Plateau on switch OFF current in flyback design

Hi kwa,

Thanks for your detailed report to fully illustrate your concern. Just a few questions regarding the waveforms:

Start-up sequence:
In your first measurement, it says there that the time/div is 1.00ms if I am not mistaken. For the succeeding start-up waveforms, were the zoom waveforms (10.0us and 100ns) taken from the same initial start-up waveform?

Steady-state with 47R load:
Did you use the same start-up sequence waveform (1.00ms time/div) and zoomed in to show the next two waveforms for the steady-state operation (4.00us and 1.00us) ?

Furthermore, are you experiencing any problems with the output voltage? Is it going into AR? One thing you can try is to probe just the drain current to confirm if there is indeed a plateau. What you are seeing might just be due to the resolution of the oscilloscope especially if you captured the waveform at 1.00ms/div and later on zoomed in to the us and ns range. Try setting the scope at that range (us and ns/div) initially and then you may also want to try removing all other probes except the drain current one. I hope this helps.

Regards,

Hi
Start up sequence... No... the subsequent waveforms were not take from the startup sequence. Instead I repeated the startup sequence, decreased timebase and adjusted the trigger to trig on peak primary current, so I would get a high resolution sampled picture. None of the pictures was taken using the scope zoom function.
Steady state waveform ... No .... again... no pictures were taken using scope zoom. The were taken by a single trig (not auto trig) at the right current trig level and at steasy state long established.
Furthermore.... I did not experience any problem with the output voltage.
What does "AR" mean ?
Sampling just the drain current.... it opted to place the current probe at the place where there is no voltage swint, so that potential current probe CMRR is not a problem. As you see I made measurements without the RCD snubber and hence measured only the drain current.
Measurement technique... it is not a measurement / scope sampling problem !
I have made these kind of measurements on flyback for the last 25 years using the same kind of measurement method. From 1W to 350W
This is most likely no 14 flyback transformer I measure.
I have never seen this current ramp down plateau before.
wkr. Kenneth

Attached you will find a similar masurement on a 80V DC 2W flyback from linear.
It has no plateau on the primary current cut-off slope. wkr. Kenneth

Hi,

I'm sorry for asking on how you measured the waveforms (especially about the oscilloscope resolution); I just wanted to get it out of the way first before I go deeper into the analysis. By AR, I mean auto-restart. I was wondering if you were having problems with the output voltage/current while you were observing this phenomenon. Allow me to make measurements of my own before I get back to you on this. Thanks.

Regards,

Any news ? Have you had the opportunity to measure the "primary side ramp down" current yourself ? wkr. Kenneth

Hi,

I did some verification on my own bench and found that my drain current waveforms are consistent with yours. I've consulted some of my colleagues regarding this; there is nothing to worry about. It is normal behavior for this specific family of ICs and has nothing to do with the design. It is internal to the IC and should not affect performance of the power supply.

Regards,

Hi.
Can you explain it a bit more ?
I understand that you mean it is not a problem, but this is the same as buying a car with a unexpected motor sound and having the salesman tell you "Not to worry, it is supposed to sound this way".
That might be the case, but only if I understand why can I rest assured.
wkr. Kenneth

Hi,

According to our team, it is most likely due to the Miller effect presence in the high voltage device. During transistor turn-off, as the drain voltage rises, it turns-on the transistor again due to the drain-to-gate capacitance and the Miller effect. During the plateau, the gate driver continues to pull the gate down which eventually turns-off the high voltage device.

Regards,