Voltage controlled switch - op-amp question

[armandine2]

I thought this was going to be fairly simple - the initial preamble to the Question ...

 (No.8, section 2, Paper II, Syllabus A - Joint Matriculation Board Summer 1991 - Physics Examination Paper - on the face of it an op-amp voltage controlled switch and a dc transient using a  split rail +/- 4.5V [not visible in that picture] power supply)

....states the "capacitor is uncharged , switch S is closed and the LED is off (unlit)".

Hmmm -

my LED though is lit.

My next move(s) -

remove the two serial psu earths and float them
try another op-amp (than Motorola MC1741    SCP1    K8734) as I have an input at a rail voltage?

Am I missing something more obvious, I suspect so?

[xavier60]

Tell us what are all of the opamp's pin voltages with respect to 0V ground.
The idea is to figure out the problem's cause by using an understanding of how opamps operate.

[armandine2]

My voltages are:

1) = -4.48
2) = -.04
3) = -4.50
4) = -4.50
5) = -4.47
6) = 3.24
7) = 4.50
= 0.00

floating the supplies didn't affect the circuit. (unsurprisingly)

[xavier60]

Except for the output pin 6, all voltages are as expected. With the Non-inverting input being more negative than the Inverting input, the output should be some large negative voltage.
The opamp is either faulty or maybe doesn't like having an input pin at negative rail voltage.
Is there a blink from the LED when the switch is turned off?

[fcb]

The 741 opamp is not a rail-rail input, and you are treating the non-inverting input like it is rail-rail. This is why it's not working.

If you look at the equivalent circuit schematic for the 741, you'll see some current-mirrors and other gubbins below the differential input pair (this may help you understand why the 741 doesn't work at input voltages close to Vss).

LM358 would work.

[armandine2]

no blinking from the LED - The switch is a jumper wire - at start up it is in circuit and as soon as the power supplies are switched on it is lit.

The meat of the exam question is around what happens when the switch is opened:

                    a) Explain why the LED comes on 6.9s after the switch is opened and then remains lit.

In my breadboarded   circuit   when the capacitor short is removed the led goes out, after a second or so,  and comes back on 5 seconds ish later.

[xavier60]

no blinking from the LED - The switch is a jumper wire - at start up it is in circuit and as soon as the power supplies are switched on it is lit.

The meat of the exam question is around what happens when the switch is opened:

                    a) Explain why the LED comes on 6.9s after the switch is opened and then remains lit.

In my breadboarded   circuit   when the capacitor short is removed the led goes out, after a second or so,  and comes back on 5 seconds ish later.

That's what I meant by " blink"
That proves that the opamp doesn't work properly while the Non-inverting input is at full negative rail voltage. It is outside of its input common mode range. Data sheets often dont directly point out this type of problem.
Put a resistor across the switch, like 33K and test again.

[xavier60]

I should have said to put the 33K in series with the switch or jumper.

[xavier60]

With the LM358, the input common mode range includes the negative rail but does not include the positive rail but so long as one input is within the specified common mode range, the output state will be true to the input logic.

[xavier60]

The 741 opamp is not a rail-rail input, and you are treating the non-inverting input like it is rail-rail. This is why it's not working.

If you look at the equivalent circuit schematic for the 741, you'll see some current-mirrors and other gubbins below the differential input pair (this may help you understand why the 741 doesn't work at input voltages close to Vss).

LM358 would work.

Looking at the schematic for the LM741, I can't see a reason for the behavior. https://neurophysics.ucsd.edu/courses/physics_120/lm741.pdf 
"No Latch-Up When the Common Mode Range is Exceeded " it says.
Maybe the part is not genuine.

[xavier60]

I bread-boarded the same experiment using a TI branded UA741. It worked properly, as expected.
I found that so long as one input remained within common mode range, no closer than 2v from the negative rail, the output state would always correspond correctly to the input difference.

[perieanuo]

(Attachment Link)

I thought this was going to be fairly simple - the initial preamble to the Question ...

 (No.8, section 2, Paper II, Syllabus A - Joint Matriculation Board Summer 1991 - Physics Examination Paper - on the face of it an op-amp voltage controlled switch and a dc transient using a  split rail +/- 4.5V [not visible in that picture] power supply)

....states the "capacitor is uncharged , switch S is closed and the LED is off (unlit)".

Hmmm -

my LED though is lit.

My next move(s) -

remove the two serial psu earths and float them
try another op-amp (than Motorola MC1741    SCP1    K8734) as I have an input at a rail voltage?

Am I missing something more obvious, I suspect so?

s closed, IN- "bigger" than IN+, opamp output is low, led is off

[armandine2]

See attached picture for the op-amp I used.

[armandine2]

For clarification I have not modified the circuit to work outside the shown exam figure. Hence finding a op-amp that works in the circuit shown I would consider to be practically solved.

[xavier60]

See attached picture for the op-amp I used.

The photo needs to be much more detailed. Does the printing look tilted? Also look closely at the surface texture. There should be no sign of it being abraded, fine directional scratches.

[xavier60]

Where did the part come from? I should have also asked.

[armandine2]

I can see a tilt on the thumbnail picture! otherwise looks ok

the chip would've been acquired in the early 1990s (latest) - I haven't been getting ICs since then (when I was a student).

[xavier60]

 The part number on old ICs were mainly screen printed and easy to alter compared to  laser inscribed markings.

[fcb]

Got loads of old chips here - and plenty of them have dodgy screen-printing. So it could be fake, but I don't think that is cause of your problem.

Change from a 741 to an LM358 - as discussed earlier, 741 does funny things if you take the inputs to close to Vss.

[armandine2]

Found a LM358N in one of my parts containers - and a schematic in an old Maplin catalogue.

Yes works fine now.

just the math now to do - my clumsy stopwatch technique had it timed at 7.7s - 7.9s.  Might try and calculate the range for 5% resistors and a 20% capacitor.