sub harmonic oscillation at low duty cycle

Hi Guys,

Im running an active clamp forward design (LT3752)

Vin 170-350V
Vout 24 @ 1.25A

For some reason at 200V input the system seems fine, as soon as the vin goes to 240V+ i seem to be getting some sub harmonic oscillation (Gate goes to small and large pulses). Now I thought this normally happens at higher duty cycles (closer to 50%), but it seems to be happening at the lower duty cycle end. Any ideas or experience with why this is the case? Is it a case of trying higher slope compensation?

If you are bumping up against a minimum pulse limit (like,
the comparator's ability to reswitch in short time) you
can fall into a pulse skipping kind of behavior.

Current mode PWM right about 50% can have trouble
with the opposite sync or output edge "kicking" the
current compare forward or back, via the supply
network.

Gate goes to small and large pulses

Click to expand...

There is such a thing as hiccup mode, where a large current burst occurs, causing overshoot of output voltage. This makes the converter skip operation for the next few cycles. The converter hiccups periodically.

Very low duty cycles reputedly are difficult to manage.

Oscillations can result if feedback signal (a) is slightly delayed, or (b) has overmuch gain applied to it. Either condition can result in overcorrections, leading to unwanted oscillations.

your volt loop is unstable, the higher Vin is raising the gain of the power stage and your volt loop is not sufficiently compensating...

or, you have minimum duty cycle issues with the IC...

- - - Updated - - -

adding some load should tell you which it is...

There is such a thing as hiccup mode, where a large current burst occurs, causing overshoot of output voltage. This makes the converter skip operation for the next few cycles. The converter hiccups periodically.

Click to expand...

I thought that hiccup mode was more associated with missing pulses or gaps in the drive waveform? This tend to be a periodic gate waveform but with long and short on time pulses is that not a feedback related issue or is this present in hiccup mode too?

Oscillations can result if feedback signal (a) is slightly delayed, or (b) has overmuch gain applied to it. Either condition can result in overcorrections, leading to unwanted oscillations.

Click to expand...

the higher Vin is raising the gain of the power stage and your volt loop is not sufficiently compensating...

Click to expand...

Is there an easy way around reducing the gain to see if this is the case? Sorry but my understanding of feedback compensation is very limited.

add some load e.g. 50% to 100% load, this will damp the power stage a little, if your osc goes away for Vin approaching 240VDC then your volt loop needs correcting...

I see this sort of behavior when there's insufficient blanking on the current sense signal. Putting a more aggressive lowpass filter on the current sense pin often helps.

add some load e.g. 50% to 100% load, this will damp the power stage a little, if your osc goes away for Vin approaching 240VDC then your volt loop needs correcting...

Click to expand...

could the input filter I have chosen have an effect on this?

bowman1710 said:

I thought that hiccup mode was more associated with missing pulses or gaps in the drive waveform? This tend to be a periodic gate waveform but with long and short on time pulses is that not a feedback related issue or is this present in hiccup mode too?

Click to expand...

Your small and large pulses could come from a combination of factors. This includes the 'hiccup' kind. I brought it up as a catch-all term which might appear in troubleshooting articles.

I must defer to the other replies because their suggestions are from a more expert viewpoint than mine.

Is there an easy way around reducing the gain to see if this is the case?

Click to expand...

I imagine the control IC always applies maximum gain internally. On the other hand a control circuit made from op amps could provide less abrupt action in your switching, and allow a means for you to adjust feedback response.

Do the irregular pulses appear suddenly at 240V? Or do they grow in amplitude as you increase voltage? If suddenly then they might arise from on/off action, which implies it's from the control IC. But if gradually then it could be from analog (linear) action, which implies other components.

I'm merely brainstorming for concepts which might allow you might track down the problem.

Do the irregular pulses appear suddenly at 240V? Or do they grow in amplitude as you increase voltage? If suddenly then they might arise from on/off action, which implies it's from the control IC. But if gradually then it could be from analog (linear) action, which implies other components.

Click to expand...

The pulses get worse as the voltage increases, it tends to become more erratic and skip more pulses ad the voltage increases. I tried changing the turns ratio but a bigger turns ratio just pushes up the voltage (i.e 290V now) at which it becomes unstable, so must be a duty cycle related issue i guess??

Although the answer could very well be in any of the replies already given, it would help if you were to post your schematic, and oscilloscope waveforms.

It is possible for the inductor to interact with other components as voltage is turned on and off. (Example, high voltage spikes if disconnected abruptly, ringing with a nearby capacitor, etc.) The effect often gets worse as you increase the power.

To narrow down the source of the problem, consider the substitution of manual control instead of the IC's control. If the problem stops, then your control IC probably is defective.