Output transformer cores

[Mario Pankov]

I was looking at transformer cores for the past few months as I am looking to start winding my own trafos. The M6 silicon steel has been a favorite to many people - it is cheap, not that hard to find and has good parameters. However, I was looking into something more exotic, no matter what the price would be.
What I found so far is a material called Metglas Magnetic Alloy 2714 which shows superb qualitues - it has the lowest core loss, permeability is many times higher than M6 so it should allow a much smaller core size, which means a reduction in the turns radius and thus less losses in the copper. This material also has a freq range up to 250Khz and is often used in aviation and medicine transformers. There are a few companies that offer C-core laminations of this material. To me it looks like the ideal core material, any thoughts?

 

[WBS]

C-core is the way to go if you've got the dough.

 

[smoking-amp]

Max flux density and power at the lowest audio frequency sets the core size, not the permeability. Metglas 2714 only has .57 Tesla saturation (annealed) and only .25 Tesla un-annealled. So this will require a 3 to 6 times larger core area than M6. (which translates to 5 to 14.7 times the core volume of M6)

 

[Bandersnatch]

hey-Hey!!!,
I'd go for 49% Ni stuff. I suspect I can get C-cores in the size I want; it is unavailable in the larger sizes of scrapless E-I. Near the saturation flux density of M6( or M4 and M3 for that matter ), higher permeability and lower loss than 3.5% Silicon steel. Nice stuff indeed, though I suspect it'll be $60-70 per pound... I'll be using it in a coil design set for a 2.5" stack of 1.5" center leg width scrapless E-I.
cheers,
Douglas

 

[Matt BH]

Hi Mario,
Could you let us all know where you can get the Metglas stuff. I have a pair of M6 c-cores but as two they are big enough for one transformer but I need two. I have been thinking about winding my own output trannys for a while, even built a lathe type lash up with a counter etc. The cores I have a are marginal after doing all the math. I really dont want to wind some trannys on the off chance. Even after measurement with my cores. I get 300W with the two at 50Hz with minimal distortion. So c-cores can be a fair bit smaler for push pull. Langford Smiths papers show this but there is very lttle info out there about it.
Cheers Matt.

 

[Mario Pankov]

There is an English company that supplies these materials. I do not have the pricing as I have not figured out what would be good as material.

I did a mistake when pasting the Metglas number, it is the 2605SA1 and I found a work by a Swedish aircraft electronics designer who said they are best for transformers used in the Grippen aircraft. What I am looking is this:

http://www.amorphousmetals.com/products/page5_1_6_2.htm

 

[alexberg]

There is an English company that supplies these materials. I do not have the pricing as I have not figured out what would be good as material.

I did a mistake when pasting the Metglas number, it is the 2605SA1 and I found a work by a Swedish aircraft electronics designer who said they are best for transformers used in the Grippen aircraft. What I am looking is this:

Powerlite C-Cores

What is the purpose of using low loss cores? Losses are extremely small - you will be getting a lot of unwanted resonances... unless you know for sure how to deal with it.
BTW
1.Package factor is less than 0.8.
2.I would not recommend flux density more than 1T.
In respect to iron cores amorph core will be at least two times bigger by the volume.

 

[Arnulf]

I found a work by a Swedish aircraft electronics designer who said they are best for transformers used in the Grippen aircraft.

Grippen is a fighter plane, not a tube amplifier

 

[Jacques Merde]

Have you considered Vanadium Permendur (2% vanadium, 49% cobalt and 49% iron), AKA Hiperco 50? It has a significantly higher saturation flux density of about 2.4T compared with 1.5 for the Metglas 2605SA1.

Hyperco 50 datasheet

I noticed that for the 2605SA1 the material properties looked great, but the published B-H curve in the C-Core Technical Bulletin for the AMCC6.3 C core did not look so swell. Specifically, B vs H is not very linear at reasonable inductions.

That high value of permeability on the material data sheet sure looks good though.

 

[kevinkr]

OT:
Jacque Merde.... As a user id, clever...

 

[Bandersnatch]

OT:
Jacque Merde.... As a user id, clever...

Now we can say we *DO* know him...
cheers,
Douglas

 

[Francois G]

What I found so far is a material called Metglas Magnetic Alloy 2714 which shows superb qualitues - it has the lowest core loss -- To me it looks like the ideal core material, any thoughts?

Lundahl offers an amorphous core in their top transformers - see Lundahl Transformers, audio transformer and tube amplifier transformer manufacturer

Here is what they say about the difference with regular M6 cores of the same size:

"The obvious measurable difference between our silicon-iron cores and amorphous
cores is that the saturation flux for the amorphous core is approximately 33% less than
for the silicon-iron counterpart. This is caused partly by a lower saturating flux level,
partly by a smaller fill-factor due to the thickness of the amorphous sheets.
As a result, power bandwidth is reduced with about 50%. (This means that if the max
output power for a standard LL1620/40mA is 25W at 30 Hz, corresponding max.
power for LL1620AM/40mA is 13W.)"

I don't know what amorphous alloy they use - perhaps Per Lundahl will be willing to email you about it.

Best wishes.
Francois

 

[Jacques Merde]

Now we can say we *DO* know him...
cheers,
Douglas

Yes, but do you understand him?