Shielded inductors.

Simple question. I'm looking for a 100uH inductor in either a toroid or a pot core. (I'd like it shielded from outside interference.) Digikey lists no pot core inductors. Are the shielded ones like this from wurth,

formatting link
the same (or similar to) pot cores.

Thanks

Reply to
George Herold
Loading thread data ...

Not really. The ring around the central disc is epoxy, so the shielding is nothing like as good as a pot core or a toroid.

I'm a fan of the Coilcraft toroids, but they're on the pricey side.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

So a bit like a pot core, but the gap is at the top and bottom, and not hidden (magnetically) inside the core. And some B-field can leak in or out at the edges.

George H.

Reply to
George Herold

Digikey has toroids:

formatting link
formatting link

Pot cores are geneally custom parts, got to roll your own. As Phil said the "shielded" upright squares have a gap. Sometimes hard to see when they color the epoxy to match the core. The leakage is easily verified: Run a switcher with them, connect an LED to a same kind of core and gradually lower it to the core in the circuit while under power. Pretty soon it'll light up.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

My experience is that this design IS very sell shielded; I've been unable to pick up the operating frequency of a DC/DC converter using a search coil. I suspect the field leakage around the 'epoxy' gap might be very small indeed, and it'd be in a space away from the printed wiring and adjacent components. Like pot cores, the dimensions of these magnetic parts are made by accurate grinding, the gap might be epoxy filled, but it's small.

Reply to
whit3rd

Thanks Joerg, Yeah I ordered some toroids. I've got an inductor that's used to block RF in a current source. (RF into a diode but down stream of the source.) The Dang thing is a open ended coil and it's picking up interference from everywhere.

George H.

Reply to
George Herold

I had a similar situation lately.

formatting link

A big source of noise in any current source is Johnson noise in the sense resistor, so I figured that I could put an inductor in series with the resistor, since inductors don't have Johnson noise. Spice sometimes doesn't like to sim this, but when it does, the current noise drops off radically at higher frequencies (when L0 obviously.) I'd need a low-pickup inductor, probably a toroid, to get away with this.

We stock some nice little 2:2:1:1 ISDN transformers, cheap, toroids, that would get me about 200 mH here. They are handy for all sorts of things.

--
John Larkin         Highland Technology, Inc 
picosecond timing   laser drivers and controllers 
 Click to see the full signature
Reply to
John Larkin

Can you share mfg and part number? If it's a part that's likely going to be around a while. I got burned with modem transformers. They are very nice for high inductance shielded cores because otherwise hundreds of mH are only available in the old open "sewing kit" style. Those are like barn doors and one must place shields.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

Well true confessions, I copied the current source from Jan Hall (RSI 64(8) Aug, 1993 pg 2133.) Here,

formatting link
(might as well use dropbox.)

Here the inductor is after the current source. I'm not even sure what it's doing. (My interference went down when I shorted it.) I'm hoping the toroid will fix things.

George H.

Reply to
George Herold

Oh I see now... duh, Jan was feed the RF in right after the 100uH inductor. in my implementation I feed it in down stream.. right into the lasre diode.

George H.

Reply to
George Herold

I have lit an LED soldered to another inductor of same type and then held adjacent.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

Probably fixing the gross drain capacitance of the PFET. Otherwise you effectively get full shot noise at high frequency, due to the diode's noise.

I'm a fan of using transformers in series with the diode. If you use two secondary windings in series opposing (i.e. like a centre tap but with one winding flipped) you can put a big bypass cap at the tap to get all the AC, but the DC magnetization cancels out. The input end is left flapping in the breeze, but it won't flap much because the diode impedance is low.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

Yes. However the gap is along the top (as you can see), and maybe the bottom. So it's not nearly as good. A toroid isn't exactly blameless, so may be about equivalent.

I suppose your tradeoff is miniaturization of this part, versus more spacing between parts for shielding, versus adding shielding material (metal plates or cans, or ferrite plates -- which are available adhesive-backed if you were wondering).

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
 Click to see the full signature
Reply to
Tim Williams

Putting the inductor after the current source is a silly idea. The output impedance of the current source can be in the hundreds of k. The 100uH inductor will be a small fraction of that, so it's not doing anything except picking up noise as you have mentioned. Similarly, the 100uH inductor in series with the 1N5711 at the other end of the coax cable is not helping to reduce the noise much. I don't know the reason for the 32 ohm resistor shunting the inductor.

The 50 ohm resistor in the VP0106 source lead generates 0.9 nV/Sqrt(Hz), about the same as the LT1028. As JL points out, putting the inductor in series with the resistor would make it much more effective at reducing the resistor noise at higher frequencies, subject to stray pickup.

I don't know what the diodes in series with the +/- 15V supply to the LT1028 are supposed to be doing. The 10k in series with the gate of the VP0106 could be much smaller. No sense generating lots of noise when you don't need to.

The VP0106 is a p-channel mosfet. These are not known for low noise. Perhaps a low noise jfet might be a better choice.

The resistor network at the LM399 is going to generate gobs of noise, and the RC filter is not going to help at low frequencies. The resistors need to be much smaller.

The LM399 is spec'd at 20uV RMS noise and is a really old design. There are much better references available.

There are some really bizarre things in this circuit. I'd spend more time on it but I have to finish doing laundry.

I'd take a look at the bias supplies for the main and vernier coils in a YIG oscillator for ideas on low noise current sources.

Reply to
Tom Swift

Joerg, I'm not sure that's "fair". You are sticking two magnetic materials together. What happens if you stack two pot cores, one on top of the other?

I should order a few and look see at the pick-up... You know, one thing really bothers me. How much worse can the interference be at some customer's place?

George H.

Reply to
George Herold

I don't think I care about shot noise. I think,(I'll have to measure), 1mA of current changes the wavelength by about 1GHz (DC), (that's at least an order of magnitude) A wavelength change of 1MHz, (which is all I need.. can see) is then 10uA (p-p) (?.. there's some factor going from DC to noise.) into something like a 1 MHz BW... 1nA /rtHz. (Hmm that number is too low, I'm missing an order of magnitude.) With a maximum dc current of 100mA. Shot noise of 100mA is.. (i^2 = 2*e*I-dc*BW) ~0.2nA/rtHz

George H. (those dang orders of magnitude... I'm gonna join a different order. :^)

Reply to
George Herold

Grin... Tom this is great. Let me just say I copied this ~2001-2, and it was ~10 year old then. It worked and I haven't paid much attention since. I've recently had my nose rubbed in my neglect, (See HV opamp thread... same instrument, different circuit.)

The first 100uH is to keep the RF modulation.. right below it, out of the current source.

Similarly, the 100uH

I use a 47 uH inductor right near the laser diode as part of a bias Tee, with ~50 ohms in series with the RF input.

I used an opa277 and not the LT1028, so even more noise. Fortunately, all these are below what I care about.

Thanks.. you can see my scribble's as I worked out the low frequency modulation via the opamps. It's weird, but it doesn't change the DC current, (which is what I care about), as the modulation is added.

Thanks, at the moment I've got more of a grounding and shielding, issue. (At least that's my hope.)

George H.

Reply to
George Herold

Are they non-progressively wound? The simple-minded approach to winding a t oroid puts a current loop in the plane of the toroid, which will pick up an external field. There are a variety of "non-progressive", "astatic" windin g schemes but they don't seem to be well known.

I'd wind my own 100uH inductor on a - gapped - pot core. They tend to come in around 1uH turn, so 100uH is only ten turns, a single layer winding.

The gap is in the centre, so the bulk of the ferrite shields the coil prett y well.

--
Bill Sloman, Sydney
Reply to
Bill Sloman

You can put a loop in the board underneath, which helps. The ones Joerg posted reverse the winding halfway, which helps, but there's still the resulting vertical turn, owing to the wire crossing the middle. Doing two layers with inverted helicity wired in parallel might be better. (You obviously have to do that so the field in the core doesn't cancel out.)

There's leakage at the seam, though, and the centre pillar tends to be vulnerable to saturation.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

TALEMA SMJ-140B

They make them in other turns ratios, too.

--
John Larkin         Highland Technology, Inc 
picosecond timing   laser drivers and controllers 
 Click to see the full signature
Reply to
John Larkin

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.