We've all seen the expensive toroidal winding machines on YouTube, and manufactured solenoids are always so neat in construction - as in: Each winding is neatly packed to its previous, without overlapping itself: like "-/////////-"

There's no explicit numerical answer here - I'm just after an indication of significance - if a winding was to be done in a factory - perfectly rolled, no overlaps, each coil immediately next to the previous one vs. one of the same specifications done at home with a few gaps, a few overlaps, and other imperfections here and there - what is the "ballpark" performance hit of that? Are we talking 0.1% or 10%?

If it's actually possible to perform some sort of calculation (with some assumptions put in there) - let’s use the example of: winding a ⌀35 mm annealed soft iron rod with ⌀1.2 mm magnet wire, wound 200 times across a 40 mm length.

  • \$\begingroup\$ Cant you actually do it better at home than a machine? \$\endgroup\$
    – DKNguyen
    Commented May 30, 2020 at 3:01
  • 4
    \$\begingroup\$ You don't necessarily need a machine to wind coils neatly by hand. I have done it plenty of times. As long as the wire is 30 AWG or bigger its not hard to do. It can even be done for cylindrical cores by just putting them in a variable speed drill press at the lowest speed and carefully guiding the wire with your finger. \$\endgroup\$
    – user4574
    Commented May 30, 2020 at 4:20
  • \$\begingroup\$ Are basically asking to compare 33 turns per layer in 6 layers (198 turns) versus (say) 25 turns per layer in 8 layers (200 turns), and see what the magnetic field density is at some small distance form the end of the solenoid is? If not, please shed more light. \$\endgroup\$
    – Andy aka
    Commented May 30, 2020 at 8:35
  • 1
    \$\begingroup\$ I've always used a machine for winding cylindrical coils with more than 100 turns. even if it was just a hand-powered drill. \$\endgroup\$ Commented May 30, 2020 at 9:57
  • 2
    \$\begingroup\$ If you are trying to maximise copper area in a restricted space to boost a motor's efficiency, being neat lets you get more turns in, or move up a wire gauge. You can even use square section wire and I've seen hexagonal wire in studio loudspeakers to improve packing. \$\endgroup\$
    – user16324
    Commented May 30, 2020 at 11:54

3 Answers 3


Compared to close-packed coils, how much space is inside a random scramble-wound coil? Not 10%, I bet it's more like 50%, and the total volume of the coil is nearly twice that of a close-packed version.

For random "scramble-wound" coils, the wire's turns/kg and the turns/km values are low, but also the average skin-depth for the entire inductor is large. That's what we want, so, for low-loss, high-freq RF coils we must avoid the close-packed windings. Buy a fancy scramble-type coil winder, and perhaps also wind your coils "pie-wound," as a stack of pancakes.

But for DC or 60Hz, a close-packed coil is much smaller, but with the same gauss/watt value as a big mushy scramble-wound coil. If physical size is an issue (motors for example, also solenoid actuators,) then those close-packed windings produce strong, miniature devices capable of high-wattage drive.

Also: vibration. Cheap, poorly-made motors will fail because the windings weren't tight enough. Some of the turns were vibrating, and this chewed through the wire's insulating varnish. Eventually a short-circuit developed. A motor with internally shorted turns will experience drag and heating, and may even "run away" into internal charring, fires. With motor coils, we want the coil to behave like a solid object, with nothing inside that ever wiggles.

Also: cooling! We can buy specialized coil-winding wire with square or rectangular section, which lets the windings pack together with minimal gaps. (Usually this wire size is well above 10AWG, intended for large transformers.) Thermally, the resulting rectangle-wire coil acts like a solid metal block, with high thermal conductivity. Fan-cool the outside, and the interior is cooled as well. On the other hand, a scramble-wound coil is full of insulating air: more like a hunk of styrofoam than a hunk of metal. It will have a smaller maximum wattage than a dense, non-scramble coil.

Below, as scrambley as possible? A few-mH value, HF tube-amp anode choke

enter image description here .

See also:



Depends on your performance parameters. Any wire looped through a high relative permeability \$\mu_r\$ core works pretty much the same magnetically regardless of position within the core. If there is an air gap in the magnetic circuit then position matters a bit more, but usually not a lot in such parts as gapped flyback transformers.

Coils wound in a "basket fashion" as in wbeaty's photo have lower distributed capacitance and thus a higher SRF (self-resonant frequency). So for RF applications, it matters a lot.

Coils for high voltage (typically the insulation on ordinary magnet wire is not rated for much voltage and cannot be relied upon for either safety or for functionality to withstand much voltage) may be wound on segmented bobbins and it's necessary to avoid the ends of the coil from being too close to each other or crossing over.

Messy winding may mean you can't use as much of the winding window that you would expect, so you're not using the core as efficiently as possible.


If we're talking about magnets or solenoids, which don't actually do any significant amount of work when turned "on" in a steady state, then:

For a given coil geometry, the efficiency (regardless of wire gauge) is determined by the amount of heat produced per ampere-turn.

For a given number of ampere-turns (= magnet strength), that amount of heat is determined by the resistance of the coil material.

If your coil is 50% empty space, then the average resistance of the coil material is twice as much as a densely-packed coil, and your magnet is therefore only 50% as efficient.

That's a pretty big deal.

  • 1
    \$\begingroup\$ I think there's a length/volume scaling problem here. Halving the density of the winding (upon a fixed form and fixed number of turns) will not double the amount of wire used and hence the resistance. The length of a single turn will increase by π times the change in diameter of that turn, so the net effect depends on the diameter of the core (and is smaller the larger the core is). \$\endgroup\$
    – Kevin Reid
    Commented May 30, 2020 at 15:18
  • \$\begingroup\$ @KevinReid I stipulated a fixed coil geometry, so if you halve the density but you want to keep the same number of turns, then you'll have to switch to a smaller gauge, and you'll end up using about the same length of wire. The resistance of the wire will double due to the gauge change. \$\endgroup\$ Commented May 30, 2020 at 16:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.