Hey all,

I'm new to the forums but I've been browsing the website for a few years now (ever since I started college ) and it has come in handy too many times to count. Anyways, I'm a electrical engineering student working on a senior design project (with a semester left to go) and I have some questions that I can't seem to get answered despite asking senior engineers and lots of google searches. I still have much to learn...

My senior design project is detecting very small holes (micron size) using IR emitters and detectors. Right now I am trying to design the op-amp amplification stage. Our received signal is on the nanoampere scale and we need to amplify this signal up to mV levels. I am thinking a TIA (transimpedance amplifier) would work well.

My question is, needing a gain of about 1 million, what's the highest frequency of pulsed light that an op amp could amplify at 1 million gain?

I have seen op-amp's with GBW of 10ghz but at 1 million gain, the highest frequency that could be used is 10khz. Is this a correct assumption?

Is there a way to have, say, a 100khz signal amplified at 1 million gain?

I hope that's not too many questions. I'm still trying to wrap my head around TIA design, op-amp characteristics, and photodiode specs and its all a bit overwhelming.

 

Our received signal is on the nanoampere scale and we need to amplify this signal up to mV levels.

I am having difficulty with this part. Amperes are for current. Volts are for voltage.

 

The first thing I would do is hit the application notes from a manufacturer of high end op amps such as Analog Devices. Others might be TI/BurrBrown/National Semiconductor, Maxim, etc.

The apnotes will cover principles to know and suggest circuits and particular parts. Most if not all of your questions should be answered with a little research there.

I am thinking a TIA (transimpedance amplifier) would work well.

Analog Devices agrees! See the link and check out the nifty circuit calculator.

I am having difficulty with this part. Amperes are for current. Volts are for voltage.

Photodiode detectors may be thought of as a constant current source whose current is set by incident light. Vary the light, vary the current. They require considerable amplification due to the small currents involved. The op-amp is run as a current to voltage amplifier.

 

Getting 100kHz at 1 million gain is possible. Yes, a single op amp with a GBW of 10GHz only has a bandwidth of 10kHz. So you just need to divide the gain among several amplifiers. For example 4 stages with a gain of 31.6, each with a bandwidth of a little over 100kHz will get you there. This requires an op amp with a GBW somewhat greater than 3.2MHz, much easier to find. Note that you will want a very low noise op amp for the first stage (such as a J-FET input type) and a high order bandpass filter centered around 100kHz to maximize the S/N ratio.

 

I have seen a instrument can measuring down to 1 um, but I'm not sure how is the resolution of that instrument, and the price is really high that it 's over US$28000.

More gain and more BW, the cost is more higher, about the gain that you can using the multiple way to reach up to one million, the definitions of the Gain for the BW is set to 1, so like as the BW of LM324 is 1Mhz, but you only can run the Gain=1 on 1Mhz, if you want to get some more Gain then you have to make the BW lower.

You can using many different ways to reach the Gain up to 10000.
exp: 3 stages as 100*100*100, or four stages as 100*100*10*10, and five stages as 100*10*10*10*10.

The points are to get the high BW and Low Vos Drift, Stable, Low Noise op amps.

 

Thank you all for the helpful replies. I will definitely look for 'application notes'. I have come across a few helpful white papers by the major name manufacturers but haven't seen any of these 'notes' yet. Thanks for the tip!

It looks like multiple gain stages will be necessary where the tradeoff will be added noise and extra components.

 

Plus the power for op amp needs a extra low ripple voltage, and using some filter capacitors more than 10000uf and in parallel.

 

..........................

It looks like multiple gain stages will be necessary where the tradeoff will be added noise and extra components.

It will require extra components, but if the first stage has high enough gain (sufficient to raise the 1st stage output noise well above the input noise of the second stage) then the noise of the subsequent stages will add negligible noise to the total.

 

It will require extra components, but if the first stage has high enough gain (sufficient to raise the 1st stage output noise well above the input noise of the second stage) then the noise of the subsequent stages will add negligible noise to the total.

That is the high priced product for the commercial.

 

That is the high priced product for the commercial.

What?

 

You have to also remember that the GBW product you are referring to is for voltage feedback amplifiers - not transimpedance amplifiers. The BW of the transimpedance amplifier is calculated differently because the input is a current source, not a voltage. Key is to keep the capacitance around the feedback resistor and input node to a minimum, and to reverse bias the detector to minimize it's capacitance. Often, a PIN photodiode is used for high speed applications so that the capacitance can be very low when reverse biased.

 

What?

I meant that the function higher and the price will be higher too, because the op want to build a product that if you buy it from the market must be a high priced product.