Maker Pro
Maker Pro

My ferrite transformer is bigger than yours... eh well...

T

Tim Williams

Jan 1, 1970
0
Well, maybe I *shouldn't* say that here, because I know some people have
rather large ferrite transformers... Still, it's big as home built projects
go.

http://webpages.charter.net/dawill/Images/Induction1002.jpg

That's 10 turns of 10AWG wire (mil spec!) doubled up, wrapped around four 3"
toroids (Magnetics Inc. OW48613TC), with a 3/8" pipe down the middle.

Works great already. I've only tested it a little, and I've already got way
more power output than ever before. 1200W heats up steel pipe pretty fast!

Here's an example inverter output waveform:
http://webpages.charter.net/dawill/Images/Induction1004.jpg
Leakage inductance is much lower: there is some, and it's noticable (this
oscillograph is at 33 kHz and still far from resonance, while the tank is
supposed to resonate at 35), but it's nowhere near as bad as my earlier dual
C core test.

BTW, I learned one possibility why uncoated ferrites suck, and one reason
why uncoated conductors suck more. ;o)

Tim
 
E

Eeyore

Jan 1, 1970
0
Tim said:
BTW, I learned one possibility why uncoated ferrites suck,

Aside from material inherent bulk resistance ( which is normally high enough not
to be a problem ), what was that ?

and one reason why uncoated conductors suck more. ;o)

Yes, that would be the case in a transformer.
 
T

Tim Williams

Jan 1, 1970
0
Eeyore said:
Aside from material inherent bulk resistance ( which is normally high
enough
not to be a problem ), what was that ?

The reason is a combination. I had wound eight turns of bare copper strap,
some of which rubbed off on the bare cores. Applied power and I got a
hearty spray of orange sparks between the start and end turns -- I guess
there was enough of a copper track between them to short.

If either had been insulated, this wouldn't have happened, of course. I
don't really have a good way to insulate strap... could spray it with
urethane or something I guess. Not great for corners though.. maybe wrap it
with something?

The present wire, which is insulated with "mineral filled teflon" or
something like that, of course works quite nicely.

Tim
 
J

James Arthur

Jan 1, 1970
0
Phil said:
Woven fibreglass tubing?

Cheers

Phil Hobbs

At the other end, ferrite cores are often epoxy-coated
for insulation & winding protection. A Dremel with a
diamond wheel knocks the sharp corners off cores, too.

Tim, I saw a pallet of ferrite toroids at All Electronics
not too long ago--they were almost dinner plates, but...
your dinner would fall through the hole.

New invention: dieter's toroidal ferrite plate. Reduces calories
AND shields delicate organic foods from harmful cell phone
RADIATION. Available at audio stores everywhere.

Cheers,
James Arthur
 
T

TheQuickBrownFox

Jan 1, 1970
0
Well, maybe I *shouldn't* say that here, because I know some people have
rather large ferrite transformers... Still, it's big as home built projects
go.

http://webpages.charter.net/dawill/Images/Induction1002.jpg

That's 10 turns of 10AWG wire (mil spec!) doubled up, wrapped around four 3"
toroids (Magnetics Inc. OW48613TC), with a 3/8" pipe down the middle.

Works great already. I've only tested it a little, and I've already got way
more power output than ever before. 1200W heats up steel pipe pretty fast!

Count the number of wires (bifilar passes) that pass through the core.
Those are turns. I do not mean each wire as I know that you doubled it
up. I mean the number of times that twisted conductor passes through the
core, including the lead in/out. That count is your turns count.
Regardless of how many times you counted when you made it, on a toroid,
each pass is "a turn" in transformer talk.

Have the copper silver plated, and run it at high frequency.
Here's an example inverter output waveform:
http://webpages.charter.net/dawill/Images/Induction1004.jpg
Leakage inductance is much lower: there is some, and it's noticable (this
oscillograph is at 33 kHz and still far from resonance, while the tank is
supposed to resonate at 35), but it's nowhere near as bad as my earlier dual
C core test.

I'll bet it likes 56 - 59 kHz.
BTW, I learned one possibility why uncoated ferrites suck, and one reason
why uncoated conductors suck more. ;o)

Coated or not, they suck for HV. A corona forms on the core. Don't get
me wrong, we used them all the time, but I did discover that some
miniature HV transformers have some strange effects that would only be
huge on a larger scale unit. Larger stuff is usually potted or immersed
though.
 
R

Richard Cranium

Jan 1, 1970
0
Count the number of wires (bifilar passes) that pass through the core.
Those are turns. I do not mean each wire as I know that you doubled it
up. I mean the number of times that twisted conductor passes through the
core, including the lead in/out. That count is your turns count.
Regardless of how many times you counted when you made it, on a toroid,
each pass is "a turn" in transformer talk.

Have the copper silver plated, and run it at high frequency.

I'll bet it likes 56 - 59 kHz.

Coated or not, they suck for HV. A corona forms on the core. Don't get
me wrong, we used them all the time, but I did discover that some
miniature HV transformers have some strange effects that would only be
huge on a larger scale unit. Larger stuff is usually potted or immersed
though.


Hi Archie:

Are you ready to attempt the puzzle yet?
 
T

TheQuickBrownFox

Jan 1, 1970
0
Yes? I know what turns are, and there are ten of them as I stated.


That would be difficult, considering the primary shown is coated in
PTFE. (Unfortunately, it's already plated with nickel, which probably
isn't great for skin effect, and definitely sucks for soldering.) The
other hardware isn't important enough to matter, and a mere plated
layer will have no effect under 10MHz. The skin depth is
approximately the entire tubing (the 1/4" copper seems to be 0.02"
thick).


Seems to work fine with flyback transformers.

Funny, I wound my own little FBT the other day. 10T of ~20AWG litz
followed by 500T of #37, on a small E-I core about 1/4" square cross
section. Got it up to 3-4kV peak, before the filament winding started
arcing through (I don't have any HV silicon, so I used a toob
diode!). That's the fault of the 300V hookup wire standing off 6kV
peak, since I didn't have space for a better insulated wire. That
37AWG wire gives a hearty sized arc, I must say.

When your turns count on your secondary climbs past 150 turns, you have
to be mindful of proximity between additional turns, and the turns down
near the lead in. Count up your volts per turn, and you can define how
many turns are safe to be near each other with single strength mag wire.

Once you are above that number, you either want to add transformer
tape, and start a new layer, OR use a segmented bobbin.

When you get up around 1400 turns of #53, proximities are very
important, especially pre-impregnation, if any. The coating on that wire
is thin, and yes, they spit highly excited electrons out pretty darn
good! I was making 15kV from 3V in (that control circuit was a feat in
and of itself). I think the transformer was making about 2500 Volts into
the multiplier. It was a switcher though, not a flyback.

This is also why one will never see a Tesla coil that has a layered
secondary. The second layer turns are several tens of thousands of volts
separated from the first layer. Far too much for the insulation of the
mag wire. It will fail every time.

We had miniature transformers that had 7 layers of #43 on it. Each
layer was kept inboard from the bobbin faces a mm or so. It would be a
flat single layer of turns, then a layer of transformer tape, then a new
layer of turns. The layers had to be kept inboard so that the tape would
seal between layers. It also allowed impregnation media to get in there.

Every time we wound with the turns all the way at the bobbin faces,
there would be a turn to turn short eminent.

We had stuff that if you breathed "coffee breath" on the multiplier
caps before potting it, it would fail.
 
T

Tim Williams

Jan 1, 1970
0
TheQuickBrownFox said:
Once you are above that number, you either want to add transformer
tape, and start a new layer, OR use a segmented bobbin.

Bank winding has other advantages, too. Take this for example,
http://webpages.charter.net/dawill/tmoranwms/Elec_Compound10.jpg
About 20mH total inductance (all six in series), while having only about
80pF parasitic capacitance. Low enough to get a plate waveform like so:
http://webpages.charter.net/dawill/tmoranwms/Elec_Compound12.jpg
That's a 200ns rise/fall time, and yes, I said plate waveform. This
inductor belongs to perhaps the world's first class D tube amplifier.

Tim
 
Top