Ian Iveson wrote:

Phil wrote

...toroids which *must* be quite labour intensive

Why?

** More steps are involved in winding a toroidal - plus their manufacture
uses a special "shuttle" winding machine that can do only one tranny at a
time.

That doesn't have much bearing on how labour-intensive they are. You could say
much the same about many products made entirely by automated machinery involving
no hand work.

There is an anomoly in the common perception of China, incidentally. On the one
hand, it is said that they depend on cheap labour, and on the other that they
are likely to baulk at the prospect of making toroids because they are too
difficult. However, obviously if those difficulties can be overcome by intensive
labour, then one might expect the Chinese to favour toroids.

Another common assumption, and one that Patrick has made several times, is that
winding a toroid requires a shuttle. Not true...fortunately, because then the
machine really would be complicated.

The shuttle I use to hand wind a toroidal is a piece of broom handle about 300mm
long
with deep notches at each end so that the wire for a winding can be loaded onto the
shuttlle so that one turn around the shuttle = 1 turn around the core. The wire for
a given winding must be preloaded onto the shuttelt with wire to spare.

But a toroidal winding machine has a circular shuttle in two halves which is clipped
together through toroid hole
and rotated on rollers to load it up before winding the wire back onto the core.
It is labour intensive and slow, but 20 times faster than the broom handle shuttle.

The toroid tranny method whatever it is isn't good where one wants to have an OPT
with say 4 primary windings for
the anode primary CFB primary ( 8 wire ends ), and say up to 24 winding ends for 12
windings which allow wasteless
low loss secondary configs for different impedance matches.

The E&I allow the winding ends to just come out each side of the wound bobbin.
OK, the leakage inductance is lower on toroidals than on E&I but then that is
because the
"traverse width" of the winding is the ccentre line distance around the core rather
than the distance across the bobbin. Capacitance may be higher with a toroid.

In any case i can wind E&I OPT very effectively with bandwidth at full power without
NFB using triodes
from 16Hz to 100kHz + very much easier than a damn toroid.



The hard parts about making a toroid are clamping the workpiece, and traversing
the core with the bobbin. The machinery is more complex and expensive, and both
the machinery and the process are less versatile. Hence automation can only be
justified by large production runs of the same transformer.

Chinese makers of transformer winding machines spammed me a couple of years ago but
the cheapest
one for E&I was forsale for aud $2,500 from an agent guy.
I said I built mine for 1 week's work and $200.
I never heard from them again.
The toroidal machines were dearer.


Custom-wound toroids don't seem to be commonly offered. If the only problem was
the need for intensive labour, then they would be the perfect candidate for
boutique production. Actually they are just hard to make, no matter how you make
them.

My attempts to get custom toroidals wound resulted in a high cost item unavailable
until months had passed
so I have given up trying to get anyone to wind my designs.
When they see what I want they go pale and vomit and I have to retreat.
I tell them I have wound all this stuff and they just don't wanna know.

Most consumer electronics tranny winding in Oz is 99.99% large number runs of mains
trannies operating at 1.2Tesla, and random wound coils, high noise, temperature, and
none I know possess or use a vac chamber and oven for varnish
impregnation.

There used to be skilled tradesmen and many women who wound spendid layer wound E&I
trannies but
not much now.
The imports of cheap nasty chinese crap is slowly but surely forcing the Oz tranny
winders to give up.
The labour in china is 26dB cheaper than Oz labour, so the chinese can gurrantee
their product for
5 years because if 10% failiures occur the replacements by the supplier doesn't
matter.



R-cores have so many practical advantages it is a surprise how few you see
being used.

Yes, and come to think of it, why haven't they *always* been used? What are the
problems?

They are still a lot more exensive than an E&I trannie.
The E&I are wound with 2 concentric bobbins, primary on one and any assortment of
secondaries on the other,
and large numbers of bobbins can be wound at the same time.
The cores are high grade GOSS with max µ = 17,000 if they were fully interleaved,
but they are
not interlaved at all and a block of E and block of I are placed around the
assembled bobbins by machine and
welded with machine welding machines. The resulting µ of the material is
sufficiently high to
give low iron losses. This production method is very cheap compared to anything
else.

So regardless of how much lighter or better an R core or toroidal might be, the
common denominator
price dominates the market so that 90% of all electronics looks like it has been
made at the same factory
with the same automation and what little labour is engaged is paid peanuts.



winding strip into anything other than a toroid is fundamentally problematic.
How are the strips kept squashed together in the straight runs, and how does the
core keep its shape during and after the winding of the strip? Presumably it
must be annealed after winding and clamping, or bonded somehow. There is also
the need for precision-cutting of strip to a shape peculiar to each core size,
without leaving a burr on the edge. I wonder if that is done before or after the
insulating coating? That's quite a lot of tricky processing dedicated to each
core size.

Ah, the wonders of the mechanised machines in factories have to be seen to be
believed....

Patrick Turner.




Maybe that is about to change.

My recently acquired ( Asian made) CRO has one for example.

http://www.filespoint.com/point/2689...00578.JPG.html


Is that a strip-wound core?

cheers, Ian