Plain Enamel wire insulation history

[salvarsan]

This is intended to stimulate some research on the origins and manufacture of plain enamel wire insulation.

What the wire manufacturors called "plain enamel" was known as yellow varnish, later oleoresin varnish, then specifically alkyd resin insulating enamel. Wikipedia covers alkyds; the varnishes had evolved to using vegetable oil feedstocks in place of rosin byproducts from turpentine distillation. It's easier to grow soy beans than pine trees.

A 1939 patent, Insulating enamel for electric conductors, discusses the general formulations for converting slow drying to fast drying enamels. The chemistry of alkyd resin formulations had progressed well beyond vagaries of natural rosins even before WW2.

It is probably a safe assumption that, by 1950, the majority of "plain enamel" insulations were vegetable oil derived semi-synthetic alkyd resins differing primarily in their proportion of unsaturated fatty acids and drying accelerants.

[salvarsan]

A 1995 article from Journal of Materials Science and Technology called "Some Dielectric Properties of Some Resin Varnishes" cuts straight to properties we're interested in -- dielectric constant and loss factors.

The alkyd resin AKA "plain enamel" specs at 4.1, slightly higher than most others in the MSW Wire Tech Book which cluster around 3.7 - 3.8.

[Mike Sulzer]

What high losses! A capacitor in a power circuit using this material as a dielectric would smoke. But I wonder what the losses are at audio frequencies? If they are similar, then they might be significant at resonance, especially in a pickup that does not use steel as cores. Does experience indicate that the the of insulation of the wire matters more in single coil pickups using alnico, or hum buckers using steel?

A 1995 article from Journal of Materials Science and Technology called "Some Dielectric Properties of Some Resin Varnishes" cuts straight to properties we're interested in -- dielectric constant and loss factors.

The alkyd resin AKA "plain enamel" specs at 4.1, slightly higher than most others in the MSW Wire Tech Book which cluster around 3.7 - 3.8.

[J M Fahey]

Enamelled wire varnish is optimized for insulation, must stick very well to wire, never crack or chip no matter how you twist that wire while winding, all very tall orders.

I bet dielectric constant is not a front row priority, if it compromises any of the others.

And since 99% of such wire runs at 50/60 Hz, as in power transformers and motors, no big deal (in their eyes) .

Output transformers or pickups are a very very small fish in that pond

Speaker voice coils, typically a couple grams each, even smaller

[salvarsan]

What high losses! A capacitor in a power circuit using this material as a dielectric would smoke. But I wonder what the losses are at audio frequencies? If they are similar, then they might be significant at resonance, especially in a pickup that does not use steel as cores. Does experience indicate that the the of insulation of the wire matters more in single coil pickups using alnico, or hum buckers using steel?

Now, that's a curious part. The first standard dielectric constant / "permittivity" test, ASTM D150, came about in 1987 and only tests at 1 MHz. 1987 was also the last year that NEMA included plain enamel copper wire (MW 1-C) in their specification.

MWS Wire's specs changed over the last 4 years. Their 2010 doc put most DC's at 6.2, formvar at 3.7 while the newest doc has most DC's at ~3.7 and formvar ~7.4. I don't know what changed. Maybe MWS' recent figures were normalized at 1MHz while the older ones were done at lower frequencies. I'll check on polyvinyl acetal resin (formvar) later.

The unencumbered 1995 JMST doc measured from 200 kHz to 10 MHz with roughly a 25% decrease dielectric constant with increasing frequency. They claim f0 = 10^5 in the abstract but all the data plots start at log10(f) ~5.3, about 200 kHz.

The standout was alkyd resin (plain enamel). Its very low loss factor, tan δ, was closest to zero at 200 kHz.

Page 121 table shows relaxation times and entropy changes (ΔS) for the 3 varnishes. Again, alkyd is the standout with the smallest entropy change -- the possible ordered states (molecular vibrational/conformational changes in this case) were limited suggesting a higher degree of polymeric cross-linkage in the cured insulation. More polymer cross linkage means a stiffer, more scuff resistant varnish than the polyester/amide and alkyd/polyesteramide types.

[David King]

What's APT? I have a spool of "HAPT and it's dickens to get it off for soldering. Fortunately it sounds great when you do get it soldered.

[salvarsan]

What's APT? I have a spool of "HAPT and it's dickens to get it off for soldering. Fortunately it sounds great when you do get it soldered.

From a quick UTFSE on "HAPT magnet wire", I found that:

Whitmore/Wirenetics sez their HAPT is Heavy Armored PolyThermaleze.

"Heavy" here means a double build, a polyester undercoat with a polyamideimide overcoat
conforming to the MW 35-C specification in the NEMA magnet wire standard. MWS Wire claims
the dielectric constant is an ordinary 3.86, close to solderable polyester, but coils wound with
heavy builds sound more alike than different irrespective of insulation.

SAPT is the single build version. These APT wires are spec'ed at 200C
max working temperature, are very resistant to moisture and most solvents.

I bet your HAPT wire coils almost never have insulation cut-through on the inner winds and will last a lifetime.

[salvarsan]

Plain enamel wire may be a minority product these days, but alkyd resin varnish is alive and well.
I bet the guy at American Wire in Sandy Hook, CT grabs a basic oil-based marine spar alkyd varnish
off the paint store shelf and doses it with a curing accelerant before use.

[J M Fahey]

Maybe

Or maybe they have a few leftover cans, from the good old days, which if unopened can last a lot.

Now once you open them and oxygen gets in, you have to use them within a few months, definitely less than a year.

FWIW I have in storage some > 20 years old silkscreen ink cans which I never used (think pink fluo, lemon yellow, etc. , what was I thinking? ) , rust has punched through and paint has started seeping out, fortunately self closing the leak.

I transferred to fresh unused cans ... as good as new

[David King]

He must add a colorant as well. Maybe we can get him to pull some purple -black wire next time

The best marine varnishes were phenolic based. There must be some phenolic wire insulations out there as well.
I see 4.9 on this chart that gives the dielectric constant for just about every material known to man.
http://www.clippercontrols.com/pages...nt-Values.html

[salvarsan]

At this point, I'm thinking that any dielectric constant that omits test frequency is suspect.
Pickup makers care about 100 Hz - 20 kHz, not 1 MHz where the behavior is different.

Until I saw the dielectric loss figures in the JMST article, they were unimportant.
Plain enamel's low frequency (well, lower than 1 MHz) loss is so small compared to
other insulations that it has to be significant.

And by the way, magnet wire insulation is not enamel, never has been. These are all
resinous varnishes, be they natural, synthetic, or things in-between. That alone halts
a lot of research at the start.

Edit:
Dielectric constant, loss, and strength are different.

D.constant is a basic measure of capacitance.
D.loss is just that: electrical energy that is lost forever as vibration, heat, and/or chemical decomposition.
D.strength is a maximum working voltage before failure.

[salvarsan]

The dielectric constant of alkyd varnish for magnet wire changes a lot with frequency --
At 60 Hz, if ranges 5.1 - 7.5 depending on formulation.
Compare this to the 4.1 at 200 kHz mentioned above.

Google books exerpts a table from Industrial Plastics: Theory and Applications By Erik Lokensgard, a 5th ed. 2008 book available for stupid money.

Search on "Industrial plastics: alkyd dielectric constant"

[Joe Gwinn]

I would submit that for pickup uses, it would be best to test in a pickup setup.

Specifically, wind some tight test coils on plastic bobbins (standardized size, no nearby metal, round bobbins would be OK), and measure the AC resistance as a function of frequency in the audio band (assumed to be well away from self resonance) The AC resistance minus the DC resistance (due to the wire) is caused by losses in the dielectrics, if there are no nearby metal masses (and their eddy currents) to interfere.

Based on low losses seen with handheld LCR meters at 1 KHz, my guess is that the losses won't be large, and because audio is well below most molecular bands, won't be all that sensitive to frequency in the audio band. My recollection is that these bands start in the 100 KHz range, but we will see in the AC resistance versus frequency plots.

Mike Sulzer's V-I test setup ought to be able to take the needed data.

[Joe Gwinn]

Audio-frequency dielectric data from Von Hippel

I recalled that I have a classic book on dielectrics, "Dielectric Materials and Applications", edited by Arthur R. Von Hippel, MIT Press 1954, 4th printing 1966, and this book has some audio frequency data for products available in the 1950s.

Formvar Type E: 100 Hz - 3.16, 54; 1000 Hz - 3.12, 100; 10 KHz - 3.08, 154; 100 KHz - 3.00, 190 (at 25 C, dielectric constant, loss tangent times 10^4) Page 334

There are a long list of other resins with names ending in "var" listed; all have similar behavior.

A number of Alkyd Resins are also listed:

Red Glyptal #1201: 100 Hz - 4.9, 760; 1000 Hz - 4.5, 600; 10 KHz - 4.1, 500; 100 KHz - 4.00, 400 (at 25 C, dielectric constant, loss tangent times 10^4) Page 348


Anyway, in the audio band, these are not frequency-selective enough to matter to the ear.