I've never done much with opamps, but I'm trying to use a CA3130 as a unity gain buffer to an ADC input. As such I am powering it with a single sided 5V supply and want the input and output to swing from 0 to 5V (at least close to this).
The problem I am observing is that when the input signal gets up to approx 3.3V the output oscillates and continues to oscillate at a high frequency through to 5v input. From 0 to 3.3v it works as expected.

I tried a 47pf capacitor between pins 8 and 1 as suggested in the datasheet but it made no difference. I also noticed that noise on pin8 begins at the same time as the output goes awry.

The oscillation has an approx 1V swing and is centered at about the expected output voltage, ie the output still appears to follow the input to some degree.

 

Do you have decoupling capacitors at the power supply pins? Does it also happen when you put a load at the output?

 

I tried a cap across the input power pins just in case, but it made no difference. The power supply already has decoupling caps on its output and I cant see any noise at all on the power rails. The output is connected directly to the ADC input pin on the microcontroller (a picaxe 28x2), should I have a resistive load as well?.

 

should I have a resistive load as well?.

You could try it to see if it makes a difference, let's say 10k.

 

I tried a 10k resistor between output and ground and it seems to have eliminated the noise but I still cant get above 3.5v output. I then tried a 5M resistor but the noise came back.

 

The CA3130A requires external frequency compensation. See Pages 8and 9 of this data sheet, including Figure 8. http://www.intersil.com/data/fn/fn817.pdf

 

I tried a 10k resistor between output and ground and it seems to have eliminated the noise but I still cant get above 3.5v output. I then tried a 5M resistor but the noise came back.

In the datasheet it says at 15V it can swing to about 13V max with 2k load resistance. Meaning with 5V also it could not swing to 5V with a load resistance . Try 100k and maybe a small capacitor of 1n.

 

I tried a 220k and a 5M and both still had noise. I then tried a 1nf (I think its 1nf, its a disc style cap with 104 on it) and no output load resistor. There is no noise but still no higher output.
I had already looked at the datasheet but having a bit of trouble fully understanding it, particularly the offset and frequency compensation. My signal requirements are very slow moving DC so all I need is to cut off any high freq response.

As I understand the datasheet, the chip can provide rail output into a high resistance load. I figured that the microcomputer input should qualify as high resistance.

 

CA3130???? Isn't that like the first ever CMOS op amp ever made? Or am I thinking of something else???

 

I tried a 220k and a 5M and both still had noise. I then tried a 1nf (I think its 1nf, its a disc style cap with 104 on it) and no output load resistor. There is no noise but still no higher output.
I had already looked at the datasheet but having a bit of trouble fully understanding it, particularly the offset and frequency compensation. My signal requirements are very slow moving DC so all I need is to cut off any high freq response.

As I understand the datasheet, the chip can provide rail output into a high resistance load. I figured that the microcomputer input should qualify as high resistance.

Try this. You know your unity gain feedback loop? Use a 220 ohm resistor instead of a wire. Next, look up the input capacitance of the amplifier and put that value capacitor in parallel with the resistor.

Sometimes, the input capacitance of the amp. forms a pole and causes spurious issues. Also use both high value and small value caps for decoupling.

 

CA3130???? Isn't that like the first ever CMOS op amp ever made?

Yes, it was designed by RCA a long time ago when they invented Cmos logic.
It is very very noisy.

 

Yes, it was designed by RCA a long time ago when they invented Cmos logic.
It is very very noisy.

Ah!, I had wanted a CA3140, but the only electronics shop near here only had the 3130. I quickly looked at the specs while in the shop and thought it would probably do given that all I want is a buffer to isolate the micro from the outside world.

 

I tried the 220 ohm resistor and .1uf cap on the feedback loop, and then saw in the data sheet that they showed an example with .1uf and 2k resistor. Neither made any difference.
Is there a better opamp to use instead of the 3130?.
I've looked on all sorts of forums and most of the suggestions dont seem to be readily available.

 

I tried a 220k and a 5M and both still had noise. I then tried a 1nf (I think its 1nf, its a disc style cap with 104 on it)

You are describing a 100nf capacitor (0.1uf).
Per the spec sheet, 5V is the minimum recommended voltage and at this voltage the notes indicate that the temp must not go below 25°C.
Try increasing the voltage slightly and then the temp slightly to see what happens.
I have used this device successfully, but at a higher voltage.

 

Sorry, got distracted and forgot about this thread/question. In the end I tried the circuit shown in the datasheet for non-inverting amp with offset adjustment and frequency compensation and still didnt end up with a stable configuration, or rail output. i also tried upping the voltage to 12V and raising the temperature of the opamp, but nothing removed the awful output instability completely.
I gave up with this chip and switched to an LM324 (because I had one) and used it to give me 0-VCC/2 swing which is all I really needed in the end.
Thanks everyone for the help, I learnt a few things about opamps along the way though.

 

Even though you found a solution, here are some of the specs on the CA3130 running off +5V/0V. Note the input common mode range. The part is not guaranteed to work if the inputs are above 2.8V. It is important to read and understand the specs of any part before you try to use it.

 

Sorry, got distracted and forgot about this thread/question. In the end I tried the circuit shown in the datasheet for non-inverting amp with offset adjustment and frequency compensation and still didnt end up with a stable configuration, or rail output. i also tried upping the voltage to 12V and raising the temperature of the opamp, but nothing removed the awful output instability completely.
I gave up with this chip and switched to an LM324 (because I had one) and used it to give me 0-VCC/2 swing which is all I really needed in the end.
Thanks everyone for the help, I learnt a few things about opamps along the way though.

Did you notice in the datasheet that the output will go closer to the negative rail (ground in your case) if you add 10k from the output to ground?

 

thanks Ron,
As I said in an earlier post, I dont understand much about opamps and dont understand the common mode limitation. If I have it configured as a voltage follower, does that mean that the common mode voltage exceeds 2.8v when the input exceeds 2.8?.
I tried various resistors to ground and even to +5 to see what effect it had. The zero output side was not a problem, its the high side and the output noise that was the biggest problem.
As I said in the first post, I got sold a pup when I had intended to purchase the CA3140 but the shop had none, only CA3130's were available. I quickly looked at the specs in the shop and they seemed to be similar which is why I ended up with them.

 

When an op amp is connected with negative feedback (voltage follower is one example), the two inputs will be at the same voltage. This voltage is called the common mode voltage, because it is common to both inputs. The CA3130 with +5V on the positive supply rail is not guaranteed to work when the common mode voltage is above 2.8V.
CA3140 looks a little worse in this respect, with Vcm(max)=2.6V.