pot core

On Saturday, May 7, 2016 at 8:01:48 PM UTC-7, John Larkin wrote: ...

...

Bifilar winding may reduce the ringing on the drain of the device that is turning off by allowing the energy in the leakage inductance to couple better to the device that has just turned on.

That may not be a significant issue though.

kevin

SMH-1811-LA to be precise. It isn't available any more. The contract manufacturer knew it was EOL and could have bought a lifetime supply, but didn't. A CM has no incentive to stock parts that it may not ever use; the principal licensee can always jump to a cheaper CM.

I'm thinking I could make a PC board, maybe just FR4 with no traces, with knurled pins pushed in and the core glued or bolted in the center. Epoxy is slow and messy; countersink the botttom and use a flathead screw maybe.

That would be easy to wind: twirl the wire around a pin, make N turns on the bobbin, end at another pin. Center taps would work fine. Then use a hot iron to solder the thermal-strip wire to the pins. The result is a surface-mountable assembly that drops into the existing PCB.

You might check the original winding connections and phase carefully, per the schematic pin-out. It looks like there might have been some fine-tuning on the 18V output, using 1/2 turns.

RL

The guy who designed the transformer tried using some half-turns, but the losses went way up for some reason.

So I have heard. There is apparently a way of winding fractional turn transformers but thankfully I have never needed it, see:

piglet

That's cool. I think that T just started the N+1/2 turn winding through one of the pot core entries and ended at the opposite one. That created a poorly linked half turn inside the core and another weird half turn outside. Whatever it was, it didn't work.

The half-turn problem happens when the gap is in the centre leg. A full turn produces a purely solenoidal field (i.e. one that goes up through the central core and back down via the sides), and so feels the effect of the gap.

A half-turn also produces this field component, but in addition produces a field that goes all the way round in the outer part of the core, _with_no_gap_. That can change the saturation behaviour completely

Cheers

Phil Hobbs

We once tried a 4-1/2 turn inductor, it fried, 4 was fine, 5 was fine. Too many years ago, I think it was a gapped core and the wire near the gap overheated.

Mikek

Improperly applied half turns usually only produce excessive leakage inductance and degraded output regulation in the affected circuit alone - no extra internal xformer loss, as such.

The excessive leakage can produce increased noise coupling to outside loops and affect circulating leakage energy in the source, even in a pot core. Properly applied, the leakage effects are generally reduced.

You might have to correct the schematic, or just the transformer puppet, before throwing it into a cad board revision, if there is no intention of using 1/2 turns on the 18V.

RL

OK, you said it a LOT better!

Jon

On Sun, 08 May 2016 10:17:49 -0700, John Larkin Gave us:

Or don't countersink it and use a flathead screw that sits low in the hole, but does not fall through. You get a nice clean hard mating ring with the screw as well. Bottom side profile would be no higher than some of the components there.

Just put PTH holes in the board (yes a new layout silly)and use a readily available PTH version. Sheesh.

Also, if you had any dexterity, you could simply turn the pins under and make a "J-lead" package out of the PTH core available. Snip the remaining lead length and voila! Back to SMD version.

On Sun, 8 May 2016 16:15:26 -0700 (PDT), snipped-for-privacy@ieee.org Gave us:

I was referring to thread years ago that JOHN was in, you illiterate putz.

On Monday, May 9, 2016 at 9:49:41 AM UTC+10, DecadentLinuxUserNumeroUno wro te:

Your choice of words didn't suggest that. I wouldn't call you illiterate - it's a silly sort claim to make in writing, when you obviously expect the " illiterate" target to read the insult - but I would point out that you didn 't entirely succeeded in making clear what you now claim that you meant.

DecadentLinuxUserNumeroUno wrote in news:92kvib5n5q7efl0qcp20m4q3u65kpq3328 4ax.com:

vib5n5q7efl0qcp20m4q3u65kpq3328 4ax.com:

eptember.org!aioe.org!.POSTED!not-for-mail

jd, you are being a jerk.

-september.org!aioe.org!.POSTED!not-for-mail

Too true. DecadentLinuxUserNumeroUno isn't actually always wrong and he doe s make occasional constructive contributions, which is a performance that J ohn Doe doesn't seem to be able to match.

"prefect"??

On Mon, 9 May 2016 21:10:51 -0000 (UTC), John Doe Gave us:

Stalk much, retarded putz?

On Mon, 9 May 2016 21:27:10 -0700 (PDT), snipped-for-privacy@ieee.org Gave us:

Frank Capra rolled over in his grave and said that he wants him put into prison for using his character name.