|
Post by wolf on Jun 24, 2009 23:06:38 GMT -5
|
|
|
Post by ashcatlt on Jun 24, 2009 23:54:23 GMT -5
We could probably spice that up. I'm supposed to be building drum tracks right now though. Just wanted to mention that all the time I was reading Chesh's rant in that first link I kept thinking about the Pultec Passive EQs, and all the money Tape Ops and Mastering Engineers are paying for then.
|
|
|
Post by wolf on Jun 25, 2009 2:52:11 GMT -5
|
|
|
Post by newey on Jun 25, 2009 5:58:54 GMT -5
I fixed Ash's link, before I paged down to see wolf's reposting of it. Both now work.
Rather than using a 5-position rotary switch, could this be done using a Strat 5-way switch and 3 different cap values, with the 2 and 4 positions giving different values through parallel connection of 2 caps at once?
Of course, this would not give one a progression of values from low to high as the rotary does. They would be "out of sequence", for lack of a better term.
|
|
|
Post by sumgai on Jun 25, 2009 14:56:54 GMT -5
newey, They would be "out of sequence", for lack of a better term. Yes, we do need a better term, because what you've described is still a sequence, just not one that you take to be linear, logical, nor possibly even desirable. A Strat-style blade switch pretty much requires that one manipulate it in a sequentical manner - rather hard to get around that little physical limitation, I'm sure you'll agree. ;D sumgai
|
|
|
Post by wolf on Jun 25, 2009 15:07:51 GMT -5
Well, now that some replies have been made to this topic, I have some questions. What is the frequency that is passed by this circuit? (Basically, what is the formula?)****Edited to add **** From what I've seen on the Internet, some websites say that the resonant frequency formula holds true whether the inductor and capacitor are in parallel or in series. ***************** If the capacitor and inductor were in parallel (a tank circuit), what new values would the caps and inductor be?
|
|
|
Post by sumgai on Jun 25, 2009 17:49:25 GMT -5
wolf, There won't be any "new" values, because capacitors and inductors will respond to frequencies in the same way whether connected in either series or parallel, so long as they are not connected to anything else in any way other than at the two ends of the "network", or as you correctly called it, the tank. Call that a two-pole network, and you'll see that if there are any other connections, then it's no longer a two-pole, it's something else. The reason the two scenarios react in the same way becomes obvious when you do the arithmetic. In essence, the inductor passes the lower frequencies, and it doesn't care if they came straight from the source, or if they came from a capacitor that modified that source. Ditto for the capacitor, it also doesn't care about the input, it only passes the higher frequencies, comparatively speaking. All that's not to say there aren't reasons to use one way over the other. A series combination will tend to transfer the maximum voltage from the source to the load, whereas a parallel combination will tend to transfer the maximum current from source to the load. For our purposes in a guitar, the series combo would be a better bet, all other things considered. (Like, can it be mounted without fear of physical degradation.) As for the formulas.... Aw, geez, we've done that so many times here already. OK, once again, but this time with some outside references: The resonant frequency of a two-pole network comprised of one or more inductors and one or more capacitors (a simple LC circuit) is: F = 1 / 2π√LC Read that as "Frequency in Hertz is equal to 1 over 2 times Pi times the square root of inductance in heries times capacitance in farads". Because the HTML symbol for "square root" can't be extended over succeeding characters, let me rephrase the formula: F = 1 / (2π(√(LC))) Along with a lot of other jizz-jazz, you can find all the substantiating documentation in The Wikipedia. HTH sumgai
|
|
|
Post by wolf on Jun 25, 2009 18:19:06 GMT -5
Okay sumgai, thanks for the information. I studied electronics a long time ago and I've forgotten a lot. (I knew about the resonant frequency formula but did not know it applied either for parallel or series circuits). Anyway, going by the formula the resonant frequencies for those capacitors combined with a 1.5 Henry inductor are:
Hz MFD 4,109 0.001 1,838 0.005 919 0.020 658 0.039 411 0.100 (Hope that table displays okay).
Okay, are there any tone options that might be worth getting for capacitor values smaller than .001 or greater than .1 mfd?
|
|
|
Post by sumgai on Jun 25, 2009 20:51:29 GMT -5
wolf,
I can't say about whether there are any other cap values that would yield good-sounding results, that's up to the individual, of course.
However, let me lay some more numbers on you. You already know that A=440Hz, but just for a refresher, that fundamental frequency is found at the 5th fret of the first string. Meaning, that the 17th is only 880Hz, and by extension the 22nd is a mere 1125Hz, give or take a few cents.
All that goes to your table: At 0.001µf, the resonance is well beyond the nearly the 3rd harmonic of the highest note on the fingerboard. (!) That's not very easy to hear, is it? With your published design, the value will only be a 3dB increase at the resonance point, although the Q of the filter says it will have some effect down to half of the peak frequency, thus it might be noticible around the 2nd harmonic.
Sure, under studio conditions you'll hear something happen, but on-stage? I doubt it. Ditto for the 0.005µf cap, that's not much of an effect either.
For those following along at home, the lower the fundamental frequency, the higher the harmonic that will fall into the resonant peak. Not to mention, it's also lower in volume in the first place, compared to the fundamental, hence the less audible will be the effect. That translates into, the lower you play, the less you'll hear the filter's effect, because it's affecting ever-higher harmonics that are also ever-weaker.
Your 0.02µf cap might make an audible difference, it's nearly what most guitars use, that's well down in the guitar's fundamental range. The others also might be interesting.
HTH
sumgai
|
|
|
Post by newey on Jun 25, 2009 21:30:34 GMT -5
Couldn't the resonant frequencies be brought down into more "guitar-Friendly" territory by changing the value of the inductor as well? Whether you'd go higher or lower . . .well, I guess I better go back and look at that formula again.
|
|
|
Post by JohnH on Jun 25, 2009 22:29:33 GMT -5
if no one else gets to it, I can put it in 5Spice at the weekend.
What I think it will show is that the series LC circuit will act as a notch filter, cutting the sound at around the frequenecy where the impedance of the inductor matches that of the cap, giving a dip. By contarst, a parallel LC filter in the same position will cut the signals away from that frequency, leaving a hump in the response.
lets see... John
|
|
|
Post by wolf on Jun 25, 2009 22:37:32 GMT -5
Thanks again sumgai If I've followed your explanation correctly, larger capacitor values will have a more noticeable effect. I did experiment with this circuit a few years ago and found that capacitor values greater then .02 mfd cause a significant loss of volume. But be that as it may (and I doubt if it ever was) here's a table with newer values and capacitors rated in nanofarads:
nfd Hz 1 4,109 5 1,838 20 919 39 658 100 411 200 291 300 237 400 205 500 184
JohnH Yeah, that would be a great idea to give this a P-Spice analysis. Thanks.
|
|
|
Post by ashcatlt on Jun 25, 2009 22:55:04 GMT -5
Keep in mind that this is a notch or band-cut filter, not a band-pass. With the smallest cap, you're rolling off many of the same frequencies that a normal tone control would when turned down. Thing is, there's a big resonant peak just a bit above the resonant frequency you've calculated. This is right around the place you'd find the resonant peak with the tone control all the way up.
It was mentioned in the thread where I posted the 5spice analysis that this sounded a lot like a piezo pickup on an acoustic-electric, and I can see why it might.
As the cap values get larger, the "resonant trough" seems to decrease according to your figures, but the resonant peak (mostly dependent on the other LCR network in the circuit) stays about the same.
With the higher values, I'd expect something like an exaggerated "metal-scoop" type sound. Are you looking for a "stuck wah" type thing?
I've done 5spice analysis here, but I can't get it to step to your specified values. All it lets me do is put in a max and min value and then it steps through in up to 10 equal increments.
|
|
|
Post by ashcatlt on Jun 25, 2009 22:58:57 GMT -5
ninja'd twice in one post! I did experiment with this circuit a few years ago and found that capacitor values greater then .02 mfd cause a significant loss of volume. This makes sense. It's a nasty deep scoop, and as your table shows you really start to dig out the relevant guitar frequencies as you go to higher caps. Since most of the overall energy is in that region, it should sound a lot like a general volume reduction.
|
|
|
Post by JohnH on Jun 26, 2009 6:57:20 GMT -5
I see we almost got there last September, with 5Spice by Ash. Heres some more: It’s quite interesting. This is a 5Spice model of the filter, set into a typical Gibson-type guitar with one humbucker switched on. The normal tone and volume controls are included, both at max, plus a guitar cord and amp. The extra filter comprises R6, L2 and C5. I can’t set the exact cap values that Wolf suggested in one graph, but by running across two ranges in a series of equal increments, it gets close to most of the values. Here it is fully engaged with R6 at zero, and C5 from 0 to 10nF. It’s interesting how the standard treble peak and dip (red curve in 1st graph only) is replaced by one at a higher frequency, as the extra inductance combines with the pup inductance to make effectively an overall lower L value: Here it is from 10nF to 100nF Increasing the series resistance R6 quickly separates the filter from the rest of the circuit, and all the action occurs at the lower resistances up to about 50k. Here are plots with a 10nF cap, and R6 values from 0 to 500k and 0 to 50k: It could be a fun gadget, especially with some boost somewhere to compensate for lost volume. regards John
|
|
|
Post by JohnH on Jun 26, 2009 18:20:42 GMT -5
|
|
|
Post by wolf on Jun 26, 2009 23:35:28 GMT -5
John H That's quite an Excel spreadsheet you created. (Yes, I downloaded the 2.2 version). I never investigated the previous versions you wrote but it sure seems like something for which you have quite a talent.
As you and Ashcatlt had said, making those capacitor values smaller makes the notch value more pronounced but it starts going into the "dog whistle" zone below 10 nanofarads.
After doing all that analysis, I hope this discussion continues for a little while longer.
|
|
|
Post by JohnH on Jun 27, 2009 19:02:23 GMT -5
John HAfter doing all that analysis, I hope this discussion continues for a little while longer. OK, lets keep talkin' Have you tried building such a circuit? One variant of this design is to make the inductor one of the pickups, instead of just a passive coil. I think much the same response shape would occur, but with more volume preserved. Mathematically, I found it interesting to realize (subject to a corrective smack on the head if anyone knows better) that at the 'resonant' frequency (actually, is it resonating?, it seems totally dead at that frequency), the impedance of the cap in series with the inductor is zero, hence (if the resistor is zero) killing off all the output at this specific frequency notch. The Math derivation comes from the 'complex number' representation of the impedances on the vertical J axis, where reactance of a cap is positive and that of an inductor is negative. Caps and inductors in series have their impedances summed, so equal +j and -j = zero (for perfect components). The spreadsheet has all this tabulated lower down, at each frequency. Columns E and F are the combined impedance of the tone LRC network. John
|
|
|
Post by ChrisK on Jun 28, 2009 15:57:22 GMT -5
I still use the LTC SwitcherCAD version of pSpice since I can stipulate the absolute step values. You may be able to use the .step param list declaration in 5spice as well. This is also how I can use absolute audio taper pots at specific knob positions for specific non-linear values. I show this in the The Passive High-Cut Tone post. In the posts in pickup tuning I show some series resistance to limit the effects. The series RLC circuit in parallel with the signal generators does have its resonant frequency when the capacitive and inductive reactance vector sum adds to zero degrees (real). At this point, the impedance is limited by the series resistance in the coil wire and externally. If the inductor is not simulated with some realistic internal resistance, quiet things are. This circuit topology is a notch filter (mid-scoop - the Varitone). A series LCR circuit has a component voltage response at resonance that is limited by the series resistance. Without it, it can become infinite. The resistance limits the excursions as well as broadens the width of the response peak. The series LCR circuit comprised of the guitar pickup, pickup and cable capacitance, and pickup internal resistance results in the voltage gain and humping above the nominal generated AC output of a pickup. This is because the output is tapped off of the capacitance in the structure. A parallel LCR circuit (the R is part of the L) in parallel with a signal generator has its resonant frequency when the capacitive and inductive reactance vector sum adds to zero degrees (real). A parallel LCR circuit has a component current response at resonance that is limited by the series resistance. Without it, it can become infinite. The resistance limits the excursions as well as broadens the width of the response peak. This circuit topology is a bandpass filter (the edges above and below resonance are attenuated). You can use a pickup for the inductor, but if it's producing an output signal, it's not just an inductor. The Fender Powerhouse Strat actually makes use of a pickup with unmagnetized magnets. These tend to be about 3 H @ 5K5 Ohms. I had posted some signal transformer links about here that can be (and are) used as guitar tuning inductors. This is exactly what Torres uses for their mid-boost kits as well a Bill Lawrence for his Q-cube thingy. From Mouser, these are a couple/few dollars. Inductors for Passive Guitars
|
|
|
Post by wolf on Aug 6, 2009 14:45:47 GMT -5
I guess this topic has been dormant for a while. Here's a variation of the circuit (on the right): Switching to parallel would change its tone characteristics and I've done this. It seems to yield a lot of the treble signal and not much else. What do you think?
|
|
|
Post by JohnH on Aug 6, 2009 21:59:33 GMT -5
Wolf - I think youd expect a strong peak at the resonant frequency determined by the combination of pickup/coil inductance and pickup capacitance/tone capacitor, and a general shunting of frequencies away from that. Is that how it seems to work?
John
|
|
|
Post by wolf on Aug 12, 2009 14:43:56 GMT -5
Just about every site I've surfed states that this circuit (or the Varitone circuit) acts by cutting the mids which leaves the bass and treble frequencies unchanged. They state that the circuit puts a "notch" in the frequency response. (Those sites also state that this is basically what a wah pedal does). From the discussions here, and my limited knowledge of electronics, a capacitor and inductor in series would pass treble and bass frequencies. However, if those frequencies are being led to ground, what remains (and goes to the amplifier) is the mid frequencies. So, even though this is a passive circuit, it is what most people refer to as a "mid-frequency boost" correct?
|
|