# OP Amp buffer stability

Discussion in 'Homework Help' started by eladta, Jun 30, 2013.

Apr 20, 2013
41
0
Hi,

I want to use the AD8638 OP Amp as a simple buffer but my circuit is unstable.
I know that according to theory we need to check the phase of the open loop gain where the magnitude of it crosses 0 dB. According the datasheet page 8 it comes to 1 MHz, and the phase is aprox. 65 degrees. Meaning stability.
I don't understand the basics, say i wanna use another freq. what than ? how do i check the stability ?

Attached are the datasheet and the circuit.

File size:
622.9 KB
Views:
16
• ###### buffer.jpg
File size:
10.4 KB
Views:
36
Last edited: Jun 30, 2013
2. ### LvW Active Member

Jun 13, 2013
674
100
Latch-up or oscillation?
Power supply?
Termination of non-inv. input?

3. ### WBahn Moderator

Mar 31, 2012
18,089
4,917
Are you bypassing the power supply close to the amp?

Are you being sure that the input voltage range, which only goes up to about 2V less than the supply voltage, is not being exceeded?

Apr 20, 2013
41
0
Sorry it took me so long to response..
Oscillation instability.
power supply of 5v with a ferrite bead

Anyway my goal is to understand the principal of stabiliy issue and not just to solve this praticular problem. If you can help me with that it would be much appreciated.

Apr 20, 2013
41
0
Yes i am bypassing the power supply close to the amp.
About your second question, if i'm understand you right there is a limitation of the input voltage which should be 2v less than the power supply ?
If so i did not knew that..

Anyway my goal is to understand the principal of stabiliy issue and not just to solve this praticular problem. If you can help me with that it would be much appreciated.

6. ### LvW Active Member

Jun 13, 2013
674
100
Yes - but you should know that there are two different forms of instability:
DC instability (no operating point) and ac instability (oscillation).
Therefore, my question: Which effect did you observe?

Apr 20, 2013
41
0
ac instability (oscillation).

8. ### LvW Active Member

Jun 13, 2013
674
100
The frequency of oscillation would be interesting - however, I suppose it will be in the range of the transit frequency (unity open-loop frequency, correct?).
It is surprising since the opamp type is unity gain stable (according to the data sheet.).
Please, show the complete circuit including power supply, possible load and input connections.

EDIT: Final question - oscillation hardware or in simulation?

Last edited: Jul 2, 2013

Apr 20, 2013
41
0
Here is the complete circuit.
What do you mean by "I suppose it will be in the range of the transit frequency" ?

File size:
27.1 KB
Views:
52

Apr 20, 2013
41
0
Both of them

11. ### LvW Active Member

Jun 13, 2013
674
100
The pos. input has no bias voltage resp. current.

12. ### WBahn Moderator

Mar 31, 2012
18,089
4,917
That's the complete circuit? Really? You have the positive input tied only to one side of two capacitors and the other sides are floating? You have a resistor that is just lying in the circuit not connected to anything? And where did you get a stock of 4.02kΩ resistors?

Apr 20, 2013
41
0
Let's leave aside the circuit and focus on the principal, I want to use the AD8638 OP Amp as a simple buffer.
I know that according to theory we need to check the phase of the open loop gain where the magnitude of it crosses 0 dB. According the datasheet page 8 it comes to 1 MHz, and the phase is aprox. 65 degrees (meaning stable). what does it mean exactly ?
say i wanna use another freq. what than ? how do i check the stability ?

14. ### LvW Active Member

Jun 13, 2013
674
100
This means that your minimal phase margin (for 100% feedback) will be 65 deg. That´s a good value (as long as one can trust such data sheet information) .
For larger gain (less feedback) the margin is even larger.
This margin is independent on the applied signal frequency.
When an amplifier with feedback tends to be unstable - it is unstable, independent on the applied frequency.

Apr 20, 2013
41
0
How do i control the feedback amount ?
Secondarily, if the input freq. is independent of the stability why do i care at what freq. that the OP AMP is stable ?

16. ### LvW Active Member

Jun 13, 2013
674
100
* Do you know the meaning of the feedback factor? That is the factor which determines the part of the output signal that is fed back to the input.
Logical, is it not?
* The question "at what freq. is the op amp stable" is not appropriate.
Its stability properties (stable/unstable) are completely independent on the signal frequency. If it is unstable in gets into self-excitement (oscillation or latch-up).

Apr 20, 2013
41
0
I understand what you are saying, but i'm still missing something.
e.g the above op amp. we see that the magnitude of the loop gain crosses 0 dB at 1 MHz and we see that the phase margin is 65 degrees. But what can we make from the fact that the freq. is 1 MHz ?
What if the loop gain would cross 0 dB at 100 MHz and the phase margin would be same. what is the difference ?

18. ### LvW Active Member

Jun 13, 2013
674
100
You would have an amplifier that can work as an amplifying device within a frequency range that is 100 times larger than before.
Example: Unity gain amplifier (buffer operation) for frequencies up to 1 MHz or 100 MHz, respectively. (These values are the -3 dB points, approximately).

Comment: The above considerations are theoretical only. In reality, you have to check the slew rate also.

Apr 20, 2013
41
0
Got it. Thanks.
In a more practical point of view, say i want to test the phase margin of the above OP AMP and see if it's corresponding with the datasheet. do you know how can you do it ?

20. ### LvW Active Member

Jun 13, 2013
674
100
Do a simulation of the units open-loop gain - and you can verify everything you want (gain magnitude and phase).
However, be aware that you must find before the correct bias point in the linear range of the amplifier (determine offset voltage before).
As another possibility you can provide negative feedback for dc only using a suitable R-C combination in the feedback path.
Do you know what I mean?

Last edited: Jul 4, 2013