Electronics > Projects, Designs, and Technical Stuff

Battling Capacitance?

(1/2) > >>

ELS122:
so I have a guitar pickup and because of its construction, it forms a capacitor between the coil turns. is there any way to battle this capacitance without rewinding the coil differently?

DaJMasta:
The amount of capacitance that creates should be quite small... and that sort of level of capacitance is not likely to interfere with anything in the audio band (capacitance of the interconnect cables would be just as much, for example).


That said, it's due to the the construction of the coil itself, so rewinding it in one form or another is going to be required.  Capacitance is dependent on spacing between the coils, so if you use a magnet wire with a thicker insulation, you will take up more space on the bobbin with the same number of turns, but you will have lower capacitance.  Theoretically, changing the dielectric (the insulation) would also change it, but you'd have to choose carefully and I don't know if there's much availability of magnet wire with insulation optimized for low parasitic capacitance.


And, then there is actually some documented stuff on it, take Peavy's "Chapter 5 The Ancient Order of the Mystic Magnetic Pickup" whitepaper available online.

T3sl4co1l:
Nope, fundamental property of the universe. Sorry. :-DD

There are strategies to avoid it.  Namely, you need, I suppose around 20kHz of bandwidth, more or less, so the pickup needs to be designed to have Fo = 1 / (2*pi*R*C) above that limit, where R is the load resistance (or source impedance, defined in several ways) and C is the equivalent parallel capacitance.  Including cable capacitance, incidentally.

There's also no fundamental reason to need humongous turns counts on pickups, other than guitar amps being designed for very high impedances, only for historical reasons as far as I know.  A pickup could be made with fewer turns of somewhat larger wire, and receive exactly the same signal power.

The internal resistance would then be lower, the wire length (and therefore capacitance) lower, and the impedance closer to that of the cable (probably in the vicinity of 50 ohms, even if not proper RF transmission line).  Bandwidth could be extended to 100s of kHz, not that you'd need it!  (Or even want it, necessarily; RFI is easily filtered out at the input, fortunately.)

The difference is, you'd probably need a different amp, or a preamp.  Which could be installed in the guitar itself and battery powered, or you just always play with a pedal that boosts the lower voltage to what most amps use.

Or if you're not talking in regards to coil design (I know a few players wind their own) but just what you have, then again the answer is as above -- reduce the load resistance.  Limited by the coil design, you won't improve bandwidth that much, and pretty quickly you'll get into diminishing returns as the pickup's inductance and internal resistance dominate, and the noise floor rises.  There exists an optimum load, where bandwidth is best without also compromising noise and gain too badly; you'll have to find what that is in the documentation (if they give it at all?), or by experimentation.

Tim

ELS122:
in guitar pickups, this is noticable problem. and If you measure the most cheap pickups they measure about 10nF which will filter out some of the highs.

I thought that maybe somehow using some clever circuit you could trick the universe into thinking that capacitance doesn't exist but it doesn't seems like it's doable

also the frequency range of a guitar is about 60-16k hz since you can play pinch harmonics that are multiple times higher than notes you can play normally

T3sl4co1l:
You can use opamps to synthesize a negative capacitance, but this is merely equivalent to a differentiator, an amplifier with excess gain at high frequencies -- no free lunch, you'll only increase the noise level up there.  But probably also make an oscillator, because inverted impedances are tricky to use.

Alternately, you can tune out the capacitance to a certain extent (limited by the higher frequency details of the winding), but this only works for narrow bandwidth; for example, a 100-10kHz pickup might be tunable to 5k-15kHz or 10k-20kHz.  Pretty useless for audio purposes.

Although given what you played with in your other thread, https://www.eevblog.com/forum/projects/overkill-pikcup-combination/ maybe a multiband pickup wouldn't be too ludicrously out-there after all. ;D

You'd tune it by simply putting an inductor in parallel, and adjusting the termination resistance as needed -- it'll want to be lower, like if the 10kHz case above is terminated with 100kΩ, tuned to 10k-20k it might need 50kΩ.

The response will simply look like a pickup EQ'd to a bandpass of whatever range it has.  Apply appropriate EQ and mixing, and you've got the original signal back.  Give or take differences in phase shift at the band edges, which can cause notches at the crossover points; a full stack could use phase shifters (maybe a 90 degree phaser pedal with fixed adjustment, rather than sweeping/phasing wah-wah), on each source, tweaked until the notches go away.

Hmm, shouldn't be too hard to compare frequency responses and make sure your pickups are flat (or within whatever response is desirable), you've got a sweep generator right there, just perform a slide. :-DD

The reductio ad absurdum could use all tuned pickups, with extra C to reduce the HF band limit, e.g. so you get a 60-300Hz band, then a 300-1kHz band, and... and mix them all together.  Mind, it wouldn't do anything special -- it would just be amusing as an electromagnetic model of the human ear, of sorts.  You could further subdivide the pickup channels into, say, single notes at a time -- and further, detect the amplitude of each channel, rather than its signal directly.  Now you get a vocoder response!  (But hey... perfect tuning every time? ;D )

Tim

Navigation

[0] Message Index

[#] Next page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod