Re: Toroids for power transformers



Ian Iveson wrote:

Patrick Turner wrote:

Just forget the operation at the rated power and consider
the situation
with NO LOAD.

***Bad idea*** , if you are at all inclined to leap to false
conclusions.

Have a look at the wave form of the **current** running
through the
transformer's inductance, ie, the magnetizing current.

If the transformer has much inductance and low hysterisis
behaviour,
you'll
maybe see a current wave not unlike the mains voltage
wave, although
there won't be so much HF content because the L obstructs
the HF.
No spikes, right?

Such a tranny might have GOSS cores, and B about
0.85Tesla.

Heat loss per Kg will be very low.

So in this case the designer would have used ample turns
per volt.

But as the turns per volt are reduced, the inductance
falls, Bmax rises
and more Im flows
and at some voltage level the current wave of magnetizing
current
will have distortion in excess of 3% which is visible as
kinks in the
up/downs of the wave. Spikes. Many poorer trannies with
low TPV and high
Bmax will show much more than
3%, maybe 15% or more harmonic distortion which shows the
iron is
becoming unable to create an opposing field
to the applied emf, or voltage propelled I flow.
The type of iron may still not get very hot, but many are
beginning to
get noisy.
As TPV are further reduced, the cos of phase x Im
indicates heat lost is
becoming high enough
cause the iron to heat up considerably, and it happens
withouty a load
connected.
The full load causes additional losses i the copper
further heating the
whole thing.

**No.**

A loaded transformer passes much greater primary current
than an unloaded one that is not saturating. This greater
current causes a voltage drop across the primary inductance,
resulting in lower magnetising current. The difference
between unloaded, and loaded as rated, will normally ensure
that the core will not saturate, even if it does when the
secondary is unloaded.

Obviously, the current will become worse with load if the tranny is
saturating.

But first you should examine the unloaded current wave to see
the quality of the design. Low Afe combined with low turns per volt give
high Bmax,
and the saturation shows up like dogs balls without having any load.

Perhaps you have not designed, wound and tested many transformers like I
have.

I thought not.


There is an important lesson here: be very careful if you
"over-specify", by buying a transformer bigger than you
need. Neither should you reckon on squeezing more than the
rated voltage from a toroid by under-loading it and taking
advantage of its regulation. If you underload, it may be
driven into the saturation region.

Just buying/ordering a higher VA tranny won't fix the problem of the
Bmax being too close to
saturation. The Bmax must remain low for quiet cool running.

If one buys a toroid with all the usual horrid problems most of them
have
then one may add 50 turns by hand around the outside, then re-insulate,
and this will increase the turns per volt, decrease the Bmax, but also
decrease
all the secondary output voltages. Its usually too difficult to
find a toroid which allows such sec reductions after adding primary
turns.
So I wind all my own trannies because the commercials are such poor
crap....

Toroids by their nature don't lose heat so easy, especially
from the core. That may be why they all use GOSS, BTW.

It is one reason why toroids are GOSS, but one could use NOSS,
and use a 400VA rated core size for where 200VA is needed, with no
change to turn numbers.



A poor transformer gets quite warm just sitting there
connected to mains
with no load.

***Bad idea*** It may get much less warm if properly loaded.

But a well designed tranny won't get hot any time, period.


Equipment which by its nature demands widely varying current
from the transformer should be careful to specify the
transformer to suit the application. "Standby" circuits
should run on an auxiliary transformer, and disconnect the
main transformer from the mains.

I often instal auxilliary mains trannies of a few VA to run fault
protection circuits and bias monitoring in tube amps.

But all my trannies don't get hot unloaded...

Neither do those of most reputable makers.

And the artful and despicable bean counter likes to have
about a 20C
rise in T because it means he's
saved enough copper and iron to buy that BMW he firqing
well wanted.

Manufacturers meet the requirements of application
designers. Accounting is a necessary part of that.

Its amazing how you defend the undefendable bean counters always trying
to maximise profits at the expense of quality.

Accounting has a function to count, but not to control production
methods.

Our
requirements may be different from the norm, and to that
extent we should not expect commodity transformers to meet
our needs. This is just as true of an EI as it is of a
toroid or anything else.

Who is "Our?"

All transformers should be good quality period.



Most GOSS don't have the poor performance as NOSS, but you
can easily
build a toroid or E&I tranny
with GOSS that WILL get scorching hot and you can't keep a
hand on it
because its well over 40C.

Most designers don't use too little GOSS in a toroid,
because the design
becomes far too noisy
before it gets too hot.

MANY toroidals I have seen and a number I have bought are
way too noisy
to use in hi-fi gear
even though they don't over heat. So although they say
GOSS can be run
at 1.3T which is approaching saturation
the noise is the problem, and the only way I have
addressed this is to
unwind the toroid by hand with a shuttle, and add 30%
more primary turns and then rewind the secondary over the
top.
The last two 300VA and 500VA took a couple of days each to
fix this way.

But they now run with only a few mA for Im, and with a
rectifier they
are tolerably quiet.
Still not as quiet as a well wound E&I.

The GOSS E&I I get now has maximum µ = 17,000 when fully
interleaved,
and losses are comparable to a spiral core of GOSS used in
a toroid.
With equal Afe for either toroid or E&I, ( or decent
C-cores ),
the weight of the toroid is only marginally lower than the
E&I.

In Oz, the problem of hopelessly inadequate toroid and E&I
trannies
entering Oz
in consumer items became so bad that all imports were
forced by govt
regulations to have a thermal fuse
mounted between the primary and secondary, so when the
item
got hot, it would go open even though there was not enough
current to
blow a mains fuse.

Sometimes one would be lucky, and able to put a link
across the leads of
the fuse after it blew,
and place a new fuse of lower T rating on the outside of
the sec.
But many makers concealed the thermal fuse to prevent the
tampering, and
you need to buy a new
tranny to repair the item.

Makers were constantly whittling down the copper and iron.

Planned obsolescence.

*Recognition of the inevitability of future progress.*

They want you back after it fails, and if you hate the
Yamaha when it
fails, you'll buy a Sony,
but their transformers are no better. The consumer has NO
ESCAPE.

My Sony CD player transformers are beautiful, and clearly
better for the purpose than anything I can buy off the shelf
from RS or Digikey. Closest would be something of Sowter
quality, with all the bells and whistles like copper band
and internal shields. Same with my cheap little Denon
CD/radio.


Yes, but its unlikely sources blow their transfromers.
Its mostly power amps or receivers which smoke their trannies.

Often because a channel has been labouring into a low load for too long,
or the ampo
is in a closed cupboard, and it just gets hot in there...

Anyway, a constant stream of mainly SS amps pass my bench for repairs
and some include those needing a new tranny.



Toroids usually fail because the windings get too hot, not
the core,

Is this different for EI? Considering how easy it is to cool
the core of an EI, I wouldn't have thought so.

Cool cores are larger than bean counters want them to be.

In E&I trannies, core losses are often arranged to be slightly
higher than copper losses, and because the wire is a bit cooler,
it matters not if the core gets hotter, as the core surrounds the wire.
However, the core blankets the wire passing through the windows,
so the T difference is negligible. Many designs have equal core and
copper losses
at rated working load of say 5% of the VA for each.

Toroids using GOSS nearly always have lower core losses.
The copper losses determines the temperature, and the core follows the
copper,
because the copper fully blankets the core.



and
because there
are crossed over turns all around the windings with no
insulation
between them,
the presssure of the mounting bolts and rubber squeeze the
hot wires
together an pinch through the
insulation in pressure spots. The shorted turn as a result
does the
rest.

Again, a matter of specification. You could just as well say
that an EI could fail because the bobbin has corners which
concentrate the pressure on windings.

The winding of many toroids I have seen is attrocious, with
very many crossed turns all piling up without insulation.

Some E&I are also "random wound", ie, not neatly layer wound, which is
far
more reliable than random or toroidal. The corner turns don't matter if
insulation is used
between all layers.
Pressure spots don't occur in E&I the way they occur in toroids unless
the toroid
is potted well after varnishing properly and thus is not subject to
dynamic stress due to a bolt
and a way too thin rubber washer.

Indeed they may fail
that way, but not very often as long as they do not stray
from their safe operating area. Manufacturers are of course
aware of the potential weak spots, and generally take
measures accordingly. Disreputable manufacturers don't last
very long in the market. It is up to all of us to make wise
purchasing decisions.

Hmm, the Chinese seem to remain in the market place as do many
other dodgy brands of gear I'd like to see hunted off.

One has to say a few prayers for the trannies if the country
they came from is Dodgymaya.

But in 1955, if you bought a Leak, and had a tube saturate, it often
wrecked
an OPT, then a PT as well...

Similar happened to many old brandname amps.

Bean counter designed crap has always been with us.

Patrick Turner.


Ian

I have often wound E&I trannies with copper ratings = 1.5
x actual
maximum draw,
or 2A/square mm, and cores of 3 times the VA rating
required.

T rise with GOSS is so low its negligible.
.