Consistency and Quality Control

[radrobgray]

What metrics do you use to measure consistency for same model pickup builds? What is a acceptable level of tolerance? I am setting out to do a run of the same pickup, they seem pretty close spec wise. I checked them with my LCR and DMM. I'm looking at +/-0.037 Ohms and +/-0.007 H.

[copperheadroads]

If your not in an temperature controlled room ,well good luck with that .

[mozz]

Machine wound or hand wound?

[Helmholtz]

What metrics do you use to measure consistency for same model pickup builds? What is a acceptable level of tolerance? I am setting out to do a run of the same pickup, they seem pretty close spec wise. I checked them with my LCR and DMM. I'm looking at +/-0.037 Ohms and +/-0.007 H.

I guess you mean +/-37 Ohms and not 37 milliOhms. Still not very realistic. I would allow at least for +/- 100 Ohms.
Inductance is far more important than resistance but I think something like +/- 5% (or 0.1H) should be sufficient.

If your LCR meter allows the measurements, I suggest to control inductance as well as Q factor at 1kHz. An abnormally low Q will indicate shorted windings.
The limits you need to determine from several runs.

[kayakerca]

I'm looking at +/-0.037 Ohms and +/-0.007 H.

That would be impressive! I remember seeing a SD factory tour video and if I remember correctly the guy doing the interview said they look to keep their winds within 3% to be accepted. Didn't specify within 3% of what though. I look to stay under 1% for any given recipe (based on my winding shop room temperature).

[eschertron]

I'm looking at +/-0.037 Ohms and +/-0.007 H.

Without knowing any better, maybe this is an engineering shorthand for "+/-3.7%" Ohms and "+/-0.7%" Henrys. Sounds a bit more reasonable?

edit: just because I couldn't let it go, a 5.4k coil wind would end up +/-200 Ohms with a 3.7% tolerance.

[Antigua]

With guitar pickups, some of the supposed mystique of vintage pickups is that they had wide tolerances, allowing you to accidentally end up with something unexpectedly pleasing, like opening up a pack of baseball cards. Whatever people's rationale is for valuing old pickups over new ones, wider tolerances were one of their features.

The best way to truly quality control a pickup would be to create a bode plot of the pickup's impedence, as that would actually reveal how the pickup would actually function, where as resistance and inductance values are are merely clues, with inductance being the better clue of the two. A bode plot can also reveal the Q factor, and I don't believe that electrical shorts within the windings is probably something that actually happens, but I have seen variations in AlNiCo pole pieces that resulted in a different resistivity, and/or permeability, which resulted in pickups having difference Q factors.

Here's an example of three Fender PV 56, a set which is supposed to be three identical pickups:



Even though there are three overlapping lines, you'd never know it due to how closely all three pickups function.

Here's is a Fender PV 59 set by contrast:



There's more drift with the 56 set than the 59. The loaded resonant peaks (with an added realistic amount of capacitance) are still close enough that they probably do all sound the same, but in terms of QC for its own sake, the 56 set shows a higher level of precision. I don't know how Fender makes pickups, so I'm not sure why there is this variety in tolerance, but it's there to be seen. Based on my own experiences, I'd guess that both sets are completely machine made, and it would difficult to get the level of precision seen in the 56 with a hand guided coil, a set of hand guided coils will look more like the 59 set, because as the wire is randomly laid, a slightly different inductance and capacitance will result in the final product, yielding a semi random resonant peak. I don't think it's even necessary to even achieve such a high precision if you make and sell pickup that are deliberately hand made, but for one's own purposes, the bode plot would do the best job of indicating how alike the pickups actually are, or are not. If you were to see wider variation that you wanted, you could find the inductance, then solve for the capacitance, and with that much information on hand, probably narrow down the cause of the problem.

It would also be good to have a gauss meter on hand to make sure that all of you pickups show the same magnetic strengths. An AlNiCo supplier could send you a botched supply, and you might see this in the bode plots, especially if you set aside a "control pickup" whose purposes is to compare magnets, with the coil remaining as a constant, but if the AlNiCo is electrically the same, but has a deviated residual flux density, you'd only know it by checking gauss strength.

[Helmholtz]

It would also be good to have a gauss meter on hand to make sure that all of you pickups show the same magnetic strengths.

Good point.

[ken]

With guitar pickups, some of the supposed mystique of vintage pickups is that they had wide tolerances, allowing you to accidentally end up with something unexpectedly pleasing, like opening up a pack of baseball cards. Whatever people's rationale is for valuing old pickups over new ones, wider tolerances were one of their features.

Leo Fender supposedly shot for a 20% tolerance on his DCR, and used uncharged magnets charged up with a charger made from an old 6volt truck starter and a battery. Supposedly he only charged up the battery when it went just about dead. I wonder what he would have thought about keeping a 3% tolerance?

There are so many variables in pickup making today, IMO much more than making any other musical electronic equipment. Nowadays, if you make amps or effects it's easy to buy 2% tolerance resistors and close tolerance caps, but for us holding tight tolerance on a 5 pound spool of copper wire can be difficult. What do you do if your brand new spool of 42 gauge goes from nom/max to min/nom or less in the middle of its third or fourth coil?

This was Leo Fender's original 1950's pickup winder. It sorta explains the 20% tolerance. Click on the photo to expand it



Ken

[eschertron]

Hey! I think I have the motor from the same sewing machine!

[radrobgray]

I want to make sure I'm winding a consistent product, I want to offer models. I'm hand winding some humbuckers and did two different sets. The finished pickups are all between 9.36-9.40k and 6.174-6.220H, all magnets are around 500 gauss according to the meter on my Mojo machine. I build the devices from this page http://kenwillmott.com/blog/archives/152 to plot the data, but haven't had time for it yet. +/-3% Ohms and Henries seems obtainable. What about Resonant Peak or Capacitance?

Here is the data from the run for the individual coils:

Set 1 - Inductance @ 120HZ, Ohms
1.382H, 4.826K
1.412H, 4.846K
1.409H, 4.847K
1.405H, 4.852K
1.412H, 4.887K

Set 2 - Inductance @ 120HZ, Ohms
1.275H, 4.557K
1.275H, 4.552K
1.279H, 4.551K
1.273H, 4.588K
1.291H, 4.561K

[Antigua]

I want to make sure I'm winding a consistent product, I want to offer models. I'm hand winding some humbuckers and did two different sets. The finished pickups are all between 9.36-9.40k and 6.174-6.220H, all magnets are around 500 gauss according to the meter on my Mojo machine. I build the devices from this page http://kenwillmott.com/blog/archives/152 to plot the data, but haven't had time for it yet. +/-3% Ohms and Henries seems obtainable. What about Resonant Peak or Capacitance?

Here is the data from the run for the individual coils:

Set 1 - Inductance @ 120HZ, Ohms
1.382H, 4.826K
1.412H, 4.846K
1.409H, 4.847K
1.405H, 4.852K
1.412H, 4.887K

Set 2 - Inductance @ 120HZ, Ohms
1.275H, 4.557K
1.275H, 4.552K
1.279H, 4.551K
1.273H, 4.588K
1.291H, 4.561K

Those are very close together, so close that I don't think there is cause to plot the impedance. If the inductances are that tight, I don't see why the capacitance wouldn't be either.

[Joe Gwinn]

The best way to truly quality control a pickup would be to create a bode plot of the pickup's impedance, as that would actually reveal how the pickup would actually function, where as resistance and inductance values are are merely clues, with inductance being the better clue of the two. A bode plot can also reveal the Q factor, and I don't believe that electrical shorts within the windings is probably something that actually happens, but I have seen variations in AlNiCo pole pieces that resulted in a different resistivity, and/or permeability, which resulted in pickups having difference Q factors.

We do see inter-winding shorts for sure, and one can hear the effect. It's a standard cause of returns. Sometimes it's wire with brittle enamel that flakes off during winding. Sometimes its the pressure of the outer windings forcing the inner windings together and/or into an edge or roughness on the magnets in a strat singlecoil. Five thousand turns of fine wire with a winding tension of 35 grams yields a total constriction force of (5000)(35)= 175 kilograms.

There have been discussions of this before, and one defense is to take inductance and AC resistance readins of event pickup coil made, and look for units with uncommonly high excess of AC resistance over DC resistance.

[Helmholtz]

...and look for units with uncommonly high excess of AC resistance over DC resistance.

This is actually equivalent to my proposal above. The Q value of an inductor is proportional to L/Rs, Rs being the equivalent AC series resistance. Often meters display the D or DF (dissipation factor) value. As D=1/Q, it gives the same info as your method.
Q, D and AC resistance measurements should preferably be taken at 1kHz and not at 100/120Hz, as the losses caused by inter-winding shorts increase with frequency.

[Antigua]

We do see inter-winding shorts for sure, and one can hear the effect. It's a standard cause of returns. Sometimes it's wire with brittle enamel that flakes off during winding. Sometimes its the pressure of the outer windings forcing the inner windings together and/or into an edge or roughness on the magnets in a strat singlecoil. Five thousand turns of fine wire with a winding tension of 35 grams yields a total constriction force of (5000)(35)= 175 kilograms.

There have been discussions of this before, and one defense is to take inductance and AC resistance readins of event pickup coil made, and look for units with uncommonly high excess of AC resistance over DC resistance.

Suppose you jave 8,000 turns of wire, and suppose a short causes a hundred of those winds to become "coil within the coil", what you end up with is 7,900 turns and a source of eddy currents in relation to the shorted 100 turns. These eddy currents would cause the pickup to become a little darker, but not a lot, maybe not even enough to notice. So why do people return these shorted pickups? Because they're a little darker?

As far as comparing the AC and DC resistance, that sounds OK in theory, but what's missing in a quantifiable control, to say "X amount of shorted turns results in an increase of Y at A hertz", because without that's you'd just be guessing. Differences in the core materials will cause eddy currents as well, so you would have to account for that somehow as well.

I could believe the enamel would flake off from the outside of the coil, after all there is nothing to keep it in place, but where is the enamel within the coil going to travel to?