Wire Wrapping: Discrete Components

[ElEctric_EyE]

The hand tool is fine and cheap and strips wire too. If you have the patience that's fine, but you may want to progress to a faster technique when you get some hours of experience under your belt.
Personally, I found that faster wire wrapping made me less prone to errors so I use a Gardner-Denver AC powered electric gun and a wire stripper that auto strips about a 1" of insulation. I just saw a similar gun on ebay for $60 used and I believe I paid about $20 for the stripper. These tools also go by the name of Cooper. Anyway, it takes about 30 seconds to cut the appropriate length of wire, strip both ends and zap it from point A to point B.
Also, I would recommend to you that you use 30AWG for all non-power/grounds. And use a thicker AWG for the powers/grounds.
There's more to say but I don't want to rush you.

[Johnny Starr]

I guess it depends on how much WW I end up actually doing. I appreciate having this forum as a resource though. The web doesn't offer much documentation on the subject. I'm sure I'll have questions though

I'll be sure and upload pics as I go to make sure I'm doing it right as well.

[GARTHWILSON]

Quote:

I think a lot of people my age are spoiled with tablets and phones; even those technically inclined. Most of my friends think I'm crazy for "wasting" my time with "that old stuff". Honestly, I would rather spend $150 on wire wrap and other parts than on the latest trends. I have fun for years with this stuff while they run the treadmill of trying to have the latest generation [blank].

I run into this kind of thing with the calculators.  People say, "Why would you use a 1980's HP-41 when you could load a simulator on your smart phone?"  I've had programs that I use all the time in the HP continuously for 20+ years, never having to reload, learn a new system, etc..  That new smart phone will be in the land fill in 2-3 years.  And support?  There are still new modules being introduced and sold for this nearly 35-year-old design.  I can control equipment on the workbench with it and take data, unlike the smart phone, and of course interface to printers and mass storage, and I just got a bar-code reader for it, free.  Not even modern graphing calculators can do that stuff.

Modern consumer stuff is meant to be quickly disposable, for people with very short attention spans.  The 6502 may be the most-supported processor ever, and it may outlive some modern ones.

I never tried the electric wire wrappers, but a friend who was used to the OK Industries tool tried one and was very disappointed in the quality of the wraps.

If you use the prototyping board with a ground plane, you won't need ground wires.  The wire size does not affect the inductance much anyway.  There's a wire-inductance calculator here.  [Edit, 4/3/14:  That domain seems to have expired.  There's another online wire inductance calculator at http://www.eeweb.com/toolbox/wire-inductance/ .]  4" of 30-gauge wire has 135nH inductance; but note that going to a larger wire has negligible effect on that, as the same length of 24-gauge wire (doubling the diameter) has 121nH.

[ElEctric_EyE]

GARTHWILSON wrote:

...Modern consumer stuff is meant to be quickly disposable, for people with very short attention spans...

Amen, brother! And it's a sad shame it's like that too.

GARTHWILSON wrote:

...I never tried the electric wire wrappers, but a friend who was used to the OK Industries tool tried one and was very disappointed in the quality of the wraps...

As far as electric, I've never tried the battery powered ones, the AC powered ones are consistent every time.

The 2 details one must adhere to:
1) The amount of pressure pushing downwards on the gun while pulling the trigger, i.e. wrapping, and also the length of time holding the trigger. Too light or lifting while wrapping results in a 'loose coil', which can easily be fixed by gently pushing down after wrapping. Pushing down to much results in nasty, sometimes so nasty an unwrap tool cant catch the wire off the post (another good reason to own a manual wirewrap tool BTW, it has unwrap built in when you turn the tool CCW).
2) Lastly too little trigger and you have an incomplete wrap. This is usually easily fixable as you squeeze the trigger while coming down on the unfinished wire. The gun usually will auto-engage the wire which is sticking straight up and parallel to the post, and finish the wrapping job.

GARTHWILSON wrote:

...The wire size does not affect the inductance much anyway....

It does affect the amount of current that will flow, and hence potential noise that will be introduced. Here is a good table I use to get a grasp on AWG and conductance.

[GARTHWILSON]

ElEctric_EyE wrote:

GARTHWILSON wrote:

...The wire size does not affect the inductance much anyway....

It does affect the amount of current that will flow, and hence potential noise that will be introduced. Here is a good table I use to get a grasp on AWG and conductance.

The amount of resistance in a 2" 30-guage connection is only about 0.017 ohms, not enough to matter at all in this kind of application.  At the VHF/UHF power transistor manufacturer where I worked in applications engineering in the mid-1980's we used very small stripline traces at close to a GHz to handle thousand-watt bursts for radars.

[ElEctric_EyE]

GARTHWILSON wrote:

The amount of resistance in a 2" 30-guage connection is only about 0.017 ohms, not enough to matter at all in this kind of application. At the VHF/UHF power transistor manufacturer where I worked in applications engineering in the mid-1980's we used very small stripline traces at close to a GHz to handle thousand-watt bursts for radars.

Garth, I guess you're right. A minimal SBC will be about a 2" spread. I was thinking back to older tech LS TTL and about 15 IC's total including old EPROMs and current hungry small/fast asynchronous RAM. In addition, since the design was expanding, the powers and grounds wound up being 'daisy-chained'.
I just wanted to prevent a bad practice from the getgo, since power and grounds are not easily undone! Especially when one considers the WW bypass caps.

[GARTHWILSON]

Quote:

Garth, I guess you're right. A minimal SBC will be about a 2" spread. I was thinking back to older tech LS TTL and about 15 IC's total including old EPROMs and current hungry small/fast asynchronous RAM. In addition, since the design was expanding, the powers and grounds wound up being 'daisy-chained'.

I remember a friend's first computer, a kit he assembled in the late 1970's.  He got the option for a 1K or 4K RAM—I don't remember which—and it gave off a lot of heat! I do remember TTL boards that took amps.

[BigDumbDinosaur]

GARTHWILSON wrote:

I do remember TTL boards that took amps.

Yeppers. The first ZMT that I worked on in 1970 had a linear power supply that produced a continuous 50 amps. It took a whole lot of 74xx silicon in those days to get anything remotely resembling a computer.

[Tor]

From time to time I walk down to the storage area here at work to marvel at the power supply unit we kept from an old minicomputer. It's not that physically big, but it provided 300 amps at 5 volts.

-Tor

[BigDumbDinosaur]

Tor wrote:

From time to time I walk down to the storage area here at work to marvel at the power supply unit we kept from an old minicomputer. It's not that physically big, but it provided 300 amps at 5 volts.

-Tor

That thing could simultaneously power over 1000 of my POC units.

[GARTHWILSON]

GARTHWILSON wrote:

The amount of resistance in a 2" 30-guage connection is only about 0.017 ohms, not enough to matter at all in this kind of application.

I just calculated something I've been kind of avoiding for a long time, which is skin effect.  Again, relief!  That 2" piece of 30AWG wire still has less than 0.1Ω of resistance resulting from skin effect at 50MHz.  (50MHz is the 9th harmonic of 5.5MHz, still a rather important part of getting a decent square wave at 5.5MHz.)  The resistance from skin effect will go up approximately proportional to the square root of the frequency, meaning that if it's 0.1Ω at 50MHz, it will be 0.32Ω at 500MHz.

As an interesting side note, as I was re-researching this, I found that there are golden-ears people (whom I've ridiculed before) who go for fat speaker cables partly to reduce the high-frequency attenuation from skin effect.  It didn't occur to them that the speaker coils are wound with really small wire, and it's much longer than the speaker cable in their living room or studio.  IOW, it's like trying to use a Q-tip to get the moisture out of a sinking ship!    It turns out that even if the speaker were a perfect (8+j0)Ω at all frequencies (which it's nowhere near in reality), the difference in attenuation between 20Hz and 20kHz due to skin effect in a 12-guage 10-foot speaker wire would be less than 0.1dB.

[ElEctric_EyE]

In my audio systems, I go for low AWG speaker cable especially if they're 10+ feet away from the amp. I do this for the ability of the high current amp to instantaneously drive a current load and control it, especially with a lower ohm speaker load, i.e. <=4 ohms. Nothing to do with frequency response in audio that I was aware of...

[GARTHWILSON]

In car stereos, they've gone for 4Ω and apparently even 2Ω speakers, plus bridged outputs, because the power supply voltage is so limited.  In those cases, it makes sense to have fatter wire.  When I was in school though, and we were not talking about car stereos but rather professional audio equipment where the speakers were 8Ω, I remember a guest speaker who, in his talk, mentioned "damping factor" and trying to reduce speaker resonances electrically by using fat cable.  What he was forgetting was again that the electrical resistance in the speaker coils themselves (plus the crossover's coils and capacitors) dwarfed the resistance of the speaker cable.

[ElEctric_EyE]

There is so much in your last few responses Garth regarding audio power/frequency distribution, it's probably very pertinent to this thread, so I'll continue questions regarding audio.

GARTHWILSON wrote:

In car stereos, they've gone for 4Ω and apparently even 2Ω speakers...

these lower impedance drivers usually have a higher SPL efficiency don't they?

And just FYI, I don't care for car audio. I focus on driving, and prefer 7.1 home audio.
My ultimate system would be independent amp for every speaker driver independently/digitally 'equalized', instead of passive crossovers.

[GARTHWILSON]

Yeah, it's pretty off-topic.  The topic had kind of died down, but anyone is welcome to go back to talking about wire-wrapping discrete components at any time of course.

ElEctric_EyE wrote:

these lower impedance drivers usually have a higher SPL efficiency don't they?

I haven't seen figures on SPL at 1m for 1W input, but I don't really expect a correlation.  More SPL for the voltage, yes, but the lower impedance takes more current to get there.  1W into 8Ω takes 2.83V, 354mA.  1W into 2Ω takes 1.414V, 707mA.

Quote:

My ultimate system would be independent amp for every speaker driver independently/digitally 'equalized', instead of passive crossovers.

It has been done, and may be common practice in some settings, although I have not been keeping up in the field.  Doing it digitally should enable correcting the phase problems that any analog crossover gives.  There is also the horrendous phase response of speaker elements themselves which could be corrected this way too, although you'd probably want to do something like a convolution integral and put a calibrated reference mic in front of the speaker to keep taking new coefficients to continually update the function with, because the speaker's response to both amplitude and phase versus frequency will change a lot with a mild change in temperature.  I've seen this in earphones when I've done analysis on them for doing active noise cancelling since phase response is key there, and, even though high-quality earphones are vastly better behaved than speakers are, they were still atrocious, and changed a lot as they would warm after you put the headset on.  Without DSP, the only way to make a stable analog feedback circuit was just to be careful with where the poles and zeroes of the function lie, accepting that the active noise reduction will really only be effective in the low frequencies (which works out ok, because it's the lows that don't get sealed out well by the ear cup and ear cushion).

For systems other than active noise-cancelling, the phase response is not so important as some would make it out to be.  Without realizing what the crossover is doing, or how the individual speakers' phase response is, I've seen people go to great lengths to make the speakers coaxial, with the right front-to-back spacing to make them supposedly "time-aligned" for a listener standing directly in front of them.  Without even having done a back-of-the-napkin calculation though, I expect the processing power, and maybe the delay needed (depending on method), to get it right in DSP would be impressive.  Fun stuff.