Favourite crappy toroids?

It might be easier to use some easily available inductor and just put the resistor in series. 1 ohm would drop the Q to about 2.

John

I use these Murata toroids 100uH ~100m ohm, for Smps filter, Mouser number

George H.

I'd go with John's advice and add a resistor. You don't need this inductance for Q, you need it for the Z it offers and it'll keep that Z. Simple voltage division at AC.

It's more important that the toroid won't leak (hence the compensationg turn I suggested) or, to make life real easy, just shield the whole chebang. Plated Chinese steel would suffice.

But watch loop stability at the lightest possible load. That's when things typically go haywire with LC post-filters like this.

The clamp-on filter type chokes (for common mode on power lines) are made with a lossy ferrite; could you use a lossy material for the filter toroid?

Or (I"m not sure what the situation looks like) could you use a toroidal transformer and put the damper resistor on the secondary (so at DC it looks like a short, but the AC gets damped)? Extreme case, maybe just solder a shorted ring around the toroid (like one sees in shaded-pole motors).

Fair-Rite has lots of 'lossy ferrite' materials for filtering apps, and available in toroid form factors. You just try to get any Q out of them...What is the topology?

Thanks.

It's a symmetric pi network at the moment, relying more or less on the supply output impedance to add another RC pole. It's for an application in the guts of a semiconductor manufacturing tool, and IME the tool core of a fab can be a fairly horrific EMI environment--worse than a welding shop, not as bad as a radar installation. They may be better nowadays, but I sort of doubt it.

Plus I'm a belt-and-suspenders guy all the way when it comes to front ends. (That's arguably a character flaw, but it does help prevent lame failures.)

I don't think a lossy toroid core will help, because the problem is down below a kilohertz, where even crappy ferrite still looks pretty good.

Cheers

Phil Hobbs

Lossy magnetic material is equivalent to resistance in parallel to inductance. You are looking for resistance in series, isn't it?

VLV

Yes, ideally. I want the 80ish dB rolloff from 1 to 100 kHz, whereas a lossy ferrite doesn't start looking resistive much below a few megahertz, AFAICR.

If I'm going to have to add a 1-ohm resistor, I might as well just use a three-section RC with the same total resistance, say 0.33 ohms and 180 uH AlPoly, three times in a row. That's the other way of getting rid of the inductive pickup, of course.

Mumble frotz.

Phil Hobbs

That's really not bad. The poly caps have absurdly low ESRs. Toss in some 10u ceramics to be really secure.

John

The two are equivalent.

For any inductance in series with a resistance, there exists a different inductance, in parallel with a different resistance which produces the same complex impedance.

That is, (Xs + Rs) == (Xp || Rp)

Agreed, though... at low frequencies, even a poor ferrite won't be all that lossy (i.e. low Rs, or high Rp) and won't spoil the Q by all that much.

No.

It is not possible to satisfy this condition at two different frequencies at the same time.

VLV

True... but the OP was asking about adding losses in order to spoil the Q at a specific resonance frequency (between the coil inductance, and the filter capacitance). At that one frequency it wouldn't matter whether the Q-spoiling resistance was in series with the coil, or in parallel... you could achieve the same total complex impedance either way.

Whether resistance-in-series or resistance-in-parallel would be better at higher frequencies would depend on the details of the rest of the circuit, I imagine.

Wind some/all of the turns with nichrome wire.

tm

I think Vladimir's point was that with a lossy ferrite, you have asymptotically a 1-pole rolloff rather than 2 poles, which isn't as good for low-frequency supply decoupling. For high frequency stuff, it's often the bee's knees.

Cheers

Phil Hobbs

ld

Phil did you look at the murata toroids?

The 470uH is ~300 m ohms within a factor of two of what you wanted.

George H.

The third section never buys you much.

George H.

Well, it gets me another 30 dB at 100 kHz--90 is a lot comfier than 60.

Cheers

Phil Hobbs

Nichrome?

Just adding resistance to the wire is the same as adding it externally, and much more of a pain--you can't readily solder to nichrome, for a start.

I was hoping for smaller wire diameter and a smaller core, but it looks like that doesn't exist, so I'll probably fall back 10 yards and punt, i.e. use a 3-section RC filter.

Cheers

Phil Hobbs