The output of my sensor is a sinosoidal (40KHz) voltage of 20mV p-p and I need to amplify this signal WITHOUT distortion and/or clipping to 5V p-p.
I have opamps available with me that have GBP ranging from 500Khz to 1Mhz.

Assuming I keep a 20% margin on the GBP specs, the 1MHz opamp reduces to 800KHz GBP.

Keeping this at mind, would it be possible to achieve the reauied amplification by:

1. Using single opamp
2. Using cascaded opamps from the SAME IC?

 

Use two cascaded in a single package (a dual op amp).

 

What's the slew rate?

If the output can't change quickly enough then it won't be able to do it regardless of the GBWP.

__________________

I do not answer private messages asking for help because no one else can: benefit from advice I may give or correct me if I'm wrong.

Please ask on the open forum if you have a question and I'll be happy to help, if I know the answer.

 

Very old opamps with poor high frequency response have a GBP of only 1MHz.
A lousy old LM358 and LM324 has a typical open-loop gain of only 12 at 40khz and its output makes a sine-wave with a max of only 3Vp-p.

A TL07x opamp has a GBP of 3MHz. It has a typical open-loop gain of 80 at 40kHz and its output makes a sine-wave with a max of 33Vp-p.

__________________
Uncle $crooge

 

What about the LF351 or 358?

I would also like to read more about cascaded amps - specialy their issues with noise vs signal... could you point to some nice resources?

 

Look at the datasheets.

Work out the gain and slew rate you require.

SR = 2πF×Vpk

If the slew rate you need is higher than that specified on the datasheet then the op-amp is no good.

Look at the graph for the open loop gain at the desired frequency, if it's lower than the gain you require, then you'll need to use two stages. The overall gain in a two stage system is the gain is equal to Av1×Av2 therefore to work out the gain of each stage (assuming you want to make them both the same) simply take the square root of the overall gain.

__________________

I do not answer private messages asking for help because no one else can: benefit from advice I may give or correct me if I'm wrong.

Please ask on the open forum if you have a question and I'll be happy to help, if I know the answer.

 

The LF351 was invented by National Semiconductors who stopped making it 11 years ago. ST Micro might still make it.
It is almost the same as a TL081 except it is noisy and distorted. They both have the Opamp Phase Inversion Problem where the output suddenly inverts if an input voltage is lower than the input common-mode voltage range (a few volts above the negative supply voltage).

__________________
Uncle $crooge

 

Correct me if I am wrong, but I am arranging the opamps in order of their intrinsic noise:

LF351 > TL084 > TL074 > TL064

(Wow - it seems more "recent" opamps are more noisy?)

I was also wondering on the use of cascading to "get around" GBP restrictions, but really "something's gotta give" - there MUST be restrictions on cascading opamps - where do I find about them?

 

Yes, you've got them in the right order, with the noisiest first.

No, you've got that backwards, the newer op-amps are less noisy.

There are no restrictions providing you don't put them all in the same feedback loop.

One thing you'll notice is that errors tend to add when you cascade amplifiers.

Also cascading doesn't actually increase the GBWP, it just increases the overall gain.

__________________

I do not answer private messages asking for help because no one else can: benefit from advice I may give or correct me if I'm wrong.

Please ask on the open forum if you have a question and I'll be happy to help, if I know the answer.

 

A TL074 is a TL084 that is selected for low noise.
A TL064 is low power so its noise is higher (and the max gain is lower).

There are many newer opamps with extremely low noise.

An opamp has a bandwidth of only to 5hz when its gain is set very high.
Its bandwidth is much higher when its gain is set low.
Two cascaded opamps have a wide bandwidth because each has fairly low gain. Their gains are multiplied.

Opamps also have a frequency limit called power bandwidth. The slew rate of the opamp affects the max power bandwidth frequency.
An LM324 or LM358 has a power bandwidth to only 2kHz.
A lousy old 741 opamp has a power bandwidth to only 9khz.
A TL07x has a power bandwidth to 100kHz.

__________________
Uncle $crooge

 

Op amps are all about trade-offs. What audioguru calls a lousy opamp might, in my field of work, be a fantastic one I just selected an op amp for a system that can only slew to its rails at frequencies lower than 3KHz (and that's at 3V rails). It ain't cheap either. But it has extremely good DC specs and only consumes 20uA quiescent current (compared to 5ma for a more typical op amp). My system should have a shelf life of over a year, instead of hours (FWIW, it's an OPA336 SOT-23 single op amp from TI. Two dollars and change at Digi-key for a single unit).

It's all about tradeoffs.

 

Which spec in the datasheet will tell me that?

Which spec in the datasheet will allow me to compare "noise" ?

Can you tell me more about that .. or somewhere I can read and find out?

Can you tell me more about that .. or somewhere I can read and find out?

 

The low power TL064 has a typical gain of only 6,000.
The normal TL074 and TL084 has a typical gain of 200,000.

The low power TL064 has a typical input noise voltage of 42.
The low noise TL074 has a typical input noise voltage of only 18.

Many manufacturers copy Texas Instruments who invented these ICs. But the datasheets from Texas Instruments have much more detail.

__________________
Uncle $crooge

 

Why do I need to keep the open loop gain at mind?

Can you help me locate these figues in the datasheet please?

 

a high open loop gain reduces distortion and widens the bandwidth when negative feedback is applied.

The datasheets for the TL064 and TL074 state the equivalent input noise voltage and show it on graphs. The first page of the TL074 datasheet also says its value.

The inputs use jFETS with an extremely low bias current so the noise current is almost zero.

__________________
Uncle $crooge