feedback circuit question

Thread Starter

franticAAC

Joined Jun 25, 2012
30
Hi, i've to design a feedback for my isolated push-pull converter.
The feedback is an OR between voltage and current sense.
In that way my DC/DC should work as a laboratory power supply.
I've drawn the schematic in figure below

fb.jpg

Apart from compensation network and from the voltage and currents involved I have some doubts about how i expected the power supply of Operational Amplifier. I'm not sure that the differential stage for current sensing could work in such configuration.

Could you help me?
Could you suggest me the most suitable configuration?
Thanks
 

AnalogKid

Joined Aug 1, 2013
10,987
The basic plan is ok, but...

The diodes connected to the opamp outputs are reversed, and the signal polarity might be reversed depending on what the "feedback pin" wants to see.

For the high-side current sense amp you need one with an input common-mode voltage range that extends above the positive supply input. You can buy one fully integrated with the shunt built in from Linear Tech, Maxim, TI, and maybe Analog Devices.

ak
 

Thread Starter

franticAAC

Joined Jun 25, 2012
30
In real schematic the diodes are in a right way... i make an error in redrawing it.
What do you think about TSC103 for current sense?
Thanks
 

ErnieM

Joined Apr 24, 2011
8,377
Where to start….

You need an isolated 12V supply for the op amps or this isn’t an isolated supply.

Both amp’s output will be stuck at zero (assuming the OPA2140 goes positive with increasing current) as they are inverting amps, and if you put something above zero in they try to go below ground.

For current sensing google “high side current monitor.” Several firms make nice custom ICs just for that. Cheap, accurate and they work with fewer external parts. The TSC103 is a good part but does not sense down to zero volts without a negative bias supply. I do not know of any part that does work to zero.

A definition of what you mean by “laboratory power supply” is necessary. If you want an adjustable voltage and adjustable current limit this is not the way to go, you need two separate loops for that.
 
Last edited:

AnalogKid

Joined Aug 1, 2013
10,987
There is nothing in the original post that suggests that the opamps are powered off the primary side of the converter, so this isn't the place to start.

Both amps have a non-inverting input reference voltage of +2.5V. The outputs won't "try to go below ground" unless the inverting input is above +2.5V. That is how opamps work. And when the output do try to go negative, they will have almost 2.5 V to play around in.

A lab supply usually has independently adjustable output voltage and output current limit, with the current regulator overriding the voltage regulator when the current limit setpoint is exceeded. Frequently this is done by stealing drive current away from the pass transistors through an LED, so you get a visual indication when CC control kicks in. In #1 above it is done by overriding the voltage feedback signal, another common technique. In fact, if he replaces the two opamp output diodes with LEDs he will have visual indication of which loop is in control. The design has two separate loops diode-ored into a single pass element controller. Only one loop is in control at a time, the same effect as two fully independent loops with separate pass elements.

Note:
Something I forgot to mention before is that there is an error in both error amplifiers. As drawn, both amps try to go to infinite gain at DC. There must be a DC path of some kind from the output to the inverting input to close the feedback loop; yours are blocked by capacitors. The loop might be closed through the primary side control of the bulk DC, with the two opamps operating as integrators, but I think you are asking for oscillation troubles to start out this way.

ak
 
Last edited:

Thread Starter

franticAAC

Joined Jun 25, 2012
30
Something I forgot to mention before is that there is an error in both error amplifiers. As drawn, both amps try to go to infinite gain at DC. There must be a DC path of some kind from the output to the inverting input to close the feedback loop; yours are blocked by capacitors. The loop might be closed through the primary side control of the bulk DC, with the two opamps operating as integrators, but I think you are asking for oscillation troubles to start out this way.
I don't understand very well... could you explain me better?

Have you any application note or example of smps that has this kind of feedback from which take ideas?

I need to sense current up to 1A not negative ones...
 
Top