# Identifying unusual three terminal component

I was desoldering some components from a board and found this three terminal device I have no idea what it is. The labels on the PCB were F1, F2,... which is usually reserved for fuses, but I've never seen a fuse that looked like that. My guess is that F stands for filter or something.

That 223 is literally the only label on the package, and it's a small rectangular box about 0.8 x 0.8 x 0.3cm.

• If you include a picture of the pcb it came from (surrounding parts), or a schematic, it might be easier to know what it is... – MartinF Mar 11 at 12:45
• Have you checked continuity of the pins? – HandyHowie Mar 11 at 12:49
• I'd guess it's F for filter, and it's a T type LC EMI filter. media.digikey.com/pdf/Data%20Sheets/… is the only reference I can find for a leaded version of similar size. – Phil G Mar 11 at 13:17

## 1 Answer

(joke) 'tis black art stealth technology, and was top secret until spies traded all secrets during star wars era in spite of appearance. 223 is a secret code with underscore

Now commonly known as a Pi Filter or CLC filter which is bidirectional and 22nF for EMI ingress+egress reduction above 0.5 MHz.

--|÷ö|-->--=

The funny part about rusky-yank spy wars was Japanese always banned exports of new technology, as they developed these new components which were not for sale like all other new miniature Japanese technology, for at least 5 years to have an edge on product technology. This was by economic design, as it was gov't funded.

Here is an early 70's technology graph paper I used to solve for 1 unknown making assumptions for other knowns. ( We still used slide rules in EE schools in early 70's and new calculators were not allowed in exams.) I got this graph paper in my 1st job in Aerospace R&D in '75

• That seems to be it! So the 223 is the sum of the internal capacitances, right? How about the size of the inductor? – Gabriel Golfetti Mar 11 at 15:08
• No it is the value of each C, L is tuned for impedance matching in the 100~500kHz breakpoint region. Compute Zc in this range then estimate L – Tony Stewart Sunnyskyguy EE75 Mar 11 at 15:10
• e.g. 50 Ohm 50uH – Tony Stewart Sunnyskyguy EE75 Mar 11 at 15:12
• RLCQf 5 variables, 1 known, 1 assumption ( 50 Ohm line ) measure DCR of coil in part to estimate Q if using a voltage source. with no load ( Q series gain at breakpoint) i.e. LPF with peaking at resonance. e.g. if DC is 5 Ohms the Q ~ 10 so it "looks" likes a BODE plot of RED line but with gain ~10 so series R added reduces Q like 100 Ohm at each end, But L is still a guestimate based on part size and expected f range. Questions? – Tony Stewart Sunnyskyguy EE75 Mar 11 at 15:38