E) was developed using Cadence 14.2 and as such should be compatible with Cadence design CIS.

What is the error message?

the problem is that OPA657 is not available at newark or digikey and rochester is not sending a quote. could i use something else instead of opa657 assuming the input signal has a range of 2nA to 350nA, and bandwidth of 3Mhz and i am using a tranimpedence gain of about 2.5-2.8 Mohms. i will be using ADS931 for my ADC.