rectifier waveform

Wouldn't that be a question for a Spice simulation ?

Rene

Ratty. 3.2V recovery spike... my LM339 circuit doesn't even notice ;-)

...Jim Thompson

This is just the kind of situation that Spice would get all wrong.

John

If memory serves me right this was given a beating in the past.

Yes.... spice will give you the right answer but you might have to spend some brane power thinking about the result you expect and beat it into submission so the 'model' is correct.

No, I'm not going too....

DNA

Yup. Keep bashing the Spice model until you get the waveforms you expected to get.

John

You're so full of it, John! Why do you think Spice will get it all wrong?

...Jim Thompson

You're shooting from the hip. The winding ISN'T floating.

...Jim Thompson

Exactly, because the xfmr secondary is floating around most of the time. Put some R's on there and Spice will probably get it right.

Best regards, Spehro Pefhany

Got a Spice sim? Show us the waveform.

John

Show me your measured data and I'll show you my simulation ;-)

I must admit some surprise at the size of the step!

...Jim Thompson

"John Larkin" schrieb im Newsbeitrag news: snipped-for-privacy@4ax.com...

Hello John,

I tried a few minutes ago with LTspice. No problem. It looks correct! I can send you a screenshot if I get your email address.

Yes I know that sometimes there is a litlle bit help necessary. This depends on the diode models, the load, the timestep ....

Best regards, Helmut

The used diode model is from Diode Inc. The long '.model'-line may be broken. Please repair it before you run the circuit. Name the file with any name ending with .asc . Then run it with LTspice.

Version 4 SHEET 1 1036 680 WIRE -320 208 -320 96 WIRE -320 352 -320 288 WIRE -48 96 -320 96 WIRE -48 352 -320 352 WIRE 48 352 -48 352 WIRE 48 352 48 144 WIRE 80 96 -48 96 WIRE 80 96 80 48 WIRE 80 256 80 96 WIRE 128 48 80 48 WIRE 128 144 48 144 WIRE 128 256 80 256 WIRE 128 352 48 352 WIRE 240 256 192 256 WIRE 240 320 240 256 WIRE 240 352 192 352 WIRE 240 352 240 320 WIRE 352 48 192 48 WIRE 352 144 192 144 WIRE 352 144 352 48 WIRE 352 224 352 144 WIRE 352 320 240 320 WIRE 352 320 352 288 WIRE 352 352 352 320 WIRE 448 144 352 144 WIRE 464 320 352 320 WIRE 704 144 608 144 WIRE 704 192 704 144 WIRE 704 320 608 320 WIRE 704 320 704 272 WIRE 704 352 704 320 WIRE 816 144 704 144 WIRE 816 192 816 144 WIRE 816 320 704 320 WIRE 816 320 816 272 FLAG 352 352 0 FLAG 448 144 V+ IOPIN 448 144 Out FLAG 464 320 0 IOPIN 464 320 Out FLAG 608 144 V+ IOPIN 608 144 Out FLAG 608 320 0 IOPIN 608 320 Out FLAG 704 352 0 FLAG -48 96 b FLAG -48 352 a SYMBOL Misc\\\\signal -320 192 R0 WINDOW 123 0 0 Left 0 WINDOW 39 24 132 Left 0 SYMATTR SpiceLine Rser=0.1 SYMATTR InstName V2 SYMATTR Value SINE(0 17 50) SYMBOL polcap 336 224 R0 SYMATTR InstName C1 SYMATTR Value 4700µ SYMATTR SpiceLine Rser=0.01 SYMBOL res 688 176 R0 SYMATTR InstName R1 SYMATTR Value 20 SYMBOL current 816 192 R0 WINDOW 3 31 73 Left 0 WINDOW 123 0 0 Left 0 WINDOW 39 32 100 Left 0 SYMATTR Value 0.5 SYMATTR SpiceLine load SYMATTR InstName I1 SYMBOL diode 128 64 R270 WINDOW 0 32 32 VTop 0 WINDOW 3 0 32 VBottom 0 SYMATTR InstName D1 SYMATTR Value DI_1N5400 SYMBOL diode 128 160 R270 WINDOW 0 32 32 VTop 0 WINDOW 3 0 32 VBottom 0 SYMATTR InstName D2 SYMATTR Value DI_1N5400 SYMBOL diode 192 240 R90 WINDOW 0 0 32 VBottom 0 WINDOW 3 32 32 VTop 0 SYMATTR InstName D3 SYMATTR Value DI_1N5400 SYMBOL diode 192 336 R90 WINDOW 0 0 32 VBottom 0 WINDOW 3 32 32 VTop 0 SYMATTR InstName D4 SYMATTR Value DI_1N5400 TEXT -376 -208 Left 0 !.tran 0 100m 0 100u TEXT -376 -168 Left 0 !.options plotwinsize=0 ;data compression off TEXT 736 104 Left 0 ;LOAD TEXT -320 64 Left 0 ;Transformer TEXT -376 -104 Left 0 !.MODEL 1N5400 D ( IS=63.0n RS=14.1m BV=50.0 IBV=10.0u CJO=125p M=0.333 N=1.70 TT=4.32u ) RECTANGLE Normal 912 384 640 80 RECTANGLE Normal -128 384 -384 32

it bounces up and down.

lumping all inductance into the xfmr (so none elsewhere), then at any point when current is flowing through the rectifiers, the transformer voltage is clamped. one winding will be a diode drop above +Vdc, the other a diode drop below -Vdc, and it swaps over every half cycle. its a source of common mode noise - although its primarily LF, the rectifier recovery generates HF CM noise.

when no current flows through the rectifiers, the transformer voltage will be reflected primary volts (which, by definition, is less than output voltage + diode drops) and it will be bouncing around as a function of whatever strays are going on.

Try synthesising a 3-phase waveform with three half-bridges (for simplicity assume split DC bus). no matter how hard you try, the sum of the three half-bridges will always = +/- Vdc/2, and it bounces around not at 50Hz, but at the switching frequency, which is a BIG problem with EMI. Even at a modest IGBT switching frequency (10kHz now for big ones), because the amplitude is about 300V peak, there is still rather a lot at, say, 150kHz (ooh, around 20V or so). fails EMI, ever so slightly.

Cheers Terry

When the peak output of the transformer is less than two diode drops above the power supply capacitor voltage, the diodes are not conducting, so there is nothing other than leakage and capacitances forcing the winding to any particular voltage relative to the "Ground" from the bridge rectifier (-) terminal.

For example, at the instant the voltage on the winding is zero volts, either end of the winding could be anywhere between -0.6V and + 15.6V with no appreciable current flowing. Within that range, it's floating, no?

Best regards, Spehro Pefhany

I'm guessing that there is no such thing as "correct", since the waveform will vary wildly from case to case. And that if it looks like a reasonable, symmetric waveform, it's probably wrong.

Can you post the waveform to alt.binaries.schematics.electronic? If not, please email a pic to jjlarkin ampersand highlandtechnology period com.

Thanks,

John

Hello again,

after I have read Terry's posting, I have to admit that I forgot the stray inductance in my example. Adding 100uH for the transformer secondary results now in additional short spikes with about 100kHz repetition rate. Their amplitude depend on the stray inductance, stray capacitance, load current and the recovery time of the diodes.

John, thanks for asking this question. It's more than just a simple rectifier.

Best regards, Helmut

"Helmut Sennewald" schrieb im Newsbeitrag news:e5cvr8$heo$03$ snipped-for-privacy@news.t-online.com...

There are a bazillion treatments of this circuit in texts and on the web, but I haven't seen one that mentions what the transformer terminal waveforms might be. I guess that's because it usually doesn't matter. This all got started when someone wanted to derive a timekeeping-quality 50-Hz clock from the secondary, and then it gets interesting.

I'm thinking of the side effects of...

Primary-to-secondary capacitive coupling

Diode reverse leakage asymmetry

Diode reverse recovered charge asymmetry, whether they snap or not.

all of which poke charge or current into the winding+diode capacitances and none of which will appear on a basic Spice model.

Leakage inductance could matter, too, especially if the diodes snap.

This is more interesting than fixing toilets, which I'm supposed to be doing today. Why is it that, when you buy a "universal" toilet rebuild kit, you wind up cutting/drilling/gluing/redesigning things to make them fit?

John

"John Fields" schrieb im Newsbeitrag news: snipped-for-privacy@4ax.com...

Hello John,

I have tried to get similar waveforms in my simualation. After I increased the stray inductance to 500uH, it looks very similar. The remaining difference are bursts with about 100kHz repetiton rate when the diodes switch off in every cycle. Maybe your scope doesn't show that because of the choosen sampling rate. It would require a sampling rate of 1MHz or faster. The spikes depend on the diodes too.

Best regards, Helmut

So who's using a "basic Spice model" ?:-)

That's how I got a 3.2V step.

The old house syndrome ?:-)

I just finished installing a replacement fluorescent fixture in the laundry. Original belched the guts of a ballast... messy :-(

And the sockets had cracks, so I just changed-out the whole damn thing.

Nasty getting two 8' ladders in there so we could hoist it up to the

...Jim Thompson