TOPSWITCH JX excess EMI from primary ringing.
At least we think it is from primary. It appears on secondary also, but seems to be common mode.
75W supply 100 to 240VAC input, flyback, output 24V 3A plus local power. Using available Wurth 750311269 transformer with 6 to 10 uH primary side leakage specified.
Primary side snubber is a standard RCD snubber with diode limiter (SMCJ188 TVS), 4.7nF capacitor with 16k discharge resistor, plus a series resistor of 68 ohms which seems to damp higher frequency ringing, but not the 90 mHz problem we have.
There is 2300pF from bus + to secondary common. (2x 4700PF "Y2" capacitor in series)
What we have is several cycles of 90 mHz ring following the beginning of the flyback pulse. This is the main EMI issue we have.
It has been very resistant to any fix so far, despite attacking it on both primary and secondary side. Primary side has been effective to some degree, secondary not, so we assume primary is the origin. We especially think this since it is very dependent on the load placed on the supply, much more at full load.
We are about ready to snub right across the drain to source of the TOPSWITCH.
Thoughts?
Comments
Hi,
Looks like you tried several things to minimize the 90MHz radiation emission issue, good.
If I understand well, the ring appears on the drain to source voltage during the turn off composed by leakage inductance and parasitic capacitors is the root cause of the emission issue. Some other things that I would think about are as following:
1. Optimize the diode RC snubber on the secondary side, and also do not forget RC snubber for the bias winding diode if there is any.
2. Is fast diode or slow diode is option or the primary clamping? The one you used in the ultra fast diode, sloe diode may helps to smooth the ringing.
3. A small capacitor (10pF) can be put in parallel with the diode of the primary clamp diode to help.
4. I have seen ferrite bead put in series with the primary clamp diode helps for radiation emission too.
Hope this is helpful.
Yes, had not considered primary diode capacitor (your 10 pF).
YES IS uf diode, but supposedly softer.... no a/b time spec given, so unknown HOW soft it is.
Yes bias....but low current, and ring is load dependent, high load high ring.
Slower diode is a good idea, we just have to find one.
Was considering bead on transformer lead.
Will investigate
Thank you
Using a slower diode (RS1M) does help generally, but is definitely not a cure.
There is another new issue, and that is a startup problem, for which I made a new thread..
This is still not completely resolved, but is considerably better. I think good enough for the system EMI filter to take care of.
We have it reasonably well under control by the use of a snubber across the topswitch drain-source, plus making the secondary snubber on the 3A 24VDC output rectifier lower impedance.
The drain-source snubber is 33 pF/ 33 ohms (no diode), and the secondary 1 nF/3.3 ohms. Plus use of the RS1M 500 nS snubber diode. We are not using a diode on the 33pF snubber, it does not seem to affect temperature.
The startup problem seems to have been an un-noticed layout problem, our fault.
Hi,
Very glad to hear that you got a solution. Good job.
It would be better to be sensitive enough. - Steven C Wyer
Snubbing across topswitch mosfet had some effect, but not enough. Rise time is helped, ringing is not well damped even so. 220pF and 1500 ohms seemed to be best, but does not damp the ring.
rectifier (STT802) has 1NF and 56 ohms, which appeared to be optimal for it.
Rectifier for primary snubber is STTH112, allegedly a soft recovery device.