[Reverie]

Hey all,

I've been contemplating what wire to wind my Metglas toroid Inductor cores with for a Tripath project, for some time. I am after AWG16 wire for this design (yes, very thick, but these are large cores).

Just as I thought I had found the ideal source and product, someone pointed out to me that 'magnet wire' insulation must be very thin, otherwise even the slightest bit of air between the wire and core will ruin it.

So basically, my questions to you are thus:

1. Just how thin is acceptable, and how much difference does it make? Enamelled wire must have some thickness, even if it is very thin. I have seen some silk/cotton AWG16 silver, with insulation 1-1.5mm thick, is this any good?

2. Secondly, does anyone know of any source for high quality AWG16 copper (hopefully OFC/OCC) or silver (again, high purity preferred) with a suitable insulation?


I could always get some AWG16 wire of a preference with another insulation, and then strip it, and enamel it myself somehow?

Thanks for any suggestions or comments, I'm very grateful 

Peter.

[shep]

http://www.ahfartaudio.com/
very reliable people. Don't know if this might be what you are looking for though.

[Steve Eddy]

1. Just how thin is acceptable, and how much difference does it make? Enamelled wire must have some thickness, even if it is very thin. I have seen some silk/cotton AWG16 silver, with insulation 1-1.5mm thick, is this any good?

Nah, that's way too thick. Typical insulation thickness on 16 gauge enamaled wire is only about 0.0016" thick (about 0.04mm).

2. Secondly, does anyone know of any source for high quality AWG16 copper (hopefully OFC/OCC) or silver (again, high purity preferred) with a suitable insulation?

Not aware of anyone who has anything as large as 16 gauge other than regular ETP coper. www.wires.co.uk shows they have 10 meters (a little over 30 feet) of some 1mm 99.99% silver enameled wire, which is about the equivalent of 18 gauge.

se

[Gordy]

http://www.parawire.com/magnet_wire2.html

[Steve Eddy]

http://www.parawire.com/magnet_wire2.html

Yeah, there's any number of places you can get 16 gauge magnet wire, but it's just made with regular ETP copper. The original poster was wanting something a little more "exotic," like OFC or OCC or high purity silver.

se

[shep]

I would ask F...T (see link) if they can source smaller gauge OCC. They are very helfull.

[Reverie]

I would ask F...T (see link) if they can source smaller gauge OCC. They are very helfull.

'Hi Peter,

I know him, he is from france, intersting guy. I do not offer enamel wire, all I have are either PVC sleeve or teflon sleeve. , more
for speaker cable, interconnect, or internal wiring.


robert'

Good idea, but sadly no.

I'm going to start to look for people who'll enamel wire for me...

[Steve Eddy]

I'm going to start to look for people who'll enamel wire for me...

Unless you're wanting to have the world's most expensive toroid transformer, I don't think you're going to have much luck there. I looked into that some years ago and wasn't able to find anyone interested in firing up their huge machines unless I was wanting to run 100 pounds of wire. And even if you find someone willing to do a smaller run, it'll be bloody expensive just to make a couple toroids for your amp.

Good luck.

se

[Reverie]

Yep, that's looking like the most probable conclusion of this problem.

So, it leaves me with 2 compromises:

Either:
A) a core with high quality wire and good yet thick insulation.

[ Worth noting what Duelund do on their inductors:



-silk and cotton permeated with natural plant resins.

Although admittedly, their scale is considerably larger than mine, so gaps between wire and core may count for less. ]


or
B) a core with lower quality wire and poorer yet thin insulation.

[ What usually is done it seems. ]

[Steve Eddy]

Yeah, I see your dilemma. Though while technically the thinnest insulation possible would be the ideal, at the end of the day what counts is how it sounds and a thicker insulation of natural materials may sound quite good and perhaps even better than something more technically ideal. Wouldn't be the first time. And hey, if nothing else, your winding capacitance would be less, all else being equal.

If you don't want to pay $600 to have some specialty wire wrapped with silk, I have some pure filament silk braid (which I use for jacketing on my cables) that would work if you opt for option A. Though not sure you could get 16 gauge through it. I don't have any 16 gauge on hand, but 18 gauge is a bit of a tight but doable fit. Measured on 18 gauge wire, it has a wall thickness of 0.5mm which is rather better than the 1-1.5mm you'd mentioned originally.

You sure you want to use 16 gauge solid core? The larger the gauge the larger the skin effect losses which will be much more significant at the switching frequencies of the amp. I've a smaller silk braid that should work for 22 gauge solid core. Don't have any 22 gauge but it's a bit loose on 24 gauge.

I see you're in the UK. I get the braid from the manufacturer in France. Drop me an EMail or a PM and I'll give you the details and you can see if you can get it from them directly.

se

[Reverie]

Thanks for your offer Steve Eddy! I'll PM you soon.

As for the gauge of wire, that is a very interesting point you bring up. I was under the impression that the thicker the wire the better, as the less resistance and general wire material there is. I made my design with the help of the Metglas design tool. Here's all the data I can offer:



The program automatically chooses the wire gauge; I assume based on current settings (5-6A max, ['~2.5A continuous' although it didn't ask for this] in this case).

In designs the support staff made for me, the inductor's inductance changed considerably between full and no load (15uH-20uH in some cases), what I liked about this larger core design was the fact that it kept a constant inductance much better (15uH-15.2uH). If you think it better, however, I will switch to a small gauge, and use an inductance metre to find the correct number of turns. 

Thanks for your help. 

[Steve Eddy]

I'm no expert on inductor design but if those who do know about this recommend 16 gauge for your needs, I'd tend to stick to that.

You could retain the 16 gauge and address the skin effect issues by using litz wire. But then you're rather back where you started with the wire quality vs. insulation quality/thickness dilemma.

Cardas sells chassis wire in a variety of gauges ranging from 9.5 to 23.5 which is a litz type construction. Perhaps you could use some of this, stripping off the Teflon jacket.

se

[jb]

I’d be wary of using Cardas wire in an inductor. Although it is very high quality copper, drawn in an oxygen-free atmosphere, oven annealed, etc., it contains multiple strands of different gauge wire, each individually insulated with a polyurethane varnish, where the fatter strands are wound around the thinner ones. This results in different electrical path lengths between the different strands because the portion of the signal conducted by the thinner strands travels a more or less straight path down the center of the composite wire while the portion conducted by the fatter strands follows a spiral path around the periphery.

George Cardas’ wire construction method is similar Bruce Brisson’s method. Brisson also used different gauge strands but the thinner ones are wound around the fatter ones. Although both methods are patented and claim to do wonderful things for audio, both are contrary to normal litz construction where the individual strands are all the same gauge and, whether woven or rope-lay twisted, the strands have identical path lengths. I think an audio signal traveling different paths to the same destination, whether in interconnects, speaker cables, or an inductor, produces audible smearing. And, as a practical matter, it’s very difficult to strip the Teflon insulation from a length of Cardas hookup wire without nicking the individually insulated strands. (I’ve tried.)

[Reverie]

I'm no expert on inductor design but if those who do know about this recommend 16 gauge for your needs, I'd tend to stick to that.

You could retain the 16 gauge and address the skin effect issues by using litz wire. But then you're rather back where you started with the wire quality vs. insulation quality/thickness dilemma.

Cardas sells chassis wire in a variety of gauges ranging from 9.5 to 23.5 which is a litz type construction. Perhaps you could use some of this, stripping off the Teflon jacket.

se

Well, I wouldn't say that the Metglas program really knows what I'm doing, but I did speak to Micheal Mardis about the idea of Litz wire on the core, and his view was:

"Why use Litz Wire on the inductors?
Litz wire is normally use in RF to carry high frequncies better than conventional wire (it's very flexible, too).

But the reason the inductors are in the output circuit is to REDUCE/REMOVE the RF. So you want to send the signal thru a low pass filter to remove the RF, but use Litz wire to increase the high frequency transmission? Doesn't make sense to me.  "

Also, many (well, at least 1 I know of) comercial TK2050 amps use solid core inductors with very successful results. So, for those reasons, I was planning on keeping with solid core.

If however, I can't find any way of getting AWG16 wire in a suitable form, I may drop down to AWG18. I am also uncertain as to the impracticalities of AWG16 yet too.

Whatever the case, thanks for all your help, I'll PM you back soon!

Peter

[jb]

"Why use Litz Wire on the inductors?"

Why not? Jan, of 41Hz.com, who sells Tripath amp kits, now supplies Litz wire with some of his kits.

AMP2 kits are now shipped with:
- Litz wire, 15 strand wire, for toroids. Litz wire is a thick wire made up of thin individually insulated strands. Compared to single strand wire, it has lower loss at high frequencies (which is good for switching output amplifiers) and is easier to wind. It is especially used when thick wire is required, as the "skin effect" gradually becomes more important.

Previously the builder had to supply the wire for the coils.

NOT included in the kits:
Wire for the toroids. About 3 meters (9 feet) of enameled 1.2 mm copper (16AWG) wire is needed (1.5 meter per toroid). However this thick wire is a challenge to wind so 3 meters of insulated "Litz", multi-strand wire of equivalent area is preferable.

Since you are winding your own coils you could try different kinds of wire. Additional cores are not that expensive and you could use some really exotic wire, like Isonel/Platinum litz.

[Steve Eddy]

Well, I wouldn't say that the Metglas program really knows what I'm doing, but I did speak to Micheal Mardis about the idea of Litz wire on the core, and his view was:

"Why use Litz Wire on the inductors?
Litz wire is normally use in RF to carry high frequncies better than conventional wire (it's very flexible, too).

But the reason the inductors are in the output circuit is to REDUCE/REMOVE the RF. So you want to send the signal thru a low pass filter to remove the RF, but use Litz wire to increase the high frequency transmission? Doesn't make sense to me.  "

Well, off the top of my head I would say that the inreasing resistance of the wire at high frequency would reduce the effacacy of the inductor. An LC filter would be more effective than an RC filter.

Also, many (well, at least 1 I know of) comercial TK2050 amps use solid core inductors with very successful results. So, for those reasons, I was planning on keeping with solid core.

It may be that litz wouldn't give you an appreciable improvement in sonics. But from a purely technical standpoint, litz would be the way to go. Though if the Cardas wire isn't a true litz as jb mentions above, then there wouldn't be any point in using it.

If however, I can't find any way of getting AWG16 wire in a suitable form, I may drop down to AWG18. I am also uncertain as to the impracticalities of AWG16 yet too.

Well it would certainly pretty damn stiff wire to try looping through a toroid. If nothing else, a fine wire litz would certainly make that task much easier. 

Though from your data, you're only needing something like 10 turns so that helps.

Whatever the case, thanks for all your help, I'll PM you back soon!

You're quite welcome.

se

[Steve Eddy]

Hey Peter, if you like Teflon, check these guys out:

Phoenix Wire, Inc.

se

[Reverie]

Hey Peter, if you like Teflon, check these guys out:

Phoenix Wire, Inc.

se

I've emailed them to see what they say .

Regrettably, they do say their speciality is a much smaller gauge of wire though.

[Reverie]

Why not? Jan, of 41Hz.com, who sells Tripath amp kits, now supplies Litz wire with some of his kits.

'- Litz wire, 15 strand wire, for toroids. Litz wire is a thick wire made up of thin individually insulated strands. Compared to single strand wire, it has lower loss at high frequencies (which is good for switching output amplifiers) and is easier to wind. It is especially used when thick wire is required, as the "skin effect" gradually becomes more important.'

Hmm, I guess the theory of wanting something that *doesn't* hold high frequency doesn't stand then?

Since you are winding your own coils you could try different kinds of wire. Additional cores are not that expensive and you could use some really exotic wire, like Isonel/Platinum litz.

I am becoming increasingly interested in this idea.

Where would you suggest I source some litz if you'd discourage me from the Cardas?

I'd be very interested to know what suppliers you'd suggest.

Thanks again,
Peter.

[jb]

I'd be very interested to know what suppliers you'd suggest.

I like Cooner Wire in Chatsworth, CA. There is also HM Wire, MWS Wire, New England Electric Wire, and others but they are all big suppliers and deal in custom builds of thousands of feet. You might be able to sweet talk one of them to give you a sample if they have any suitable wire in stock.

Because you only need a few feet, I suggest you make your own by twisting together several strands of magnet wire. For example, 9 strands of 25.5 AWG magnet wire will give you 16 AWG. I think Percy Audio still has some Ohno Continuous-Cast copper magnet wire in stock. Chimera Labs used to sell the same wire cryoed for a reasonable price. That would give you the wire quality you are seeking. Otherwise, there are numerous sources for ordinary magnet wire, including Radio Shack.