PAFs that sound like sh*t

[R.G.]

OK, we've established that you're not an A-hole and have some credentials, so thats a good thing.

Well, my wife tells me that I'm an a** a lot, so that may be premature, but I'll try to keep it civil...

Most of the steel in a PAF is cold rolled steel.

It pretty much had to be. Gibson being cheap, CRS is about the only commercial alloy to use, and I'm sure that they didn't buy special alloy.

If you say "steel" with no other modifier to a metal merchant, you get either CRS or hot rolled steel, HRS. These both start as furnace melts of very much the same stuff, steel with less than 0.8% carbon. Hot rolled is - yep, rolled out to shape while it's hot. This lets it get put into shape while it's hot enough to not work-harden from the shaping. It typically has a thickish coating of black mill scale from being worked hot in air. CRS is poured and rolled to rough shape, then let cool. It's worked down to final shape in rolling mills, which roll it to its final shape. This work hardens the outside, so the finished piece has a hard skin and soft interior.

I was referring to hardness in the mechanical sense. But mechanical and electrical hardness are very tightly linked, because both depend on the internal grains being pinned in place by defects and distorted by either mechanical stress or magnetic stress.

You make steel softer both mechanically and magnetically by annealing it. Magnetic is not exactly the same as mechanically soft, so dead soft iron is good, but there are some heat treatments that can improve it.

In addition, it's better magnetically if you get the carbon out entirely. Electrical iron is iron with as close to zero carbon as it's simple to make, and up to 3-6% silicon instead. Getting the carbon out reduces pinning defects around the grains to harden it, and adding silicon runs the electrical resistance of the steel up so eddy current losses go way down. This stuff is what is use for transformers and motors. Even so, bending or stressing electrical iron hardens it magnetically. It is possible to read the changes in electrical iron in a transformer after dropping it on a concrete floor *once*.

SethLover says over and over they used "soft iron." I have asked my screw maker guy what the fluck did "soft iron" mean in 1959?

The steel industry was pretty standardized then, so I bet it's pretty simple to find that out even if it's not exactly what we mean today. Today, "soft iron" is actually steel with carbon under 0.8% carbon. Low carbon steel can't be hardened to high hardness, and the lower the carbon, the lower the final hardness. Over 0.8%, you can harden steel to as hard as it can possibly be made, maybe Rc65 or a bit over if you do it perfectly.

I bet Seth meant "low carbon steel", probably what we call "MS" for either mild steel, which I understand is a misnomer; to the steel industry, "MS" means Mild Steel. Same stuff.

Obviously they didn't use pure iron pole screws or you would see the slots in the heads all chewed up over years and you just don't see that. Plus I don't think you could talk a screw maker into using soft iron for pole screws.

You're right. They're not pure iron. Too expensive.

I sent him pole screws from P90s from the mid-50s, the early 70s and Tim Shaw's 80s "repros" from the early 80s. He looked at them under a microscope and gave me his opinion. I had pole screws made in that alloy and the next lowest carbon ones, they both sound different, noticably. I also emailed a bunch of labs asking them if they could take a vintage screw and tell me what alloy they were made out of. I got a bunch of answers and methods some real expensive some pretty basic, never followed through on that because the answers were somewhat vague, and I just can't whip out a bunch of vintage screws to destroy without certain results. Maybe you have some ideas on how I can cheaply identify these alloys for dead certain, maybe a Rockwell hardness test?

Almost every decently equipped machine shop - as opposed to fabrication or welding shop - has a hardness tester. This is a widget that presses a diamond point into the metal and measures how hard you have to push to get diamond to go X deep.

I think what you want is a laser spectrometer test. This gadget zaps the surface of the sample with a laser, vaporizes a tiny bit, and reads the light spectrum of the glowing vapor. Each element glows differently, so it reads you out the elements in the sample.

Here's the rub - you have to do both, and maybe several hardness tests in different places.

Low carbon steel work hardens when you force it to bend past its elastic limit and hold a bend. That causes a noticeable loss in its maximum flux density (B) before saturation and an increase in its coercivity (H). It makes it a better permanent magnet, which makes it a poorer linear magnetic "conductor". So you need to know both the composition and where it's hardened. Shearing steel to shape work hardens it, and so does bending it.The hardening is local to the places where the work is done to it.

It is possible to soften steels by heating them to somewhat under melting, then cooling slowly. It *is* possible that the pole screws were batch softened in a furnace. There are heat-treat furnaces that could have had a few hundred to maybe a thousand screws in them for annealing; a furnace for up to about 1K screws at a time is cheap. In fact, I'm looking at buying a pottery kiln, which will do fine steel heat treating, used. They're for sale in my neck of the woods from free (Just get this thing outa my garage!!) to a few hundred.

Gibson could have done that. Probably not, sounds like a lot of trouble for the level of interest they've shown so far, but possible.

In my opinion the keeper bar and the slugs are different alloys than the screws. I only know that what alloy is used in each of these areas affects the tone very noticeably. If you want a good example go to the hardware store and buy some 3/16" rod off the shelf and make some slugs and put them in your bucker, horrible tone.

Now you're onto it. I think getting keeper and slugs of different alloy or softened is realistic. Ordering up even a single sheet (need a rail car for it) of 1002 (under 0.2% carbon) made up into slugs and poles would make a zillion pieces, and you could easily enough require your vendor to anneal them in an inert atmosphere (i.e. nitrogen or welding argon) and have you 10k pieces by the first of next month.

However, that 3/16" rod from the hardware store is probably low carbon, CRS. You could anneal it to be much better than it is now.

I'm very interested in learning what the B/H curve thing and how this test is performced, what gear is used. I know they use these curves in describing magnets as well but my magnet supplier is mostly a sales guy, middleman and he couldn't tell me and I couldn't find any info on the net, probably for lack of knowing what search terms to use, and not being an engineer, I'm a graphic artist and probably old as you except I'm still working my ass off in the hamster wheel :-)

"B" is the flux density, the amount of magnetic field juice coming out of each unit of the pole face of the magnet. "H" is the magnetic forcing needed to get "B" to come out the end. B/H is literally "how much magnetic field per unit area do you get out of the end of the magnet per unit of pushing on the material magnetically". The push is done with amperes per meter (useless for us) or ampere-turns, which is not useless.

To read the B/H curve, you rig up a small sample of the material so you can pump ampere-turns through a winding and read voltage off either the same or another winding. Imagine having a hunk of that 3/16" rod from the hardware store. Drill a hole in the end, cut off a washer, wind two coils on the resulting washer. Drive one coil with an audio amp and use a resistor to sample the current that flows. Read voltage off the second winding. The voltage on the second winding happens to be proportional to the change in flux, and the change is obviously proportional to the max flux, even when it saturates. Run those two signals into an oscilloscope in X-Y mode and you're looking at the B/H curve.

That presents the idea that the B-H curve is a single line, possibly curved, where you put in one value of H, get out one value of B. Not so. In general, The BH curve is a loop. Force steel to saturation in one direction, let it relax. The B value coming down is different at each H value than going up. Every time you force it to higher and lower B, the area inside the loop is used up as heat - that's hysteresis loss. The wider the loop, the bigger the loss.

Soft magnetic materials have skinny loops, not much hysteresis. Hard materials, especially permanent magnets, the B stays really high when H goes to zero. These materials have an almost square B/H loop and you have to work like crazy to get them to flip back and forth. Magnetic hardness and permanence go hand in hand. Magnetically soft stuff is good for signal linearity; hard stuff gives you a field to generate signal.

If you want to discuss off forum just hit my website and email from there and put PICKUPS in the subject line.

I'll drop you a line.

[Dave Kerr]

So, the upshot is, just about every facet of material handling matters to varying degrees. Using tin snips to shear sheet steel, hammering out bent corners, deburring with a hand file, roughing up the surface with sandpaper - they could all produce some effect on B and H.

This reminds me a bit of another hobby of mine, beer making. I'm pretty good at it, and can consistently make excellent beer. Consistency between batches, however, is virtually impossible for the homebrewer - yeast will act differently when subjected to subtle temperature differences, protein degradation in the mash varies noticably with small changes in the time/temp mash schedule, and on and on. Backyard brewers can make great (and not so great) stuff, while the pros can bang out almost identical batches day after day. I may not care for the taste of Budweiser, but can appreciate it as an engineering marvel.

So, a backyard metallurgist could torch that hardware steel rod to red hot, stick it into a bucket of sand and if he's lucky perhaps have an improved material. A pro will devote countless hours in optimizing that process using tight controls, swap those materials out in a known good coil until he likes what he hears, and then refines the multitudinous other variables until he's reached his own Mare-ville.

The pros, Possum in particular, have been preaching that attention to quality of steel parts is crucial, this annealing suggestion adds another level of refinement.

[NightWinder]

Well, my wife tells me that I'm an a** a lot, so that may be premature, but I'll try to keep it civil...


It pretty much had to be. Gibson being cheap, CRS is about the only commercial alloy to use, and I'm sure that they didn't buy special alloy.

If you say "steel" with no other modifier to a metal merchant, you get either CRS or hot rolled steel, HRS. These both start as furnace melts of very much the same stuff, steel with less than 0.8% carbon. Hot rolled is - yep, rolled out to shape while it's hot. This lets it get put into shape while it's hot enough to not work-harden from the shaping. It typically has a thickish coating of black mill scale from being worked hot in air. CRS is poured and rolled to rough shape, then let cool. It's worked down to final shape in rolling mills, which roll it to its final shape. This work hardens the outside, so the finished piece has a hard skin and soft interior.

I was referring to hardness in the mechanical sense. But mechanical and electrical hardness are very tightly linked, because both depend on the internal grains being pinned in place by defects and distorted by either mechanical stress or magnetic stress.

You make steel softer both mechanically and magnetically by annealing it. Magnetic is not exactly the same as mechanically soft, so dead soft iron is good, but there are some heat treatments that can improve it.

In addition, it's better magnetically if you get the carbon out entirely. Electrical iron is iron with as close to zero carbon as it's simple to make, and up to 3-6% silicon instead. Getting the carbon out reduces pinning defects around the grains to harden it, and adding silicon runs the electrical resistance of the steel up so eddy current losses go way down. This stuff is what is use for transformers and motors. Even so, bending or stressing electrical iron hardens it magnetically. It is possible to read the changes in electrical iron in a transformer after dropping it on a concrete floor *once*.


The steel industry was pretty standardized then, so I bet it's pretty simple to find that out even if it's not exactly what we mean today. Today, "soft iron" is actually steel with carbon under 0.8% carbon. Low carbon steel can't be hardened to high hardness, and the lower the carbon, the lower the final hardness. Over 0.8%, you can harden steel to as hard as it can possibly be made, maybe Rc65 or a bit over if you do it perfectly.

I bet Seth meant "low carbon steel", probably what we call "MS" for either mild steel, which I understand is a misnomer; to the steel industry, "MS" means Mild Steel. Same stuff.


You're right. They're not pure iron. Too expensive.

Almost every decently equipped machine shop - as opposed to fabrication or welding shop - has a hardness tester. This is a widget that presses a diamond point into the metal and measures how hard you have to push to get diamond to go X deep.

I think what you want is a laser spectrometer test. This gadget zaps the surface of the sample with a laser, vaporizes a tiny bit, and reads the light spectrum of the glowing vapor. Each element glows differently, so it reads you out the elements in the sample.

Here's the rub - you have to do both, and maybe several hardness tests in different places.

Low carbon steel work hardens when you force it to bend past its elastic limit and hold a bend. That causes a noticeable loss in its maximum flux density (B) before saturation and an increase in its coercivity (H). It makes it a better permanent magnet, which makes it a poorer linear magnetic "conductor". So you need to know both the composition and where it's hardened. Shearing steel to shape work hardens it, and so does bending it.The hardening is local to the places where the work is done to it.

It is possible to soften steels by heating them to somewhat under melting, then cooling slowly. It *is* possible that the pole screws were batch softened in a furnace. There are heat-treat furnaces that could have had a few hundred to maybe a thousand screws in them for annealing; a furnace for up to about 1K screws at a time is cheap. In fact, I'm looking at buying a pottery kiln, which will do fine steel heat treating, used. They're for sale in my neck of the woods from free (Just get this thing outa my garage!!) to a few hundred.

Gibson could have done that. Probably not, sounds like a lot of trouble for the level of interest they've shown so far, but possible.


Now you're onto it. I think getting keeper and slugs of different alloy or softened is realistic. Ordering up even a single sheet (need a rail car for it) of 1002 (under 0.2% carbon) made up into slugs and poles would make a zillion pieces, and you could easily enough require your vendor to anneal them in an inert atmosphere (i.e. nitrogen or welding argon) and have you 10k pieces by the first of next month.

However, that 3/16" rod from the hardware store is probably low carbon, CRS. You could anneal it to be much better than it is now.


"B" is the flux density, the amount of magnetic field juice coming out of each unit of the pole face of the magnet. "H" is the magnetic forcing needed to get "B" to come out the end. B/H is literally "how much magnetic field per unit area do you get out of the end of the magnet per unit of pushing on the material magnetically". The push is done with amperes per meter (useless for us) or ampere-turns, which is not useless.

To read the B/H curve, you rig up a small sample of the material so you can pump ampere-turns through a winding and read voltage off either the same or another winding. Imagine having a hunk of that 3/16" rod from the hardware store. Drill a hole in the end, cut off a washer, wind two coils on the resulting washer. Drive one coil with an audio amp and use a resistor to sample the current that flows. Read voltage off the second winding. The voltage on the second winding happens to be proportional to the change in flux, and the change is obviously proportional to the max flux, even when it saturates. Run those two signals into an oscilloscope in X-Y mode and you're looking at the B/H curve.

That presents the idea that the B-H curve is a single line, possibly curved, where you put in one value of H, get out one value of B. Not so. In general, The BH curve is a loop. Force steel to saturation in one direction, let it relax. The B value coming down is different at each H value than going up. Every time you force it to higher and lower B, the area inside the loop is used up as heat - that's hysteresis loss. The wider the loop, the bigger the loss.

Soft magnetic materials have skinny loops, not much hysteresis. Hard materials, especially permanent magnets, the B stays really high when H goes to zero. These materials have an almost square B/H loop and you have to work like crazy to get them to flip back and forth. Magnetic hardness and permanence go hand in hand. Magnetically soft stuff is good for signal linearity; hard stuff gives you a field to generate signal.


I'll drop you a line.

BAd ass RG! HEy, I just record, clip and play every steel- Thats great info to know, but you have to try it in real situations- PICKUPS!! Thats exaustive!!! Never stop learning

[David Schwab]

However, that 3/16" rod from the hardware store is probably low carbon, CRS. You could anneal it to be much better than it is now.

Yes, I believe it is. They call it "plain" steel. I made a number of bass pickups using plain steel rod from Home Despot when I was testing out pickup ideas, and they sound very good. My test was to bring a magnet with me and see how well it stuck to the rod!

I do find the tone muddies up fast as you wind more wire on the plain steel blades without a strong enough magnet.

I'm going to have to try annealing some.

BTW R.G., I've been reading your web site since it's been in existence. Good stuff there!

[R.G.]

So, the upshot is, just about every facet of material handling matters to varying degrees. Using tin snips to shear sheet steel, hammering out bent corners, deburring with a hand file, roughing up the surface with sandpaper - they could all produce some effect on B and H.

Pretty much. Sanding and filing tend to cut off the surface in such fine curls that only a few ten-thousandths (a 'tenth' in the machinist's trade) are affected. Sanding less than filing. Bending, hammering, shearing, yes, they work harden. Bending and hammering affect more of the material than shearing, unless it's a really dull shear. The effects are local too. Bend one end of a pole piece into alignment, you harden only that area that bent. The other end is unaffected.

This reminds me a bit of another hobby of mine, beer making. I'm pretty good at it, and can consistently make excellent beer. Consistency between batches, however, is virtually impossible for the homebrewer - yeast will act differently when subjected to subtle temperature differences, protein degradation in the mash varies noticably with small changes in the time/temp mash schedule, and on and on. Backyard brewers can make great (and not so great) stuff, while the pros can bang out almost identical batches day after day. I may not care for the taste of Budweiser, but can appreciate it as an engineering marvel.

I'd agree with that. Frankly, I think a good British bitter is more of a marvel than Budweiser, but as you say...

So, a backyard metallurgist could torch that hardware steel rod to red hot, stick it into a bucket of sand and if he's lucky perhaps have an improved material.

That's just a hair over the edge of the needed controls as I understand it. The actual temperature varies with the alloy, but interestingly enough, it's just a bit over the temperature where the steel is no longer attracted to a magnet at all. Above the Curie temperature, the thermal agitation upsets the spin pairing that makes ferromagnetism.That happens to be nearly the critical temperature for annealing. There is a required hold time above critical, and if you go too high for too long, the steel is effectively ruined and must be remelted to be useful. Hardening is about that easy though: heat with a torch until it is not attracted to a magnet, count your favorite number of seconds, and then jam it FAST into a quenching bath of oil, water, brine, etc. The outside is now as hard as that steel can get - and will shatter like glass under impact, and may crack just from the stress of cooling. You then have to re-heat it to 500-600F to temper/soften the superhard layer back down to tough enough to use.

A pro will devote countless hours in optimizing that process using tight controls, swap those materials out in a known good coil until he likes what he hears, and then refines the multitudinous other variables until he's reached his own Mare-ville.

Dead right. And the pro with his own setup can order known steels, anneal them to his specs and then experiment with selective work and temperature hardening. The big advantage is consistency - a pro can standardize his steel to the same place all the time. Annealing your steel will REALLY separate the men from the boys. Or at least the machinists from the boys...

The pros, Possum in particular, have been preaching that attention to quality of steel parts is crucial, this annealing suggestion adds another level of refinement.

This is going to sound like I'm preaching, and it's not intended that way. Knowing something in detail enough to state it in numbers, like for example the heat-treat history of your steel, is absolutely necessary to really knowing something. Even finding out that you like A1004 steel from a certain supplier will not tell you why the next batch is off, after your supplier changed the as-shipped heat history or rolling history. When you can run a full anneal and B/H curve on a snippet of the new sheet, you have the numbers in hand to figure out what to do - or at least when to change suppliers.

That's something that a notebook with a history of several hundred pickups made and tested would not give you. You'd change suppliers eventually out of suspicion and a lot of experimentation. If you know the numbers and the relationships, you can nail it in only a few tries.

I don't know that a full anneal is the right way to go. I only know it's one tool to use that has had a lot of effect in other electronic and audio applications.

The right sound may be to cut a slug of 3/16" rod, anneal it to soft, and bash it with a sledge hammer until it shortens by 0.020". Or it might be to heat it above critical, then quench it in "superquench" which is an internet recipe for a quench that hardens "unhardenable" mild steel.

Frankly, it's possible that work hardening history has been a hidden effect hiding differences inside otherwise identical steel parts.

It's also possible that this is not as huge a deal as I think. Remember - I know a lot about other things, not pickups... But I think it's a possible place for the researcher to look.

[R.G.]

Yes, I believe it is. They call it "plain" steel. I made a number of bass pickups using plain steel rod from Home Despot when I was testing out pickup ideas, and they sound very good. My test was to bring a magnet with me and see how well it stuck to the rod!

That's not a dumb test at all. In fact, depending on how good your feel was, it's maybe the best test you could do on a Home Depot floor. Was there a lot of variation?

I do find the tone muddies up fast as you wind more wire on the plain steel blades without a strong enough magnet.

That's interesting. I'd have to think for a while to come up with a reason for that change.

I'm going to have to try annealing some.

Take it as a point to experiment with. Annealing resets low carbon steel to its virginal magnetic condition. But then pumping a static magnetic field through it puts it back on a non-original B/H curve anyway. I don't know that annealing is The Way. It's a trick to try. But it is very reproduceable once you know what steel you have. It's a guaranteed reset to a known state to work from, not pot luck at the diner.

BTW R.G., I've been reading your web site since it's been in existence. Good stuff there!

You're very kind. Thanks. It's good to know I may have been some help.

[Joe Gwinn]

I'm going to have to try annealing some.

If the steel is plated with anything bright and shiny (other than Chrome), beware the fumes while heating the steel to a red heat for annealing. Do not breath the fumes. They can make you quite sick. Look up "zinc fever". Older stuff may be plated with cadmium, which is far worse.

[NightWinder]

David. The stuff from depot files way differently that say 1002. The shavings are different too . I have some annealed low carbon too, and it is harder to file than C.R.L.C.
This is interesting that the annealed has a stronger pull from the magnet, it wants to go right back to it as the C.R.L.C.(crlc breaks the same, but has less pull, if that makes sense?)

[Spence]

There's a thread on MLP forum about PAFs. Here's a quote from one of the senior members ( BCR Greg ). :

"I have three PAFs sitting here that have virtually NOTHING in common except the construction. The pickups were wound rather haphazardly, some are hot and fat, others weak but articulate. I have owned a significant number of PAF equipped guitars, and work on them for others on a regular basis. The only thing that is consistent on PAF pickups is the inconsistency."

Maybe someone forgot to use the Leesona the day those were made.

[R.G.]

rocket science? LOL.......Its just a vague honest aproach. Undersdtandably too. I fully get it : )

A kid at the last NAMM had on a T-shirt I liked. It said:

Come on! This ain't rocket surgery.

Another said:

I am a bomb technician. If you see me running, try to keep up.

[R.G.]

David. The stuff from depot files way differently that say 1002. The shavings are different too . I have some annealed low carbon too, and it is harder to file than C.R.L.C.

Really soft low carbon stuff is viewed as "gummy" by machinists and is more difficult to machine cleanly than modestly harder stuff. Cuts less cleanly. If I remember right, it does a soft pressure weld to the cutting edge, spoiling the edge.

There are alloys which have lead in them for free machining. 12L18 is an example. Machines great. Don't know what the magnetic properties are.

This is interesting that the annealed has a stronger pull from the magnet, it wants to go right back to it as the C.R.L.C.(crlc breaks the same, but has less pull, if that makes sense?)

I would guess that the permeability of the annealed is higher, so it "gathers" a magnetic field more effectively and over a longer distance. The CRLC has the same change in reluctance (which accounts for the same delta-E which makes the pull) but maybe has a lower permeability, which gathers flux less efficiently at a distance.

You know, now that I think about it, that "gathering flux at a distance" theory (which is just what popped into my mind after your comment) would have an effect on field focus in a pickup, wouldn't it?

The really bad thing about hardware store steel is that it may be *anything*. All it's guaranteed to be is steel, and cheap.

[R.G.]

Good one, Joe. Zinc and cadmium are really bad news in heat treating all right.

I suppose if one had some steel that you just HAD to use which was plated, you could file the surface. That should take off plating, because it's purely a surface coating. But still, it would be smarter to junk it as far as heat treating went. Steel's too cheap and livers are too expensive.

[Dave Kerr]

I had some zinc plated steel, gave it a dunk in muriatic acid for a bit and thay seemed to remove it.

[Steve Conner]

Hey, R.G., what do you say to the hypothesis that any guitar pickup can be modelled by a second-order lowpass filter cascaded with a comb filter and gain coefficient? That's where I'd start if I was writing a DSP PAF simulator. ;-)

If you agree with the above hypothesis, you can then devise a set of experiments to do parameter extraction on any pickups you might be interested in.

BTW, I still covet the Bare Knuckle Warpig anyway :P

[Joe Gwinn]

Good one, Joe. Zinc and cadmium are really bad news in heat treating all right.

I suppose if one had some steel that you just HAD to use which was plated, you could file the surface. That should take off plating, because it's purely a surface coating. But still, it would be smarter to junk it as far as heat treating went. Steel's too cheap and livers are too expensive.

The easiest way is to etch the zinc or cadmium off with battery acid.

More to the point, given the variation in what is sold as "steel" at hardware stores, and the small amount of metal in a pickup, I would suggest that people simply buy cold rolled steel rectangle stock from MSC or the like.

For instance http://www1.mscdirect.com/CGI/NNSRIT?PMAKA=82059148 for 1/8" thick 1018, or the fancy stuff (available in thinner gages) http://www1.mscdirect.com/CGI/NNSRIT?PMAKA=06200075, which is very accurately ground to size.