coupled-inductor vs common-mode choke

K = 0.5 maybe?

Do you have one to measure?

Normally they are never bifilar would since UL/IEC demand 2mm creepage phase to neutral

You can test this to circumvent the requirement, but that takes a lot more engineering effort

Advantage of separation is lower capacitance which results in higher self resonance frequency

Cheers

Klaus

Yes, for AC-line units, but this one has a very low 80V insulation rating, evidentially meant for dc-dc converter input stage isolation, etc.

The low inductance seems to indicate it would be used for signal or low power inputs

Or high-current dc-dc converters, 5uH 4A max. OK, not that much current. This doesn't look like an exceptionally attractive 6x7mm part. But maybe we should examine the competition, before judging. It is quite small. I have a possible application. Maybe I can squeeze one onto my MPX-16H board, in place of L4.

I LOVE how tight you make your schematics, too ! I try to do that too rather than spread them out over many pages.

I am guessing that DG-409 chips aren't available in SMT ?

I have been taken to task for it, folks reminding me of other larger paper sizes.

No, the DG409 is, as are most alternates. But I intentionally chose DIP packages because these parts are exposed to long wires off board, with static-damage possibilities, etc. So the user should be able to quickly self-repair. This has long been my philosophy for off-board connections.

I remember the pain and suffering when a National Instruments board failed due to one chip's static damage, we had to rewire the signal to another pin and change the software, what a pain!

BTW, I'm giving away free blank MPX-16H boards. Next pass I'll have a batch machine assembled, but they'll cost about $150 each.

The 9uH unit has SRF of 77MHz, but they specify its impedance at 100MHz. What's the point?

Best regards, Piotr

Resonances are symmetrical when plotted on a log frequency scale. A choke has the same reactance at 2*SRF as it does at 0.5*SRF.

Same with bypass caps.

Cheers

Phil Hobbs

OK, yes I definitely understand the resoning for through hole on those then. Hopefully they are socketed.

We have used multiplexors like the CD4051 or 74HC4051 family and when their pins are exposed directly to the outside world, would seem to break quite a bit in the field. This is why I always add some series resistance or buffering and at least small capacitor to ground on the pins to reduce that possibility. That has worked for me for about 20 years now where before that, in other products, they would break quite often.

Yep. PITA !

Assuming there's only a single resonance, of course!

Which is an okay bet with any capacitor I've seen. Not as good of one with inductors.

Typical CMCs will have a few peaks and valleys up in the 10s of MHz range, where the windings resonate in helical waveguide modes (slowed by the velocity factor of the core, and damped by its loss), at least when single layer; multilayer windings of course are easier to model as inter-layer capacitance.

Ferrite beads are usually simple, though they sometimes have anomalous humps at medium frequencies, that you wouldn't expect from the shape of the main peak.

Data line chokes are usually simple in common mode, but have peaks and valleys in the differential mode; which is to say, they have reasonable transmission line behavior up there, which is actually a good thing in this case.

Tim