Core Electronics Forum

Some Problems When Using LM318

Recently, our laboratory needs to build an instrument. This instrument has four voltage outputs vi1~vi4, and the four output voltages are in the range of 0~5V. I need to do the following calculations and then put the result data acquisition card: vo=(vi1 +vi2)-(vi3+vi4) The attached picture shows the circuit I designed. The actual op-amp I chose is LM318. It is said on the Internet that the speed of this op-amp is relatively fast and tested it with an oscilloscope by referring to the datasheet. I found that the output noise was very large, and the output voltage was not the same as expected.
What should I do? I know less about this, so hope for some good ideas.
Good listening!

2 Likes

No picture.
Cheers Bob

1 Like

Hi Guo,

While my experience of picking parts (and designing circuits) is limited (working on it…), the Op-Amp you chose doesn’t seem to focus on low noise at all, using TI’s parametric search seems to come across units with much lower noise ratings.

Keen to see how you go once you compare some datasheets.
-James

3 Likes

Hi James
I think Guo may have lost interest. I did reply stating there was no picture or circuit attached but there was no other answer forthcoming.

I don’t know why he picked an OP amp with 3 compensation points. Only complicates the issue. And what he measured remains a mystery.

To do what he wants is fairly simple but I think requires 4 OP amps and a dual (+/- 12V) supply.
I would suggest a low power Quad rail to rail unit such as TL074. and a handful of 10k resistors.

If Guo shows any ongoing interest I can sketch up a basic circuit for him to experiment with and massage at will. But I am not going to waste any time if he is not interested enough to even reply.
Cheers Bob

3 Likes

Hello!
Sorry that I can not reply in time. Because I am busy recently, which have other things to do. So…
And this is the simplified picture below.
lm318

2 Likes

Hi Guo
That will probably just go to 5V and stay there. You are providing positive feedback so will rapidly switch to max.
You say the inputs are 0 to 5V,
That means when 2 inputs are added the possible max is 10V yet you only have a 5V supply.
You have no negative supply so the output can never go negative if you say have 2V minus 3V the result would be -1V. The O/P would not even get to zero. You cannot do what you describe with 1 OP amp.
I will post you a circuit to play with and a description. Will try to get it on to-morrow. Just give me a bit of time.
Cheers Bob

3 Likes

Thank u for ur serious reply.
Yep.
It will be so helpful, and I am looking forward to the circuit you post.
Also, I am sorry to take ur time.

2 Likes

Hi Guo
The following circuit will solve the equation as you gave it. V0 = (Vi1 + Vi2) – (Vi3 + Vi4)
Jpeg.


Alo in PDF which may be better quality.
Add_Subtract.pdf (285.5 KB)

As I see it 4 OP amps are required. You cannot add and subtract with 1 Op amp at the same time as you tried to do. I think you got the + and – inputs reversed anyway. Any feed back should go to the – input.

A summing amp will invert the signal ; say Vi1 =3V and Vi2 = 2V, OP 1 output will be -5V. A unity gain inverter is used to change to a +V output.

What I have done here is add the 2 pairs of inputs, resulting in -ve voltages at OP amp outputs. Invert one of them and apply to another summing amplifier (which will invert again) which is the same as subtracting one from the other. The result will satisfy your formula.

Example: Vi 1 = 3V, Vi 2 = 2V, Vi 3 = 2V, Vi 4 = 1V

Put these numbers into your formula VO = (3+2)-(2+1)= 2

Now apply these numbers to the circuit.

Vi 1 + Vi 2 = 5V, inverted this becomes -5V at OP 1 output.

Vi 3 + Vi 4 = 3V, inverted this becomes -3V at OP 2 output.

Invert this -3V with OP amp OP 3 which now has output of +3V

Apply this -5V and +3V (which summed results in -2V) to summing amp OP 4 which will invert to result in +2V.

Formula satisfied.

Power supply. Because results are +ve and -ve and potentially reach 10V a split pos and neg 12V supply is required. No matter what you do there is no way around this. Do not attempt to artificially split a single supply or you will find yourself in trouble with Ground connections later. Also be careful with a couple of batteries and boost converters. Most of these converters have a common negative or ground connection and cannot be used here. If you can find any of these things with a completely isolated output this should be OK.

Accuracy: Will be largely dependant on tolerance and quality of the resistors. The OP amps themselves will be very accurate. Use quality resistors of at least 1% or even 0.1%.
There is one unknown. The source impedance of the inputs. This is effectively in series with the input resistors and can affect the gain/loss of the first OP amps. The input resistance if each input is 10kΩ. If this is too low for your application you can increase up to 100kΩ (all resistors with 2 input OP amps, R1 to R8) but I would not go above that. If still a problem you may have to consider another Quad OP amp as an interface configured as unity gain voltage followers.

I don’t know how conversant with OP amp behaviour you are but there is one trap worth mentioning. DO NOT expect to measure anything at the OP amp input pins. It may sound strange but this point is at ground potential and you will measure 0 or zero. I don’t have space or time to go into this here but if you were not aware of this I would suggest doing some reading on OP amp fundamentals. This is a trap many (including myself) have fallen for thinking the amp is shorted or faulty when it is quite normal.
Cheers and good experimenting Bob

3 Likes

Bob, thank you so much!
It is so~o~ o detailed for me to get it.
And I have studied a lot about this.
Guo.

Hi Guo
No Problem.
Study OP amp behaviour and you will see how this circuit works.
Cheers Bob

1 Like