Wah Inductors and magnetism

[gmoon]

GBlekas:

First, I'm glad you're happy w/your mod. It's not surprising that an inductor with a DC resistance of only 15 ohms would need something to lower the 'Q', and I'm glad you "discovered" this empirically (that's the best way to learn something).

I see from your photos that I have the exact same LCR meter (labeled BM-4070L, but I'm sure it's identical). That little minus sign? Means turn the unit dial down. I tried measuring a 550 mH inductor just now for capacitance, and also got -3.xx at the "20uF" setting. Keep turning that thing down, and it outputs some small value at each setting, all the way down to 75pF at the lowest setting ("2000pF"). So it's not the right tool for this metric...

That calculator? It's figuring responses in an LCR "tank circuit"--the R and the C aren't the DC resistance and parasitic capacitance of the inductor itself. They are external components. The calculator isn't exact enough to let you enter the variables for the inductor that deviate from an "ideal" inductor. In the calc, L, C and R are fixed, and therefore so is the resonant frequency.

So that's also why we couldn't understand how you found the "resonant frequency" for the inductors alone--that's dependent on the all the components (L and C, anyways) in the "tank," not just the inductor. A wah doesn't wah unless the resonant frequency changes, after all...

[GBlekas]

GBlekas:

First, I'm glad you're happy w/your mod. It's not surprising that an inductor with a DC resistance of only 15 ohms would need something to lower the 'Q', and I'm glad you "discovered" this empirically (that's the best way to learn something).

I see from your photos that I have the exact same LCR meter (labeled BM-4070L, but I'm sure it's identical). That little minus sign? Means turn the unit dial down. I tried measuring a 550 mH inductor just now for capacitance, and also got -3.xx at the "20uF" setting. Keep turning that thing down, and it outputs some small value at each setting, all the way down to 75pF at the lowest setting ("2000pF"). So it's not the right tool for this metric...

That calculator? It's figuring responses in an LCR "tank circuit"--the R and the C aren't the DC resistance and parasitic capacitance of the inductor itself. They are external components. The calculator isn't exact enough to let you enter the variables for the inductor that deviate from an "ideal" inductor. In the calc, L, C and R are fixed, and therefore so is the resonant frequency.

So that's also why we couldn't understand how you found the "resonant frequency" for the inductors alone--that's dependent on the all the components in the "tank," not just the inductor. A wah doesn't wah unless the resonant frequency changes, after all...

Interesting on a number of levels.
The wrong meter readings end up being very close to the value of the cap used in the classic wah circuit to ground.
Plugging this incorrect info into the calculations makes it appear that the issue with the inductor was headroom related and further calculations lead me to believe that simply adding a resistor should correct the issue. (keep in mind this pedal has been sitting on the shelf for a few years now because it didn't sound right)
So, I take all this incorrect data and try correcting the wah and for once the pedal actually sounds right...but none of this is right?
It worked like a charm but it's all wrong....that's rather funny!

I think the Q resistor was a 33k....so where would you go from 33k to increase headroom?

[gmoon]

Actually, the irony here is that the results aren't wrong (even if the method was a little off.)

Plugging in a capacitance of 4 or 5uF gave your tank a resonant frequency in the audio range. A little low, but definitely in the ballpark. For better or worse, close enough to audio to not ring the WTF bell...

An LC tank works fine without any additional 'R'. You entered a small resistance value (the DC resistance for the inductors, which is fine). Increasing that resistance will reduce the 'Q' of your tank...which translates essentially into a wider band.

It's a subjective thing, but I've already said that too narrow a band (too high a 'Q') sounds nasty--to me anyway. Adding series (or parallel) resistance is one way to fight the icepick effect.

Despite the errors, you still recognized an outlier in the group--and a correlation between R and the bandwidth.

RE: 33K parallel R w/L -- that's going to be somewhat dependent on the DC resistance of the individual inductor... Your DC resistance measurements are pretty accurate, I think.  "Headroom"? I don't know. But higher R values will increase the 'Q' (narrow the band; limited by the actual DC resistance of the inductor) and lower will decrease the 'Q'.

[gmoon]

Ahhh hit quote by accident, and duplicated the response...

[Paul Marossy]

GBlekas:

Interesting that adding the 22 ohm resistor works for you. I'm curious about something, though. Have you ever checked out what is going on with Q1 in terms of voltage and maybe the Hfe in comparison to other good sounding "film can" wahs?

The reason I ask is because Q1 will only amplify what it sees coming back to the base from the inductor, and if the transistor gain is too low, then that might limit the "headroom" of those frequencies passing through besides the one selected by the operator. Just a thought based on what I've seen in my own wah pedal modding (you know, change the emitter resistor and it changes the sound some).

I'm just trying to think of reasons why an inductor that appears to be built to spec doesn't seem to sound right.

[GBlekas]

GBlekas:

Interesting that adding the 22 ohm resistor works for you. I'm curious about something, though. Have you ever checked out what is going on with Q1 in terms of voltage and maybe the Hfe in comparison to other good sounding "film can" wahs?

The reason I ask is because Q1 will only amplify what it sees coming back to the base from the inductor, and if the transistor gain is too low, then that might limit the "headroom" of those frequencies passing through besides the one selected by the operator. Just a thought based on what I've seen in my own wah pedal modding (you know, change the emitter resistor and it changes the sound some).

I'm just trying to think of reasons why an inductor that appears to be built to spec doesn't seem to sound right.

From my tests Q1 & Q2 need to have the same HFE or there will be white noise at the output.
HFE's in this wah were both 250 but I think the sweetest sounding are in the 150 range with anything below being too weak.

At 1 minute Clapton shows his wah and I swear I can hear white noise suggesting mismatched transistors in that wah but I could be suffering from EJ syndrome.
The switch isn't too quiet either....
http://vimeo.com/8119784

Changing Re1 only works to a point. I like the way it fattens it up but not distortion so while I would never use a 510 ohm I would only go down to a 390 sometimes depending upon the tone of the wah.

Funny thing is the Clydes swept the upper end mostly and while not ice picky they were painful at home but worked great in the band mix.
Dunlops and mods I have heard often seem to sweep a much wider range of tones and while cool at home they aren't great in a band mix.

As far as the inductor goes I think it was just an unhappy accident that it ended up this way.
By far this was the worst film can I had ever heard up until I added that resistor to it.
This is the first time I had ever tried adding the resistor but typically I have found that any inductor below 30 ohms doesn't sound good to my ear.
Most inductors pulled from Dunlops and Vox wahs would read around 19 ohms so I have a box of those so perhaps I should try the series resistor with one of those to see how that works out.

[GBlekas]

For better or worse, close enough to audio to not ring the WTF bell...

Never heard that saying before but it's really funny!

[Paul Marossy]

From my tests Q1 & Q2 need to have the same HFE or there will be white noise at the output.
HFE's in this wah were both 250 but I think the sweetest sounding are in the 150 range with anything below being too weak.

Interesting. I'm sure that the old wahs must be all over the map as far as the transistors go, depending on what wahs you want to talk about. The Italian ones seem to be the ones with the most variations in parts from what I have seen.  

Changing Re1 only works to a point. I like the way it fattens it up but not distortion so while I would never use a 510 ohm I would only go down to a 390 sometimes depending upon the tone of the wah.

Yeah, just was pointing out that messing with the gain of Q1 does also changes things some.

As far as the inductor goes I think it was just an unhappy accident that it ended up this way.
By far this was the worst film can I had ever heard up until I added that resistor to it.
This is the first time I had ever tried adding the resistor but typically I have found that any inductor below 30 ohms doesn't sound good to my ear.
Most inductors pulled from Dunlops and Vox wahs would read around 19 ohms so I have a box of those so perhaps I should try the series resistor with one of those to see how that works out.

Huh, you might be able to use those things up after all.  

[gmoon]

For better or worse, close enough to audio to not ring the WTF bell...

Never heard that saying before but it's really funny!

Think it was just off the (cough) cuff...

Still-- it would be cool to see real data on the original premise--magnetic offset effect on wah inductors. Maybe RG, Merlin, PRR, (and I'm leaving other EEs out, for sure), etc., can give you some pointers about a test rig. For that matter, there's probably not much real data on those inductors at all, other than DC resistance and inductance.

Seeing as you've got an impressive vintage collection of those thingies...

[joegagan]

in richard petersen's 1963 patent ( wow -wow trumpet sound for organ) that was used for the maestro boomerang, he included several drawings of ways to alter the frequency. one was an LDR being activated /modulated by a light bulb. another was metal or a magnet moving across the inductor actuated by a knee lever, since this was originally an organ design. cool stuff.

[R.G.]

Still-- it would be cool to see real data on the original premise--magnetic offset effect on wah inductors. Maybe RG, Merlin, PRR, (and I'm leaving other EEs out, for sure), etc., can give you some pointers about a test rig.

As far as I can tell, the magnetic offset data originated with me in the first half of the 1990s. At the time, you could have counted DIY effects makers on your fingers - well, maybe you'd need toes, but there were pretty few. I had a friend who ran a used-vintage-guitar shop and he knew that I was a mania...er, enthusiast  about old pedals and would call me when he got something unusual in. He loaned me the original wah pedal I tested, along with the advice that if I messed it up, I bought it... gulp. It was reputed to be one with one of the "magic inductors" as well as they were known at the time.

Back then, I had access to a fully-equipped EE lab, and so I tiptoed in and set up some testing. I ran AC inductance, variation of inductance with frequency, self-resonance, Q at a variety of frequencies, all kinds of stuff and nothing showed up as markedly different from a (then) modern and decidedly un-magic wah inductor; at least nothing that wasn't explained by the usual measured R-L-C data.

For some reason I decided to run a spectrum analyzer on it. This was before PCs had sound cards capable of spectrum analysis, so I use a lab-grade HP setup feeding a high-purity sine wave to the inductor through a high resistance and measuring the spectrum of the voltage across it. What I found on the garden variety inductor was that for small signals, the resulting spectrum was a pure-ish sine that was not significantly different from measuring the sine generator itself until I pushed the signal level up to where the inductor started into saturation. Then I got distortion product spikes at 3x the input frequency, which is exactly what one would expect from a saturating inductor. It's clipping as the inductor can't support a voltage that high on its B-H curve.

When I tried the magic inductor in the same rig, all was the same until clipping started. What was distinctly different was that the first harmonic that came up was the **second** harmonic, not the third. Second harmonics cannot arise from symmetrical distortion products, they can only arise from asymmetrical distortion. I was so confused on this that I consulted one of my EE mentors about it. His response was "It sounds like there's an offset inside the inductor. That's what DC generators rely on to start up, remanent magnetic offset." A little more dinking turned up the verification. The inductor was going into saturation on one polarity of the signal first, and adding a little DC offset to the incoming signal could make this worse in one direction and better in the other.

The core itself was magnetized, and signal ran into saturation earlier on one polarity than the other. Looking at the circuit, I found that it puts a bit of DC current through the core for biasing Q1. The core material was fairly cheap and had some significant remanent flux, and there you are. The core was magnetized. Better ferrites don't have much remanence.

Those tests were almost 20 years ago. I posted the result to usenet (there was no web to browse or browsers to do it with at the time), and it appears that it's passed into myth and legend with the real info lost. It was a lucky accident that I had both a magic and non-magic inductor to measure and the equipment to do it. I don't know if soundcards could do the same test today. Maybe. Depends on the signal levels, I guess.

For that matter, there's probably not much real data on those inductors at all, other than DC resistance and inductance.

But it's not hard to get, given access to the inductors. I personally don't see much difficulty in getting a lot more info given the time and inclination to do it. I'd run inductance versus frequency versus signal level with AC signals, distortion spectrum on the results, probably self resonance and pulse-inductance tests on them, plus whatever else suggested itself.

There really isn't any magic there - only a technology you don't yet understand.

[joegagan]

thanks for all the info, RG.

you may recall, due to serendipity, we spent time at the mecca, your maniac/enthusiast friend's studio and home while we used to tour, he and D were very gracious to host us when were down there. his collection is strange and mindblowing. roland funny cat? sure, got one of those.  a dozen weird 60s italian made fuzzes? got that. drool worthy 50s cowboy guitars? check.

[R.G.]

you may recall, due to serendipity, we spent time at the mecca, your maniac/enthusiast friend's studio and home while we used to tour, he and D were very gracious to host us when were down there. his collection is strange and mindblowing. roland funny cat? sure, got one of those.  a dozen weird 60s italian made fuzzes? got that. drool worthy 50s cowboy guitars? check.

It is a small world. Sad to say, he's recently moved to New Mexico, a town where a car passes through only a few times a day. Took his collection of esoterica with him. I'll miss him being here.

[gmoon]

Thanks for the info, RG. It fleshes out a bit what's in the geofex article. Makes me wonder if most of usenet isn't archived somewhere...

As far as I can tell, the magnetic offset data originated with me in the first half of the 1990s. At the time, you could have counted DIY effects makers on your fingers - well, maybe you'd need toes, but there were pretty few.

Compared to today, that's probably true. But it's gotta be a slight exaggeration--I built my first DIY projects in the late 70's, all thanks to Craig Anderton's wonderful book. And his column in Guitar Player ran for years, concurrent to the books he published. Plus the electronics rags always had a handful of guitar projects.

So there were more than a few builders out there. I wonder if there were fewer in the mid-90's, than the early 80's though...

@GBlekas:
What R.G. describes requires some substantial gear (and expertise) to tackle, eh?

You have any colleges or unis close by? There's gotta be an EE student with access to all this stuff who's also into guitar. Or a prof, for that matter... You might contact someone in the electrical engineering program.

[GBlekas]

Thanks for the info, RG. It fleshes out a bit what's in the geofex article. Makes me wonder if most of usenet isn't archived somewhere...

As far as I can tell, the magnetic offset data originated with me in the first half of the 1990s. At the time, you could have counted DIY effects makers on your fingers - well, maybe you'd need toes, but there were pretty few.

.

@GBlekas:
What R.G. describes requires some substantial gear (and expertise) to tackle, eh?

You have any colleges or unis close by? There's gotta be an EE student with access to all this stuff who's also into guitar. Or a prof, for that matter... You might contact someone in the electrical engineering program.

While I have been into electronics since 4 years old (believe it or not) R.G. Keen has been the internet guru for many, myself included,  and I am trying to digest what he has said. I am sure Joe Gagan and I aren't finished here.....but, we are persistent buggers and simply don't give up or go away that easily! :>)

That being said I am trying to figure out how to prove that offset using what ever necessary and maybe even learn something in the process.
I have a signal generator and Oscilloscope and looking into exactly how to set this all up to prove or disprove so any advice would be greatly appreciated.

I have ears so I know magnetism near the inductor smooths the highs so it's doing something without a doubt. I also like what it's doing as it sounds pleasant to the ears.

So, how would I set up a test rig?

R.G. said he saturated the Fasel inductor and it went into harmonic overtone quicker than a newer inductor but does this imply hooking up signal directly to the inductor or is the inductor in the wah circuit?  Is this set up so you can use the same wah circuit popping in different inductors to see how they behave under signal or is this signal hooked up directly to the inductor under test?

Just wanted to say thanks to you all for the input as it really is appreciated!

George

[gmoon]

Excellent-- I was hoping you'd soldier (solder?) on...

[R.G.]

Compared to today, that's probably true. But it's gotta be a slight exaggeration--I built my first DIY projects in the late 70's, all thanks to Craig Anderton's wonderful book. And his column in Guitar Player ran for years, concurrent to the books he published. Plus the electronics rags always had a handful of guitar projects.

So there were more than a few builders out there. I wonder if there were fewer in the mid-90's, than the early 80's though...

Of course it's an obvious bit of hyperbole. There were definitely more than twenty. But the number of electronics tinkerers in general, and music tinkerers in specific were dramatically smaller than we have now, when there are over twenty *thousand* members at this one forum. The number of electronics experimenters went down dramatically in the 80s and 90s, as witness the magazines went out of publication. The 60 were probably the earlier peak, and the late 70s were a period of decline. I think that would match your perception. The internet reinvigorated the field with massively more available information and a new sense of community.

... I am sure Joe Gagan and I aren't finished here.....but, we are persistent buggers and simply don't give up or go away that easily! :>)

Good. It's a good place to do some detail work.

That being said I am trying to figure out how to prove that offset using what ever necessary and maybe even learn something in the process.
I have a signal generator and Oscilloscope and looking into exactly how to set this all up to prove or disprove so any advice would be greatly appreciated.

Proving it or not is easy. Find a magic inductor, one with reputed magic powers of tone, then take it all apart and use a hall effect magnetometer to measure the remanent field. The only reason I didn't do that was that I was promised a quick and messy death by the guy who loaned it to me if it didn't work right when he got it back. 

Another measurement technique is to replicate the experiment. Feed a high purity sine from a high impedance source into the inductor and run a distortion/spectrum analysis on the result, which is what I did. I just glossed over the several hours of tinkering with various setups which showed nothing at all, but then I always did enjoy just being in labs. 

I have ears so I know magnetism near the inductor smooths the highs so it's doing something without a doubt. I also like what it's doing as it sounds pleasant to the ears.

That fits with the asymmetrical magnetic offset theory. A touch of second harmonic has a sound reputed to be smoother and sweeter, and very hard to pick out by ear as even distorted at all.

So, how would I set up a test rig?

I suspect that a rig intended to feed a 100Hz sine as a current into the inductor and then measure the voltage with a sound card and program to run an FFT would do it. It's where I'd start.

R.G. said he saturated the Fasel inductor and it went into harmonic overtone quicker than a newer inductor but does this imply hooking up signal directly to the inductor or is the inductor in the wah circuit?

I isolated the inductor outside the wah. There are too many confusing variables inside the wah circuit. Besides, if the magic was actually in the inductor, it should be there with the inductor outside the wah circuit, too, right?

It could be argued that feeding the inductor from about  5V through a 470K resistor to fake the bias current going through it is necessary. I didn't do that one that way at the time (as I can remember - it's been nearly 20 years) but it's probably a good one to run for verisimilitude

The other thing I liked was feeding a small secondary winding in a non-magic inductor with a small current. This forces a magnetic offset on any core, and replicates sticking a magnet on the inductor, but has the disadvantage of acting like a transformer secondary. You need to feed this from a very high source impedance to keep it from loading down the main-inductor primary. I had mixed results with this back when. And since wah pedals are not my primary interest, I moved on.

[Paul Marossy]

When I tried the magic inductor in the same rig, all was the same until clipping started. What was distinctly different was that the first harmonic that came up was the **second** harmonic, not the third. Second harmonics cannot arise from symmetrical distortion products, they can only arise from asymmetrical distortion. I was so confused on this that I consulted one of my EE mentors about it. His response was "It sounds like there's an offset inside the inductor. That's what DC generators rely on to start up, remanent magnetic offset." A little more dinking turned up the verification. The inductor was going into saturation on one polarity of the signal first, and adding a little DC offset to the incoming signal could make this worse in one direction and better in the other.

Aha! I think THAT is the key element right there - the 2nd harmonic. IIRC, it was also a red Fasel inductor that was tested, not the "halo" inductor in the larger than life "holy grail" sell everything you have to get one Clyde McCoy wah pedal.

So then maybe it's possible that a few of the Clyde McCoy wahs have some inductors that have this DC offset happening in them as well. I am still not convinced that all you have to do is come up with the exact same inductance, DC resistance and make an exact physical reproduction to get a magical sounding inductor. Maybe those "magical" sounding wahs are actually wahs that have what should be considered faulty inductors, based on the information that RG provided above. And yet people find that desirable. Kind of like people prefer tubes over solid state devices. Tubes are imperfect devices with much higher THD than solid state devices, but I like the sound of them better.

Interesting stuff anyway. 

[Philippe]

Resistor selection being the key ingredient to good wah tone, isn't it amazing how inductor make/design/mojo/hyperbole has become the key distraction/element of all things pertaining to wah?

[R.G.]

Resistor selection being the key ingredient to good wah tone,

It is? How exactly is that?