Electronics > Projects, Designs, and Technical Stuff

Half-Bridge 0-30V, 0-20A Feedback loop problem

<< < (16/16)

T3sl4co1l:
You sense the inductor current and close a loop on it, so that inverter current can never be outside a perfectly well defined range.  The current setpoint is then driven by the voltage error amp (which has a well defined output range -- it saturates to +V/-V and some resistors can map this to the 0-max range current setpoint).  The two loops also only have one reactive component each (inductor, filter caps) so are very easy to compensate, whereas the voltage mode loop has an LC in it which is impossible to compensate properly.

You will particularly see the difference in transient source/load testing, including short circuits.

You can still add this in, by adding a current sense amp to the inductor (a high-CMRR diffamp, often purpose made e.g. INA195, or perhaps a Hall-effect sensor which is slower but even higher CMRR and voltage range), using the SG3525 for current error amp, and add an external op-amp for voltage error amp.

Tim

paladyn:
David Hess
Variable output voltage is not comfortable for smps but a lot of regulated power supplies are working in that way. Many comercial regulated power supplies like 0-30V 0-5A is working based on DC/DC converter


UPDATE:

T3sl4co1l
Do You have practical experiance with this type of control ? Becouse it seems this type of control is very interesting.
In my sytuation power supply is splited to 2 PCBs. One is power stage and second is for control so i thing will not problem to change voltage mode control to average current mode. I can do next PCB and I will have average current mode.

But.... New type of control need new knowlage how to compensate it and how to properly design schematic for this type of control.

Update 2:
I removed CMC to check how much he change characteristic of bode plot and I have to admit that it changes a lot. Attachments below:

Output = 5V

Output =12V

1.5V on error op-amp (non inverted pin) -> compensator was calculated for 1.5V reference voltage on op-amp.


I also spotted that changing reference voltage on op-amp(non inverted pin) little change crossover frequency and phase maring but not so much. Crossover frequency was calculated to 4kHz and as you can see gain on 4kHz is about 0dB and phase margin is about 67 deg. In my opinion feedback loop should be stable. Noises above 20kHz can be caused by 14BIT resolution of ADC on Analog Discovery 2. Am I right ?

T3sl4co1l:
I don't worry about compensation, analytically; I put in a typical value and test for step response.  The op-amp might be fed by a 10k series resistor, and might have an R+C feedback (compensation) network, say 1-220k and 100pF-10nF, with 10k and 1nF being typical, depending of course on modulator gain, inductor value, sensor bandwidth and op-amp bandwidth.  And for voltage, maybe the same resistors and say 47nF, then adjust from there.

Tim

paladyn:
Yesterday i made experiment I changed lower resistor(RF2) in output voltage divider to regulated resistor and set constant reference voltage 1.5V on non inverting op-amp as I assume during compensator calculation. Many publication says that lower resistor(RF2) not impact to compensation but practical test showed that is not true. Behavior of feedback loop is very similar when you change lower resistor(RF2) or when you change Vref voltage.

When i changed output load from about 1ohm to 4.5ohm and set low output voltage like 5V, feedback in some point almost reach 180 deg.

My conclusion is that it is not possible to have only one compensator for all output voltage. If i want regulated output voltage from 0-30V i need to use on example 3 different compensators calculated for 3 different vref voltages.

Can you tell me whether average current mode control allows regulated output voltage with out complex compensators ?

Navigation

[0] Message Index

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod