741 opamp used to control linear regulator

cupoftea said:

We think this could be due to the 741 opamp going unstable, (due to the feedback loop components around it) then when its output goes near the positive rail, its latching there, and turning the pass transistor fully on, resulting in our output overvoltage........dduring the fault condition, the pass transistor is actually oscillating ON/OFF/ON....etc, at high frequency...but this is resulting in the high vout.

Click to expand...

You are apparently guessing, but no actual measurement confirms the assumption?

Surely a schematic review can clarify if it's at least possible.

Hi,

have you tried to prove your assumption by means of a Spice simulation? TI provides a Spice model, which covers for sure not the full behaviour of the realized silicon, but migth cover the behaviour of your current setup.

I would also like to see the schematic of the realized circuitry, which shows the actual implementation and the additional used components.

BR

Thanks, it is just a pretty standard linear regulator, but , as you know, we are in the days where even basic standard electronics is considered to be highly Intellectual product, and i could not deliver the full schem, unfortunately.

What would happen if a 741's output got within 0.5V of its +ve rail?

Thanks for the spice model idea....i wonder if the TI one is an updated version of the original 741?
I suspect this circuit might be using the "original 741".

Hi,

The 741 seems to be still alive. I guess I never used it. Maybe because even in the late 1980 when I started to design analog electronics I found that there are much better Opamps.
UA741 is from 1968 ... now we have 2022.

What other thing from the 1960ies is still used in modern designs? (Besides historical reasons)
I guess the 741 is still used just because schematics are copied and copied without thinking that Opamps have developed and the requirements have changed. Now we need faster response, lower noise, less overshot, lower supply voltage...

For me it's no surprise the 741 can not satisfy nowadays needs.

Btw:
* there are ready to buy linear voltage regulator ICs.
* and I personally don't like the idea of dissipating 30W of heat with a large heatsink and a fan, (dirt, noise, failure rate...) thus I'd avoid it somehow. In detail it surely depends on the application's requirements.

Klaus

Thanks Klaus, i believe its marked "uA741", thats what i was looking for, thanks...aaah here it is...

Unfortunately the datasheet doesnt really tell what happens when its output voltage gets close to its rail?.......i once had a much more modern opamp and its output latched up to the positive rail and stayed there, whenever its output got to within 0.5V of the positive rail.

...i suspect the uA741 does the same type of thing?

cupoftea said:

Thanks for the spice model idea....i wonder if the TI one is an updated version of the original 741?
I suspect this circuit might be using the "original 741".

Click to expand...

According to the Spice-file head-note, the copyright referes to National Semiconductor. So I assume it is an "old" Spice model as National Semiconductors has ben acquired by TI about 10 years ago.

Hi,

cupoftea said:

Unfortunately the datasheet doesnt really tell what happens when its output voltage gets close to its rail?....

Click to expand...

It tells the expectable voltage level limits and that it does not latch up.
What other informations do you need?

cupoftea said:

.i once had a much more modern opamp and its output latched up to the positive rail and stayed there, whenever its output got to within 0.5V of the positive rail.

Click to expand...

I guess this indeed is not an output problem, but rather an overloaded input problem. Maybe during output saturation the input
* got saturated
* or out of specified common mode input voltage
* or out of specified differential input voltage

I don't expect the uA741 is critical in described situation.

To validate the real problem you need to monitor the input voltages, too.

Klaus

TI LM741 datasheet states no latchup -




Regards, Dana.

Thanks, also i am wondering if the reg has gone unstable (well i know it has)...and when the error amp slams to the positive rail, it saturates the three paralleled 2N3055 output transistors, and they stay on such that vin = vout?
Could that be it?....ie saturation, meaning the 2N3055's take a few us to get out of sat, and in that time...current surges through them?
2N3055

Hi,

you are using three transistors in parallel each with a maximum current gain of 70 (Ic = 4 A @ VCE = 4 V).
The u741 can source a maximum short circuit current of 25 mA.

This leads to 3 • 70 • 25 mA = 5.25 A. I'm curious what's going on when driving a larger current than the mentioned 5 A. As we do not know how your actual circuitry looks like, I would recomment a Darlington-Type configuration, to increase the current gain and reduce the current which has to be sourced by the opamp to regulate the output voltage.

Have you tried to simulate the circuitry? The 2N3055 is available in LTSpice.

BR

Some background -



A look at schematic is everything in this discussion if you want accurate advice....

Note is also possible to have unstable thermal feedback loops (an example of a stable but
transient loop) -


Of course control loop affected by T rise (G change) -



OpAmp currents required to sat pass element -



And what C load do you have at output of supply ? That cap affects stability.
Although in NPN pass element architecture generally not a problem. But analysis
is everything.....

Lastly some web circuits also place bypass cap directly on OpAmp output,
thats generally speaking not a good idea, reduces phase margin. Unless OpAmp
of choice comped for high C loads.

One more item, there have been more than one revision to 741 design, so even though
TI datasheet says no phase reversal when violating common mode I would not bet my life
on that. Set it up on bench/proto and check that yourself.




Regards, Dana.

No problem to make a stable linear power supply with uA741 and 2N3055. Thousands of products have been designed in the seventies. The problem isn't with uA741 or 2N3055 as such but how they are used in the circuit, e.g. how's 2N3055 driver designed, what's used uA741 common mode range.

Useless talk without substantial information. There are too many possible ways to make the circuit fail.

Hi,
Which of the two linear regulators attached is best, top or bottom?....LTspice and PDF scheam attached.
The drivers of the bottom one are not so slow, as the Darlington wont saturate as much?

2N2222 driving three 2N3055s ?

Performance wise, not much difference.

Brian.