Dear all,

I have used LM318 high speed op amp for voltage follower , input -> 3Vp-p , 70kHz frequency . I've hard wired the output and input(-ve) pins - [6,2] .

I have no oscillations and works fine but only after bypass power supply with 0.1u ceramic capacitor .

But in datasheet they warned not to hard wired the output and input -ve pin , why? will this create problem in long run ?

If I keep the 10k and 5pf if feedback , the signal is distorted and gain reduces to 0.4 .
Is there a alternative ic for LM318 ?

Thanks in advance .



Attached the --> Distorted signal (when i make according to datasheet (not hard wired) ) ; and expected one

 

t in datasheet they warned not to hard wired the output and input -ve pin , why? will this create problem in long run ?

Likely to avoid overcurrent in the input protection diodes for fast risetime signals.

Distorted signal (when i make according to datasheet (not hard wired) )

What distortion are you referring to?

 

Power supply bypass is always a good idea. Stray capacitance on pin 2 can reduce the gain. Is this built on a plastic bread board? If so, these have large stray capacitance between columns.

 

Dear all,

I have used LM318 high speed op amp for voltage follower , input -> 3Vp-p , 70kHz frequency . I've hard wired the output and input(-ve) pins - [6,2] .

I have no oscillations and works fine but only after bypass power supply with 0.1u ceramic capacitor .

But in datasheet they warned not to hard wired the output and input -ve pin , why? will this create problem in long run ?

If I keep the 10k and 5pf if feedback , the signal is distorted and gain reduces to 0.4 .
Is there a alternative ic for LM318 ?

Thanks in advance .

View attachment 141169

Attached the --> Distorted signal (when i make according to datasheet (not hard wired) ) ; and expected one

Hi,

The short answer is that the device is not compensated for unity gain and that improves the response time. The designers probably assumed it would be used for gains greater than 1.

If you dont need the full bandwidth you could try increasing the gain to maybe 2 and then using a divider on the output. You may even get away with a smaller gain like 1.2 or something but you'd have to test.

Dont know why you are seeing such reduced gain with the 10k and 5pf, we'd have to look into that with some careful tests.
The internal circuit contains some 30 transistors so you'd have to draw them all into a simulator to go farther with the idea of trying to figure out if it is something inside, which it must be due to the lack of unity gain compensation.

There are a lot of op amps these days that will do 10MHz. You should look on some of the manufacturers sites and use their search tools.
Are you really working with only 68kHz? You may be able to get away with a slower op amp.

 

Thanks for all your replies,

Does Lt1362 works as the alternative to lm318 ? For unity stable gain .

 

Thanks for all your replies,

Does Lt1362 works as the alternative to lm318 ? For unity stable gain .

Hi,

They do state that it is unity gain stable, but there are some catches (read the data sheet).

It is faster than your previous choice so careful layout is probably a good idea.
Depending on load, a good quality low ESR bypass cap is a good idea too.

 

he short answer is that the device is not compensated for unity gain and that improves the response time. The designers probably assumed it would be used for gains greater than 1.

Why do you say that?
The TI data sheet states:

 

If you have a circuit, any circuit, and it oscillates which of the following is true:
1. No compensation is necessary.
2. Must add compensation network.

Could be either.
There's more than two possibilities here.
It could be a bad layout, inadequate decoupling, excessive output capacitance, or something else.

Aaqash, make sure that each op amp power pin is directly decoupled to ground using 0.1μF ceramic caps with as short leads as possible.
And for best results the circuit should have a ground plane.

 

Could be either.
There's more than two possibilities here.
It could be a bad layout, inadequate decoupling, excessive output capacitance, or something else.

Hello,

You are ignoring the context.

This is a multiple choice question, where you only get two choices. In a multiple choice question you cant invent your own questions and there is a good reason for that. It is to narrow down the possibilities for the scope of the question and answer. We have a clear context here, it's either compensated or it's not.

So i will repeat the question:

If you have a circuit, any circuit, and it oscillates which of the following is true:
1. No compensation is necessary.
2. Must add compensation network.

You only get two choices and that prevents you from avoiding the obvious. The obvious is that if you have oscillations and you only can either compensate or not, then surely you can not choose #1 because that does nothing.

 

You are ignoring the context.

And what context is that?

This thread is about oscillations in an op amp circuit, which could be from any number of reasons.

So i will repeat the question:

No need to be condescending.
I understood it perfectly the first time.

You only get two choices and that prevents you from avoiding the obvious.

And what is the obvious?

The obvious thing is that you are giving two arbitrary choices as answers, neither of which is necessarily the complete answer to the TS's problem

And those two choices have little to do with my response to you, which you quoted in post #8, where I was pointing out the error in your response to the TS in post #4, about the op amp compensation.

 

If you have a circuit, any circuit, and it oscillates which of the following is true:
1. No compensation is necessary.
2. Must add compensation network.

That list is surprisingly tunnel-vision-esque.

3. Add/improve the power decoupling and routing. LM318, LM380, LM386
4. Address other power/ground quality issues TDA2002
5. Address other power/ground quality issues CD4049UB acting as a microphone preamp ...

Meanwhile, back to the TS ...

The LM318 was my first "video" opamp. It was ahead of its time in terms of gain and bandwidth, and it has "personality". But I never had unexpected gain issues. Consider this:

ALL electronic circuits - tubes, transistors, hybrids, IC's, analog, digital, DC, microwave - whatever - are based on the fundamental assumption (read: requirement) that the circuit is powered by theoretically perfect power sources. The vast majority of the time it is a voltage source with a ZERO-POINT-ZERO ohm output impedance at all frequencies of interest, including high-order harmonics. Because wires and pc board traces have inductance that appears as a non-zero impedance in series with the power supply, the standard approach is to add one or more capacitors as close as possible to the device power pins with the shortest possible trace or routing length to the device other power pins, GND pins, and system ground. Not doing this affects stability, gain, distortion, and stability.

ak

 

This squabbling is not helping the TS at all and is entirely off topic. If you want to discuss personalities please take it out of the public forum.

Some posts that are entirely off topic have been removed.

 

In reference to the question:
If you have a circuit, any circuit, and it oscillates which of the following is true:
1. No compensation is necessary.
2. Must add compensation network.

That list is surprisingly tunnel-vision-esque.

It would be in a thread where we were talking about the general solution to oscillation, but this question came in response to another question that was in the context of if there was already compensation or not.

It's very simple. If a circuit needs compensation it needs compensation.
What was being asked of me was to repeat everything that had already been said and i did not feel that was my responsibility.

I wont reply anymore to posts that dont address the specific issues being talked about, but if you'd like to discuss it more in PM's that's fine with me.

 

This squabbling is not helping the TS at all and is entirely off topic. If you want to discuss personalities please take it out of the public forum.

Some posts that are entirely off topic have been removed.

Hi,

Many thanks as this stuff is getting out of hand lately.
This back and forth stuff is over for me for one.