Op Amp Vs and Vo

Thread Starter

SamR

Joined Mar 19, 2019
5,031
I've been working with an old book by Yannis Tsividis "A First Lab in Circuits and Electronics" published by John Wiley & Sons some time ago. No date given. While working on one of the Op Amp exercises I set up the following circuit using an LM741. I added a 560Ω load resistor to the output and the needed rail capacitors. I used my dual individual power supplies connected to provide +/- 5V to the rails but the -5V drifted to -5.03V even when intially set to 5.04V it supplied -5.0V at the rail.
1604369754588.png

IMG_0755[1].JPG
I set the Vref to 2.50V on pin 2 and ran the voltage to pin 3 up and down. And was surprised that the closeness to the positive and negative supply rails was different. So I pulled up the LM741 PDF from Fairchild up and looked for what the Voltage offset from the rail should be.
1604370706392.png
That is for +/-15V but 14V/15V*100% tells me it should only be a ~7% offset. But the numbers I was getting didn't support that. I knew I had 2 different lots of 741s and I was seeing 2 different sets of outputs so I tried the 2 TI uA741CPs from DK and got a 3rd unique set of values. The TI chips gave me a 12% offset on the + rail and ~38% offset on the - rail.

IMG_0756[4].JPG
So my question is, how much offset from rail should there be? And, Is there a parameter on the PDF that I am missing?
 

Attachments

OBW0549

Joined Mar 2, 2015
3,566
I've been working with an old book by Yannis Tsividis "A First Lab in Circuits and Electronics" published by John Wiley & Sons some time ago. No date given. While working on one of the Op Amp exercises I set up the following circuit using an LM741.
741s are good for solder practice and not much else. Why, oh WHY, do people still insist on using those damned ancient things? It was designed 50+ years ago and its only distinction is that it was the first op amp that wasn't an absolute, titanic pain in the ass to use-- like the uA709 or, God forbid, the uA702.

I added a 560Ω load resistor to the output....
Not that it matters much, but why?

I used my dual individual power supplies connected to provide +/- 5V to the rails but the -5V drifted to -5.03V even when intially set to 5.04V it supplied -5.0V at the rail.
Note that op amps don't give a flying fat rat's ass about the precise value of their supply voltages. If they did, people wouldn't use them. As long as the supply voltage is adequate for the innards of the op amp to function (for a 741, +/- 5V is not) AND it doesn't exceed the op amp's maximum ratings AND it's enough to give you the output voltage swing you need in your circuit, that's all that matters. The exact value of the supply voltage is of no importance so long as it meets those requirements.

I set the Vref to 2.50V on pin 2 and ran the voltage to pin 3 up and down. And was surprised that the closeness to the positive and negative supply rails was different. So I pulled up the LM741 PDF from Fairchild up and looked for what the Voltage offset from the rail should be.
That spec is not for a "voltage offset"; you're using the wrong term. That spec is for exactly what it says: its the maximum output voltage swing you can expect under the stated conditions (supply voltage, load, whatever). For example, with a +/- 15V supply and 10 kohm load, the 741 will typically drive up to about +/-14V, but might only give you +/-12V.

That is for +/-15V but 14V/15V*100% tells me it should only be a ~7% offset. But the numbers I was getting didn't support that. I knew I had 2 different lots of 741s and I was seeing 2 different sets of outputs so I tried the 2 TI uA741CPs from DK and got a 3rd unique set of values. The TI chips gave me a 12% offset on the + rail and ~38% offset on the - rail.
Again: this is NOT an "offset" and it makes absolutely ZERO sense to compute it as a percentage of anything. It is also not a precise quantity so don't expect it to be. All it's telling you is how close to the rails the op amp's output can be expected to swing so you don't ask more of it than it can deliver.

So my question is, how much offset from rail should there be? And, Is there a parameter on the PDF that I am missing?
You're not missing a parameter; you're misinterpreting the parameter that's there.

And stop calling it an "offset."
 

Thread Starter

SamR

Joined Mar 19, 2019
5,031
That would explain a lot and I'm not ready to retest all of them @ +/-15V.

I knew I had 2 different lots of 741s and I was seeing 2 different sets of outputs
And I was going to separate them into 2 lots so I got my microscope out to check the markings on the 2 lots to see how they differed and got quite a surprise. I lost count on the number of various makes but there were also 2 different models, uA741CN and LM741CN all mixed together. And the 2 lots did not separate out by model number! Even though the results showed 2 very distinct lots there were multiple makes and some of each model in each lot. Although the lot that was going to -5.03V was almost entirely uA741CN having only one LM741CN in it.

I obviously need to be much more particular when bagging and marking parts for storage.
 

Thread Starter

SamR

Joined Mar 19, 2019
5,031
Note that op amps don't give a flying fat rat's ass about the precise value of their supply voltages.
What surprised me was that I was getting -5.03V at all since I had the negative supply set @ -5V. Which it wasn't as it had drifted to -5.03V but the question remains as to how it could swing all the way to the rail voltage?

Why, oh WHY, do people still insist on using those damned ancient things?
I was probably much younger when I bought these things (in the 70s) and using an ancient book so I actually had what part it called for... I did buy 10 from AliX so that may have been mixed in with what I already had.
 

OBW0549

Joined Mar 2, 2015
3,566
What surprised me was that I was getting -5.03V at all since I had the negative supply set @ -5V. Which it wasn't as it had drifted to -5.03V but the question remains as to how it could swing all the way to the rail voltage?
Must be a measurement error of some kind. Was your 560 ohm load resistor connected between the output and common (i.e., the junction of the two supplies), or was it connected to the -5V supply? If the latter, it probably was pulling the output down all by itself.

If you want some better, more modern, op amps then this discussion might be useful.
 

Thread Starter

SamR

Joined Mar 19, 2019
5,031
The rail 2 supplies are floating and their common pt is connected to the bench PSUs used for Vin and Vref supplies negative which is grounded. So the 560Ω Rload is grounded.

At this point, I'm doing baby steps trying to learn the basics of op amps. I'll check the link but I have been skimming threads and buying some of the recommended op amps so I have about a dozen different models on hand. That particular thread predates my coming here so I'll check it against my inventory and see if I need to add some to it.
 

crutschow

Joined Mar 14, 2008
34,280
Use a minimum of ±10V for the 741 supplies before you try to make sense of anything you see in its operation.

If the output is equal to one of the rail voltages then something is seriously wrong.
 

Thread Starter

SamR

Joined Mar 19, 2019
5,031
I thought 5V was low but that was what the exercise called for? Basically, it was to understand what a Vref was for a comparator and the sharp Voltage change @Vref and it did demonstrate that.

If the output is equal to one of the rail voltages then something is seriously wrong.
I suspect all of those op amps are bad even if they did work for the exercise as they should have. This will be a good basic test circuit for op amps in the future with at least +/-10V.

I did run the rails up to +/-15V and used one of the chips that were going to -5.03V and it went to -15V and 13.something. Thinking I should put a mV sine wave into it and see what it does on the scope... Hmmm... No time like now.
 

Thread Starter

SamR

Joined Mar 19, 2019
5,031
+/-10V on the rails as a voltage follower with 10Hz @20.0Vpp (SG max V) going to pin3. Starting to clip on +rail ~18Vpp but not quite reaching the -rail. So the scope tells a different tale than the comparator exercise although it is going close to the -rail without clipping.

SDS00041.png

Now one that only went to -3.74V and I get the same thing?
SDS00042.png
Interesting...
 

dl324

Joined Mar 30, 2015
16,839
That is for +/-15V but 14V/15V*100% tells me it should only be a ~7% offset. But the numbers I was getting didn't support that.
You're interpreting the datasheet incorrectly (and, as others have mentioned, this isn't an offset voltage).

Excerpts from Fairchild LM741 datasheet:
clipimage.jpg
clipimage.jpg
clipimage.jpg

The offset voltage is the deviation of the opamp from ideal input characteristics. What you're talking about is output voltage swing.

The output voltage swing isn't a percentage of the power supply voltage. With a 560 ohm load, the opamp is "heavily" loaded. The datasheet only gives parameters for as low as 2k. So the worst case voltage swing with +/-5V supplies is 0V; typical would be +/-3V. It would be even worse in your case.
+/-10V on the rails as a voltage follower with 10Hz @20.0Vpp (SG max V) going to pin3. Starting to clip on +rail ~18Vpp but not quite reaching the -rail. So the scope tells a different tale than the comparator exercise although it is going close to the -rail without clipping.
You're violating the common mode input voltage range. You can't count on the opamp to work under those conditions.

From the datasheet, the maximum guaranteed input range with +/-10V rails is +/-7V (3V below the rails). Typical would be +/-8V.
clipimage.jpg
clipimage.jpg
 

Audioguru again

Joined Oct 21, 2019
6,671
Still using LM741? Be prepare for a tongue lashing from the audio guru.
I must admit that in about 1972 (48 years ago) I tested an MC1458 which was a dual 741 opamp and its performance was horrible.
Then the lousy LM324 and LM358 opamps were introduced and I tested one but never used it.
When the TL072 was introduced I used many of them for years.
I have never seen a lousy old 741 opamp.
 
Last edited:

Thread Starter

SamR

Joined Mar 19, 2019
5,031
With a 560 ohm load, the opamp is "heavily" loaded.
K, I did notice the >2kΩ but didn't understand that as a minimum. The 560Ω load was from an exercise using an LED indicator that I had left on as a load. Should I be trying to measure the output with a load or not?

You're violating the common mode input voltage range.
This I don't understand... Yes, I was purposely overdriving it to see if it would clip on the bottom. Is that what you meant by violating the input range?
 

Thread Starter

SamR

Joined Mar 19, 2019
5,031
Inverting or non-inverting? You can increase gain to make the output clip.
I did it just as a voltage follower. I should try it with gain as an amplifier instead. Didn't consider that but that would certainly be better. Let me get some coffee in me and wake up a bit and I'll give it a go.
 

dl324

Joined Mar 30, 2015
16,839
K, I did notice the >2kΩ but didn't understand that as a minimum. The 560Ω load was from an exercise using an LED indicator that I had left on as a load. Should I be trying to measure the output with a load or not?
You can, but you don't need 10mA. With 15V supplies, 2k would give a load half of that.
This I don't understand... Yes, I was purposely overdriving it to see if it would clip on the bottom. Is that what you meant by violating the input range?
When you exceed the allowed input range, the opamp won't be operating in its linear region.
 
Top