Low noise amplification of +-500nA to +-4.5V, >100 SNR, 15MHz BW

Hi Anders,

what is the detector capacitance of photodiode?

Kai

A very relevant question. The photo diode has a capacitance of about 70pF at 5V.

Hamamatsu S12968-4

Hi Anders,

Somehow I am not able to find datasheet for Hamamatsu S12968-4. Do you know the size of sensor area?

Question:

I am having trouble designing an amplifier circuit for a photo diode. I have very little light, so even with the best SiPN diodes, I only get about 500nA current from it.

You may need to both optical and electrical help in detecting the light signal.

1. Are you able to place an optical lens to collect more light into the sensor? For low light level condition, smaller sensor area will be better, if you are able to focus the incoming light into the sensing area. I assumed that that incoming light source is collimated.

2. What is the wavelength spectrum range of incoming light source? different sensor has different peak detect efficiency. You need to pick a sensor with the maximum response in the interest of your light spectrum. Also, sensor may have filter in front of it. You would prefer to minimize light absorption through various optics under extreme low light condition. 

3. You may need to use chopper TIA for get rid of 1/f noise (unless the incoming light source is mechanically chopped).  

4. Increase the light integration time, if possible. 

5. Get a better Op Amp for the sensing application. Please let us know more about the current sensing ranges from your detector, and we may be able to simulate it and provide you with the best recommendation. 

Best,

Raymond

Hey Anders, so with 70pF, 9Mohm gain and 15Mhz target, you need a GBP of 22GHz to do this in a single stage - probably best to break it into two stages

using the LMH6629 as a transimpedance stage, targeting 20MHz, I can get 20kohm first stage gain, means I need 450X more gain. 

Maybe the OPA855 should be applied at the first stage, with that I can get 45kohm gain - are you sure about the 70pF, pretty high for a 15Mhz target. 

Hi Raymond, thanks for your input!

The sensor area is 3.2x3.2mm. We have a laser beam that is very small compared to the area, but it is beening scanned over a area, so the size is needed.

1. This is an option, thou very expensive. We might be able to narrow down the beam to a smaller area, but not by much and prefferably not at all.

2. The wavelenght is deep UV light. And the detector has been selected not for its peak at just our wavelenght, but for the maximum responsability compared to others at just our vawelenght. Just to say, case has been taken...

3. I cant find much information regarding chopper TIAs. The light we have is indeed mechanically chopped. I guess? The 15MHz comes from the light sweep speed and the flanks of mechanical changes we want to detect.

4. I guess since we have 100%, this cant be done. With the 15MHz BW and seeep and such...

5. Yes, this is what I want. I can pay what ever. Price is not the issue, just the best solution that makes it work.

Thanks again!

Hi Michael,

From where did you get the GBP equation? I would like to read more about it.

I am currently doing two stages of 4GHz. Stages are not my limit. I am then finishing off with a AD8139 diff amplifier of 2-4x. Using two transimpedance stages based on LTC6268-10. For my evaluation..., that did not do it. Way too much noise.

We were first aiming for 15pF, but the sweeping needed to increase the detector surface. The capacitance is 70pF according to datasheet.

Price for the solution is not an issue here. We tried PMTs, and they were too noisy, even the best Hamamatsu we could get.

Thanks for your interest!

Hi Anders,

Your sensor seems to be a larger than what I expected for laser application. I would think that it should be 1mm x 1mm or slighter larger. 

Question:

1. This is an option, thou very expensive. We might be able to narrow down the beam to a smaller area, but not by much and prefferably not at all.

Yes, you will need sapphire or UV fused silica or MgF2 optical lens to handle this application. 

2. The wavelenght is deep UV light. And the detector has been selected not for its peak at just our wavelenght, but for the maximum responsability compared to others at just our wvaelenght. Just to say, case has been taken...

Since this is deep UV, I think that the sensor window material may absorb over 50% of incoming light (check with sensor's spectrum response). You will be better off to remove the sensor window and keep the optical compartment clean. This may increase input incoming light intensity and signal strength by 50% or more. In deep UV region, I assumed around  200nm region, there is no other light spectrum interference. The window material does not help you, if your incoming light signal is very weak.  

BTW, Do you purge your incoming beam path to improve the enviromental conditions - humidity and dust particles, both will attenuate (absorb and scatter) your incoming light signals in deep UV application. 

3. I cant find much information regarding chopper TIAs. The light we have is indeed mechanically chopped. I guess? 

Mechanically chopped light signal will be equivalent to electronic chopped type, which it is designed to remove 1/f noise. Since 15MHz comes from light sweep, so you do not have time to increase the integrate longer. You may not require chopper amplifier, if the incoming signal is chooped already, but you will definitely need low noise, low offset and low bias current TIA op amp.

Let me think about it Michael's suggestion, and find a solution for you. Can you finalize the min. and max. current ranges from your detector and application? You may need to have two gain stages to increase your output signals. 

By the way, have you consider photomultiplier or solid state photomultiplier as the deep UV sensor. 

Best,

Raymond

Hi Anders,

Could you post the Hamamatsu S12968-4 datasheet? I can't find it over the internet. 

After reading your application description, I am starting to understand the difficulties of the application, which is a complex one. As I indicated in my previous reply, you will need to improve the system setup and configuration in multiple fronts.  To have better Op Amp in TIA Op Amp is only one part of your solutions, especially you have tried other options. 

Among TI exerts, I believe that we are able to help you to solve your application issues. It will be good that you can provide us with more information about your design and things you have tried. This can help us to solve the issues more quickly, and avoid walking through wrong paths in doing so. 

If you want keep the information private, we may exchange the information via private E2E email. Please let us know. 

Best,

Raymond

Raymond,

1. The lens is an option if we need to make the sweep area smaller. We keep that as a backup if all other fail.

2. Everything is monitored, moist, temperature, dust. Its important as you say. The detector is for DUV, so the window is adopted for it, still not saying its not removing much of the light. Ill have a look at it, what the manufacturer sais.

3. The maximum signal we have been testing with is 500uA. So that must be close to maximum output signal. Two gain stages or more is ok. Perhaps an active filter stage for removing the 1/f noise.

We tried a number of photo multiplier, but the SNR was not good enough on any of them. We are about to evaluate SiPMs, but very long lead time on them right now for even a sample.

Datasheet:

s12698_series_kspd1084e.pdf

I want to send you the schematics of what we tested with the TIA path. How do I get your e-mail?

Hi Anders,

You may send me a friend request via E2E, and then we are able to establish the communication via private E2E email. 

Best,

RAymond