Hi,
I used evaluation design DC2337A-3.
Changed only transformer to WURTH 750813144 (for 21V output), FB resistor (R9) and R7 to 0.18.
Layout is also from DC2337A-3, only small changes due to transofrmer, and GND is polygon pour on both layers.
With no load and load 210mA, the output volatge is nice, noise is OK, No overheats, or audio noice.
I check DRAIN Waveform,
Vin 150V, yellow waveform (channel 1) is AC, D4 anode, other channel is DC DRAIN
No load
Load 210mA
I decide to follow Leakage Inductance and Snubbers section.
Start from 100pF, this give me already 2 times longer period
no load
load 210mA
I calculated Rsnubber (334k), and use it in series with Csnubber (100p) but no change in waveform (load and no load).
Cdrain is 19p
Lleak 2.12
Could you advice why I can't see improvements with RC Snubber?
Maybe I need same more changes/settings corrections?
Please advice
Slawek
If Rsnubber = 334k, definitely the calculation was incorrect. The calculated R value should be well below 1k. Check the units. Thank you!
I used for calculation whole oscillation time, for cdrain it is OK because in equation we use ratio between Tring(snubbed)/Tring. For Leak I should use time only one "sinusoid", not whole oscilation time…
Thanks, I will check it with 1k. I looked again into calculation and they looked OK according to equation in datasheet.
Csnubber = 100pF
Tring(snubbed) = 100us
Tring = 40us
Cdrain = 100pF / ((100us/40us)^2 - 1) = 19pF
Lleak = ((40us/(2*PI))^2) * (1/19pF)) = 2.12 units s^2/F ???
Rsubber = (2.12/19pF)^1/2 = 334k
I try with 1k and can see nice result, especially in section mark in white circle. Thanks, could you explain oscilation marked in pink one?
The above number is incorrect. It probably is 40ns.
Inside the pink one is called DCM (discontinuous conduction mode) ringing.
Thanks, last one, could you advise what can cause this ripple?
Yellow waveform (channel 1) is AC, D4 anode. Period 250ms.
I used for calculation whole oscillation time, for cdrain it is OK because in equation we use ratio between Tring(snubbed)/Tring. For Leak I should use time only one "sinusoid", not whole oscilation time. Than is OK