rectifier waveform

Having read what else is going on in this thread I will not give you the example I almost did in the previous post which did not come up whilst looking at the behaviour of a forward converter and the effects of leakage inductance and diode reverse recovery and how magnetising current transfers from one winding to the other one and drives the.........

Ah, but of course I'm not going to give you and example of that sort of thing because I used Spice to find out what was going on......

FUCK OFF

DNA

No, it was built in 1992, all up to code, including the seismic stuff. But things break after 14 years.

Our previous house was built in 1892, and it had a lot of problems, including flakey knob-and-tube wiring and a brick foundation. The brick chimneys broke up in the '89 earthquake, admighting daylight in a couple of places, and we liked that so we plugged the holes with skylights.

Overall, I like the new one better. Less character but a whole less hassle.

John

John, with the right model, a simulation should get it right. Meaning it should have some capacitive coupling between primary and secondary. Such that the floating secondary goes to mid line voltage, in your case approximate 55V AC

Rene

the

I used a 'kit' transformer in the sim. With a K less than 0.995, the 500kc ringing makes the sim very tedious. Real life, a 100MHz scope, 100ma and IN4005's showed no nasties, pretty much like John F's pics but smoother. A floating measurement using two probes on the 14V~ bridge input gave 17V~ and 9V~ in quadrature. Ties in nicely with a 240V supply and measured 30pF interwinding capacitance. john

A slightly 'clipped' sinewave.

Graham

Graham

How fast are the rectifiers you're using ?

Graham

You didn't say that before. They usually are.

Graham

Grrrr. My ISP has dropped binaries.

Graham

That figures. That's why 'linear' psus also need some EMC filtering. A few caps normally suffice. Try 100n across the diodes and see what happens.

Graham

Yes....

but

You design forward converters without output inductors......

I also notice that you might be an OLD fart prone to complaining about society and youff.

Excuse me, old fart who has to work until he is 68.....

What the f*ck did you do with your previous 67 years to ensure that society would be gracefull to you in your old fartdom?

Let me guess....... FUCK ALL.

In fact you protected your ass whilst being shit and now some cunts probably snurped the money you thought you invested in your pension...... and you did nothing to avoid the problem.

Oooooooooh let's talk about the good old days.

And now you are moaning about it.....

Sorry, you spent 67 years f****ng up and now I get a job wiping your arse in the Old Farts Home.....and you want to complain about how you f***ed up my life.....

Oh Dear, seems like Sir will sleep in shitty sheets tonight.

DNA

No. There is actually some small inductance there. It's more of a power inverter than a converter though.

That's funny actually. I'm a lot more in touch with youff than most ppl my age. I simply recognise where education is failing.

You're miles off my age.

STFU yourself.

Graham

Don't have abse posting ability anymore, but here's what an actual waveform looks like, in a circuit which would use a simple ST and clamped low-pass filter:

Test circuit:

Best regards, Spehro Pefhany

I was puzzled by the asymmetry until I saw the way you'd connected the transformer. If you grounded the CT I'd expect a symmetrical clipped sinewave.

Graham

Rarely!

John

"slightly clipped?"

I'm sure it'd sound OK thru monster cables.

Cheers Terry

Uh ? Always IME. So what do you think it is ?

Graham

Call it 'flat-topping' if you like.

About as OK as 50/60Hz ever does.

Graham

It could vary over a wild range of waveforms. Helmut did some really nice sims... I hope he can post them to a.b.s.e. John F has posted some actual scope waveforms which include interesting features.

I'm just now appreciating how peverse this circuit really is. How about this one:

Call the ends of the transformer X and Y (Speff has already used up A and B.) Suppose we're at the top of the cycle, and that X is the positive side. Two diodes are now conducting, call them Dx and Dy, where Dx is charging the cap + side and Dy is clamping Y to near ground. Everything is nice and symmetric.

Now, after the sinewave peaks, voltage X:Y is dropping below the cap voltage, so we are about to turn the diodes off.

First assume diodes that don't store significant minority carriers, like schottkies or fast-recovery pn diodes. They still have a severely nonlinear capacitance vs backbias curve. As the winding voltage droops, "reverse" current flows through the diode capacitance. Since the winding stray capacitance to ground is of the order of the diode capacitance, the effects on secondary voltage can be nontrivial.

The higher the reverse voltage on any diode, the lower its capacitance, so the faster the reverse voltage increases. That's a positive feedback mechanism, and if one diode has slightly less capacitance, it will lose the tug-of-war, and the corresponding transformer terminal voltage will "jump off the rail" first. This jumping off the rail is visible in JF's schottky waveform, although it is pretty symmetric in his case, and there are complex reasons for that, too.

Things get even more interesting if the diodes store lots of charge, like a slow pn rectifier. Helmut simulates some astounding asymmetric spikes, and I've seen stuff in real life that was almost as bad, but with lots more ringing.

And serious primary-secondary capacitance can make the secondary waveform even more complex, as can asymmetric diode reverse leakage, which can be significant with schottkies.

This really *is* interesting.

John