Let's discuss using small indcutors for power filtering

[midwayfair]

instead of small resistors.

Jubal81 has been using 100uH inductors where many circuits would have a tiny (10-100R) resistor for power filtering, so I started wondering about using them for that and didn't really find a whole lot.

Tayda has them for 7c (http://www.taydaelectronics.com/inductors/100uh-inductor.html), or you can get similar but better ones from Mouser for about 11c each if you get 100 (http://www.mouser.com/ProductDetail/JW-Miller/78F101J-RC/?qs=sGAEpiMZZMsg%252by3WlYCkU0cgSdNMaEC7RfDdgBSHwPU%3d --- only 4.1R resistance!). They're the size of resistors, so basically a drop-in replacement.

What are the advantages and drawbacks besides cost and tolerance of parts? To me this seems like all-upside except for a dime, but they're rarely used in anything pedal-related.

The main issues with inductors -- that they're big for audio signal levels and inconsistent for filters -- wouldn't seem to matter when using it as a power filter.

[Eddododo]

Interesting


That's all i have to contribute

[R.G.]

What are the advantages and drawbacks besides cost and tolerance of parts? To me this seems like all-upside except for a dime, but they're rarely used in anything pedal-related.

It's a matter of impedance and the frequency that you're trying to filter.

100uH = 2*pi*F*H = 6.28*100uH*F = 628 micro-ohms per Hz.

At 60Hz, the impedance is about 61 milliohms, and the 4.1R resistance dominates things dramatically. In fact, it doesn't really look inductive until F = 4.1ohm/(2*pi*100uH) = 6528Hz. At 12kHz, it's about 8 ohms, and 15 ohms at 24kHz. It doesn't really impede anything until into the RF range, at least in comparison to similar sized resistors.

Now it's a great thing in, say AM and FM radio stuff. At 1Mhz, it's 628 ohms, and 62k at 100MHz.

So there are some practical issues.

[SISKO]

100uH = 2*pi*F*H = 6.28*100uH*F = 628 micro-ohms per Hz.

Ive never look impedance that way. That is some clever unities movement.
Always learning something

[PRR]

> didn't really find a whole lot

Then you did not read and digest the inductor essay I posted 6 hours after your post??

Small inductors have dang-near NO impedance at audio frequency.

The 36-type 100uH coil you point to will be a 2 Ohm resistor for all audio purposes.

At 100Hz (which is where hum/buzz happens) the 100uH acts like another 0.06 Ohms of impedance. On top of the coil's 1.8 Ohm resistance, this is still 2 Ohms for all purpose.

At 2.8KHz the 100uH acts like another 1.8 Ohms inductive reactance, for a total impedance of 2.54 Ohms. Which, if you have a problem, 2-1/2 Ohms is rarely a a solution.

Back-track to a spec. Say a pedal is 9mA at 9V so acts-like 1K Ohms. Say you use a cheap half-wave supply and are plauged with 50/60Hz hum coming in. You want a choke which is much more than 1K impedance at 50Hz. Reactance chart gives 3 Henries. 3H is 30,000 times larger than the pointed-at 100uh, Tayda does go up to 4,700uH (4.7mH) for a whopping 15 cents. 4.7mH at 50Hz is 1.5 Ohms. This is "no effect" in 1K circuits. Anyway the 46-type 4.7mH coil has 30 Ohms of DC resistance. May as well just use a 31.5 Ohm resistor, it is cheaper and tougher.

Power and audio coils are BIG. Usually around a big lump of iron, though tweet-speakers can manage with a very big air-spool of copper.

Coils *do* have much value when fighting supersonics. But you should compute the inductive reactance and compare it to basic circuit impedances before you click "Buy". (R.G.'s expression Ohms per Hz is a nifty way to think.)

[PRR]

> At 1Mhz, it's 628 ohms, and 62k at 100MHz.

Before the radio boys cream their jeans--

Coils have Self Resonant Frequency. That datasheet has way too many types (sizes) and values to keep straight, but if I read it near-correctly, the 100uH 36-type has SRF around 3MHz. So at 1MHz it would still be near 620 Ohms, but at 100MHz it would be a *capacitor* (around 30pFd, and Q not even 5).

[mykaitch]

No argument from me BUT did have a prob with harmonics that a pi filter with two 300mH coils fixed.
Yeah, use resistiors and lots of decoupling caps,

[bool]

A 10-ish Ohm resistor in series with a 100 or 220uF inductor (filtered with a cap-bank) will make the use of switching psu's  (usually) a non-issue. This is a RLC filter with multiple center F's so use of a large enough R in series with the L is a must to bring the Q down.

You willl have to design a very good ground return path and install a good capacitor bank (MLCCs in parallel with lytics). This isn't going to be "expensive" in terms of cost, however take care with the layout.

[mykaitch]

Yes, one thing that caught me was that the filter had to be as close as possible (electronically) to the the noise source.
The actual prob was that I built a mixing desk and on inspiration fitted a Harman Kardon iDrive so that I could plug
an iPod into one of the channels. I then discovered the most awful square wave on the PSU which I think is the CPU clock of the
HK interface. It was that I needed to get rid off and it did take a while. I measured it and calculated the inductance etc and it was
rubbish. Checked ot with an on-line tool and my maths ok. Eventually I ended up with the pi. My mixer PSU is linear. I may play around
with some damping Rs but the amount you can see on the scope is only audible with no music and 1Kw amp cranked right up -- but it still bugs me that it's there at all.

[midwayfair]

What are the advantages and drawbacks besides cost and tolerance of parts? To me this seems like all-upside except for a dime, but they're rarely used in anything pedal-related.

It's a matter of impedance and the frequency that you're trying to filter.

100uH = 2*pi*F*H = 6.28*100uH*F = 628 micro-ohms per Hz.

At 60Hz, the impedance is about 61 milliohms, and the 4.1R resistance dominates things dramatically. In fact, it doesn't really look inductive until F = 4.1ohm/(2*pi*100uH) = 6528Hz. At 12kHz, it's about 8 ohms, and 15 ohms at 24kHz. It doesn't really impede anything until into the RF range, at least in comparison to similar sized resistors.

Now it's a great thing in, say AM and FM radio stuff. At 1Mhz, it's 628 ohms, and 62k at 100MHz.

So there are some practical issues.

Okay, so basically, when I plugged things into the filter calculator, I had it backwards. It thought that the only frequencies coming through were very very high frequencies above the audio range. Well, I guess this kills the idea.

Then you did not read and digest the inductor essay I posted 6 hours after your post??

As I've said many times, the time machine pedal is still in development. There are some bugs to work out. It seems to go forward in time just fine, but going backwards isn't working properly.

[samhay]

^As I've said many times, the time machine pedal is still in development. There are some bugs to work out. It seems to go forward in time just fine, but going backwards isn't working properly.

you just have to reverse the polarity no?

[midwayfair]

^As I've said many times, the time machine pedal is still in development. There are some bugs to work out. It seems to go forward in time just fine, but going backwards isn't working properly.

you just have to reverse the polarity no?

Yes, but I'm not sure which axis it's on.

[R.G.]

Coils have Self Resonant Frequency. That datasheet has way too many types (sizes) and values to keep straight, but if I read it near-correctly, the 100uH 36-type has SRF around 3MHz. So at 1MHz it would still be near 620 Ohms, but at 100MHz it would be a *capacitor* (around 30pFd, and Q not even 5).

Aw, Paul, you spoiled it! I was just waiting...   

[bool]

... yet another reason to precede the choke with a R and to close the cap bank to supply ground return loop really tight.