DCDC boost converter oscillates at low load but not high load - any ideas why?

[uoficowboy]

Hi - I have a synchronous voltage mode boost converter design using the TI TPS43000 that is giving me headaches. Its stability seems to be fairly load dependent, which doesn't make a lot of sense to me. With my current compensation network, it oscillates at low loads, but not high loads. Can anybody suggest why the load current is affecting my stability and how I can fix it? By my calculations the RHP zero (the only load dependent pole/zero that I'm aware of) is way off in the hundreds of KHz, where my gain should be well below zero. Plus - low loads should *increase* the frequency of the RHP zero, not decrease it. I should mention that all the loads I'm testing it at should still be continuous.

I'm oscillating at approximately the frequency of my LC poles. I wonder if high loads reduces the Q of the LC for some reason?

Thanks for any ideas you can offer!!

 

[crutschow]

It may be due to the change in inductor inductance due to saturation effects. What is the current rating of the inductor as compared to the maximum load?

 

[uoficowboy]

It may be due to the change in inductor inductance due to saturation effects. What is the current rating of the inductor as compared to the maximum load?

Inductor is a RLF7030T-4R7M3R4. I'm seeing oscillation at 0.3A out, 4.5V in, 5.1V out. It's stable at 1.6A out, 4.5V in, 5.1V out. I have not characterized where it loses stability. Switching frequency is about 500KHz. It's possible the inductance is drooping a teeny bit, but not a lot. Still, it is certainly possible that that is the cause.

 

[FvM]

The most common reason to get instabilities at low loads is the transition from CCM to DCM. Did you consider the respective gain variation?

 

[BradtheRad]

With a heavier load, your converter will try to maintain voltage on it. It may start entering continuous mode.

When this happens, there may still be up and down voltage swing on the sense resistor, but it is proportionately less due to the continual DC component.

 

[uoficowboy]

The most common reason to get instabilities at low loads is the transition from CCM to DCM. Did you consider the respective gain variation?

I had had a similar thought, but I believe I'm in continuous mode - from my math as long as my inductance is above about 1.5uH I'll be continuous with 300ma load at 4.5V in, 5.1V out.

 

[FvM]

The switching waveforms will tell you about CCM or DCM.

Another posible explanation would be very low stability margin so that small changes of gain or pole/zero can already cause oscillations.

If your test load is a resistor rather than a current source, the resistance will affect the output filter Q.

 

[crutschow]

You can do a rough check of the loop stability by applying a step load change and monitoring the output voltage. The output voltage should show no oscillations and only a slight or no overshoot if the loop is properly compensated.

A step change in the input voltage is also a good check while monitoring the output voltage, but that may be harder to do.