Issue with opamp tending towards wrong rail under load

[Dextrine]

Hello everyone,

I am having an issue where the output of the op amp for the following circuit works fine under no load and allows me stable output voltage.

When I close the switch and apply the load, the output voltage decreases (as one would expect), however, the op amp output doesn't increase so the duty cycle doesn't increase and the voltage stays constantly below my target.

Any ideas as to what would make an opamp not give a higher output when the non-inverting pin is higher than the inverting pin? Have any of you encountered something like this before?

Thanks for any help!

 

[Baluncore]

Some op-amp outputs invert or latch, when the input voltages are close to the supply lines and out of the common mode range.

What is the make and model of the op-amp ?
What are the op-amp power supplies connected to ?
What are those supply voltages relative to ground ?

 

[davenn]

I am having an issue where the output of the op amp for the following circuit works fine under no load and allows me stable output voltage.

What is the make and model of the op-amp ?
What are the op-amp power supplies connected to ?
What are those supply voltages relative to ground ?

and further to Baluncore's questions

what are the resistor values?
what is the load "resistor?" value ?
what is the capacitor value ?
what is the "output voltage on that point to the left of the reference voltage point ?Dave

 

[LvW]

.. and another important aspect: There is no dc negative feedback. Therefore, the circuit will NOT work in case of real opamps resp. real opamp models. Only ideal models without any offset can be used.

 

[Baluncore]

Only ideal models without any offset can be used.

I don't think that is the case. The op-amp is being used as an error amplifier, set up like a PID controller without the D. The capacitor forms a miller integrator, which Integrates the error voltage as charge in the capacitor, and so adapts to the required duty cycle. The series resistor provides the Proportional component.
The DC feedback passes through the PWM switched power supply that is not shown on the RHS of the diagram. That supply generates the “OutputVoltage” that enters the diagram on the LHS and so closes the feedback loop.

 

[jim hardy]

Baluncore probably nailed it in post 2.
What is supply voltage ?
I learned that exact lesson with LM324 . It'll work with inputs at negative rail(0volts) but it can't handle input within 1.5 volts of + rail . To use it with a 5 volt input you must power it from at least 6.5 volts.
Try making reference voltage = half Vsupply and adjust divider ratio at Measured voltage.

 

[Baluncore]

25 years ago I discovered that if the the output of an LM324 was insufficiently loaded, it produced a transient spike each cycle. When used as a high-Q filter, the LM324 ignored the input signal and locked to the internally generated transient spike.

 

[Dextrine]

The op amp is a lm2902 (http://www.ti.com/lit/ds/symlink/lm2902-n.pdf).

The opamp inputs are around 5V and the op amp power supply is provided from my dc power supply at 13.5V, because of this, i think it may be unlikely due to being outside the common mode range.

The loads i have tried are 250R and 1K.

At 250R, the opamp output drops down lower than at 1K load. this is odd since with 250R, the inverting pin is even lower so you'd expect the opamp output to be even higher.

 

[Dextrine]

Just confirmed that even without the load, the opamp output is increasing when the - pin increases. So it seems as though in all cases, the opamp output is tending to behave like the - pin.

 

[Baluncore]

The LM2902 is an LM324. What are the other three op-amps in the package being used for?

You are looking at the performance of a loop that includes an unspecified component, namely the PWM switch. Check the polarity of that PWM switch. Does an increase in the PWM control signal voltage result in an increase in OutputVoltage? or have you accidentally built two inverters and cross connected them? Without a full circuit we cannot diagnose the problem.

 

[davenn]

The op amp is a lm2902 (http://www.ti.com/lit/ds/symlink/lm2902-n.pdf).

The opamp inputs are around 5V and the op amp power supply is provided from my dc power supply at 13.5V, because of this, i think it may be unlikely due to being outside the common mode range.

The loads i have tried are 250R and 1K.

At 250R, the opamp output drops down lower than at 1K load. this is odd since with 250R, the inverting pin is even lower so you'd expect the opamp output to be even higher.

you didn't really answer all the questions

edit your circuit and show ALL component values. show ALL voltages including the pins to the op-amp

As Baluncore said, it's really difficult to help you with the lack on info from youDave

 

[jim hardy]

op amp power supply is provided from my dc power supply at 13.5V, because of this, i think it may be unlikely due to being outside the common mode range.

agreed. Perhaps you'll add that to your drawing ?

Are you reading everything with a DC meter ? Which of course reports average ?

...........................

How did @Baluncore figure this out ?

The op-amp is being used as an error amplifier, set up like a PID controller without the D. The capacitor forms a miller integrator, which Integrates the error voltage as charge in the capacitor, and so adapts to the required duty cycle. The series resistor provides the Proportional component.
The DC feedback passes through the PWM switched power supply that is not shown on the RHS of the diagram. That supply generates the “OutputVoltage” that enters the diagram on the LHS and so closes the feedback loop.

Well THAT certainly changes things.
First thing i'd do is take AC readings at same points. That'll tell you if you have built a giant closed loop oscillator.
Better yet hang a scope on it.

 

[Dextrine]

A bit of a late reply, but I figured out what was wrong. In case anyone else ever encounters a similar problem, what I found was that I accidentally had one of the op amp rails grounded to another part of my circuit. Fixing this got rid of the error. So just a heads up to anyone (especially beginners with soldering) make sure everything is connected where it should be and NOWHERE else! Thanks everyone for all the replies and all the help.