Gap, Creepage and Clearance for HF HV

acrylic or if available styrene both have excelent film forming and HV breakdown properties Paul Mathews

Get or make yourself a proper HV differential probe (like Tek P5205) if you're going to keep doing this type of work. There's nothing much to them if you want to build your own: a good fast diff amp chip and a balanced input attenuator driving a 50 Ohm buffer amp. However, the input networks have to be well matched or CMRR will be poor. Paul Mathews

Paul Mathews

Yeah..I know. I plan on redoing the layout for the power section of the smps with industry clearances. Then if nothing blows up...I'm into EMI emission problems... Oh joy :(

Most offline switchers are still operating down under 150 kHz, largely for EMI reasons. Differential mode EMI is more easily filtered at 700 kHz compared to lower freq. Common mode components at 700 kHz and harmonics and much higher freq stuff at the parasitic ring freqs will predominate. Easiest way to reduce your EMI without lowering clock rate: slow down switching transitions and take an efficiency penalty. It's all about I = C dv/dt and V = L di/dt with the parasitic Ls and Cs, so reduce di/dt, dv/dt and/or get those parasitics down. For the higher freq components (ringing), softer rectifiers and optimal snubbing. Paul Mathews

Lower the dv/dt & di/dt... :(... noooooooooooo....

The fall time on the power mosfet drain is immeasurable with my 20Mhz scope.. It's a vertical line on the screen at max sweep. It's beautiful :)

(Note: My smps prototype functions ok at 50% Vin. Parts burned out at

I'll give it a shot 1st to tame components and layout parasitic's for less buzz for those country/western AM listeners.. Failing that..Then I'll make everything sluggy..(reduce the I&V rates) That should keep the black helicopters away :)

Thanks for the pointers..Good stuff..

D from BC

Life is short. Get a better scope soon. Your rectifer circuits are probably ringing at 80 Mhz or more. Paul

Is that a ring after reverse voltage?

You mean the power rectifiers as part of the converter circuit.. Yeah...I've considered SiC diodes and currently using a hyperfast power diode.. mmm Hyperfast = hyperEMI ? Maybe I can do the dum thing and just assume it rings and do a standard fix whatever that might be..dunno..

Yeah...gotta get a better scope... The scope I have is so old...(punchline).. the manual can be read in 5 minutes. D from BC

When the rectifiers turn off, their junction capacitance forms a tank circuit with any connected stray inductances. This can have quite a high Q, and it repeats at 700 kHz in your case. Also, the ring amplitude adds to the voltage stress on your rectifiers and on the switching transistors(s), reflected through the turns ratio of the transformer. Soft recovery rectifiers help. If the inductances are low, very small RC snubbers can damp down the ringing to a single critically damped overshoot with minimal EMI consequences. There's no excuse for not having a high BW scope these days. Paul Mathews