Twisting transformer wires?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I wanted to piggy back off of Ransom Peek becuase of Poobah's post about x-raying the magnetic fields in a twisted input cable.

Does the same apply for the transformer primary/secondaries.

I just twisted those wires in my current amp (for looks to be honest) but did not realize that it may have real impact in the design. Does it? Should those wires not be twisted or does it not matter here?

Thanks,
Dominick
 
Twisting your primary wires is good... because it prevents a singular coherent field from radiating from the wires... same for the secondaries.

Twisting DC wires is good because it prevents a moving field from injecting something in.

All this twisting stuff really only works right when: The sum of all currents flowing in a group of twisted wires equals zero. That's a confusing way of saying; if there is 1 Amp going THIS WAY, there should be 1 Amp (the same amps BTW) going the OTHER WAY.

What about a center tapped secondary? If you are taking all three wires, center and legs, to the same location, just twist all three together. Also for this application you can use triaxial wire. Ping member jneutron for details... this is the hot setup.

If for some weird reason you are taking the each outer leg and a center tap to different locations, this can be confusing. So, as close to the trans as possible cut the center tap and splice on 2 wires. Twist one center tap lead with one outer leg, lather, rinse and repeat for the other side, and you're good.

;)
 
Keep in mind,

All this twisting stuff may or may not make a difference. It's not going to "open up your midrange" etc... It's all about keeping AC hum and other hash out of your signals.

The importance of this stuff increases as your signals decrease (or impedances increase)... so for things like phono & microphone signals... it becomes important... as does shielding.

Either way, harming something by twisting is next to impossible as long as you follow the sums of currents rule.

;)
 
poobah said:
What about a center tapped secondary? If you are taking all three wires, center and legs, to the same location, just twist all three together. Also for this application you can use triaxial wire. Ping member jneutron for details... this is the hot setup.
;)


Hey there.

I'm still awaiting the results of my triaxial construct. I sent it to Peter Daniel, he got it May 9th.


It pays to twist all wires which carry current to and back from a block, as they will broadcast mag fields. For low Z circuits, this is far more important to watch out for than the E fields.

Cheers, John
 
  • Like
Reactions: 1 user
I believe I remember a passage in the Jones book where he takes solid copper wire for the filaments and clamps one end of the the two wires in a vice and the other end of the two strands into the chuck of a drill and power on! Tight and clean, as it pulls its way toward the vice.

Maybe the key word here is filament wire, where it needs to be tight and tucked up along the outer perimeter of the chassis...
 
I have just received my 625VA 25-0-0-25 tranny for my project.
The primaries are thin enough to twist but each of the secondaries consists of 2 heavy gauge very stiff wire twisted together to create the one flying lead. These are very stiff almost rigid. Should I attempt to further twist them together - I'm thinking it would be very difficult.
 
mrlots2do,

I'll be perfectly honest with you. Other than using more wire and adding resistance, I don't know.

I started working in Electronics in the Navy in 63 and I was always told to make loose twist as apposed to tight. Over the years, I just got to the point that very tight twists just look like crap.

Take the jacket off a power cord or any audio cable. The twists are not tight. Even Cat 5 is not a tight twist,

I know there is a real reason but not sure what.

Just a thought from a guy who had his head smacked by seniors for making tight pretty twist.

PS I can't believe they sensored cr_p!
 
sbrads said:
Tight twist = longer wires = bad.
Loose twist = almost same length as untwisted = good.

Given a 16 guage pair, as an example...5.15 milliohms per foot, 10.3 milliohms per foot for the twisted pair...

If you take a 10 foot run and twist it until it is 9 feet long, you have increased the per foot resistance by 10%, or 1.03 milliohms per foot. If you twist it to 8 feet, you've increased it 2 milliohms per foot.

As a lumped parameter, this resistance change level is insignificant, not too many people can even measure the difference, and the other resistive elements within the circuit will wash the difference out.

The purpose of twisting is twofold:

1. Visual and wire dress: making the package neat and tidy, stiffining the pair to lower vibration.

2. Reduction of inductance. Inductance shows itself in two ways..

2A...lumped element effect. For a power transformer, this is again, swamped out by both the parasitic of the transformer, and the massive inductance of the structure. For a chipamp, this can cause oscillation without local bypass if a twisted pair is used to run the supply rails to the chip.

2B. Field generation and interception. Inductance, by definition, is the amount of energy stored in the magnetic field as a result of the current through the system. The tighter the twist, the less external field created, hence lower inductance. The looser the twist, the more external field with higher inductance.

This inductance also couples external fields into the wires. The more inductance in the wire pair, the more the pair will intercept external fields and create a voltage within the wire pair as a result.

Attached is a graph for 16 guage wire pair. Note the rapid increase of inductance as the spacing increases off the zero spacing point. The best twist will force the tightest spacing.

Also make sure that parallel twisted pairs have different twist pitches..if the have the same, they will couple to beat the band..


Cheers, John
PS...if you use the drill technique, when you let it go, it will loosen up. To get the best twist, you have to use two people, one on the drill, the other holding the wires. As the drill is turning, the wire holder allows the wires to rotate individually within in each hand, this prevents twisting each conductor, and they will not be fighting to untwist the pair.
 

Attachments

  • 16 guage inductance vs spacing.jpg
    16 guage inductance vs spacing.jpg
    66.6 KB · Views: 321
There is yet another reason for twisting wires... flexibility. Works the same way a rope does when you have a bundle or group of wires. Jacket cables are made this way so they will flex/bend when jacketed.

John is right (99.9999999%), tighter is better. But you don't have to go nuts.

Oh and no bright ideas here... DO NO put wires in a vise and twist with a drill. That might be OK for some solid core wire under professionally controlled circumstances... but generally uncool because you are literally twisting the copper. This is bad because in the case of stranded wires, you can actaully remove theis twist the factory put in. Copper LOVES to cooperate, but the old guys had a point... don't stress the buhjesus outta' the material.

When we twist, we mean, wrap in a spiral (helix). Twist your wires a few turns... then "comb" the remaining untwisted wire out straight, repeat, repeat, repeat. If you are doing it right, the twisted pair will just lie there when you let go. If you are doing it wrong, they will try to unwind when you release them.

Whew..................... I'm going to author a book on wire twisting. John, you care to "co-" on this? I smell huge $$.

:cool:
 
poobah said:
John is right (99.9999999%), tighter is better.

1 part in 109? Darn, guess I gotta work on it..;)

poobah said:
Whew..................... I'm going to author a book on wire twisting. John, you care to "co-" on this? I smell huge $$.
:cool:
No money in books now..

There is, however, in wire twisting...:D

Cheers, John
 

Attachments

  • twist fixture.jpg
    twist fixture.jpg
    82.1 KB · Views: 330
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.