Extremely low power and low loss transformers

[percy]

A little OT - I know - but since discussions about transformers are mostly in this part of the forum I thought I'd post here to get some inputs.

This is about OPTs in Crystal Radios - yes you read that right. It basically drives a high impedance (10K-20K) headphone connected to its secondary. Impedances on the primary side run anywhere between 100K-200K. Power is extremely low. So we are talking impedance ratios of typically anywhere between 5:1 to 10:1 and power in microwatts.

Generally enthusiasts use line-matching, speaker matching transformers where there is almost always a compromise either in impedance matching or signal loss or both. Now for the actual question -

Say if I'd like to have a custom transformer built, do you follow the same guidelines of transformer design or are there any specific guidelines that should be taken into consideration for this special purpose/application ?

[Sch3mat1c]

Tens of thousands of turns for inductance, lots of insulation and very careful layout to reduce capacitance. Since power is low you can spec the thinnest wire the guy can handle.

Tim

[smoking-amp]

For very low power, you should use a high Mu ferrite core. These have very low losses compared to steel alloy lamination, and the high Mu will reduce the number of turns to get an accepatable inductance for low frequency. The E cores or pot cores are easiest to wind.

Don

[Sch3mat1c]

Well, low power just means you get to use a smaller core. That's why they don't use foot-wide transformers to charge your cellphone. I'd be far more worried about permeability (particularly initial) than loss. (Hey, 10% loss is 10%. You simply don't notice it.) Use mu metal instead of silicon steel or, GOT forbid, ferrite. I'm not aware of any ferrite which has a mu comparable to a solid steel piece, let alone something like mu metal. That'll save turns, reducing resistance and capacitance.

Tim

[Larry Lomax]

I'm not shure of the result, but try to put a little magnet on the magnet path of the tranformer (like the MEG, yes that pice of junk) but not a strong NdFeB but a soft one capable of polarize the iron over the origin of the BH curve. This, maybe, could help not to waist power on reversing polarity on the histeresys cicle.
And it could be improved by overunity amplification!
Do not believe at this, it is just a joke!
Sincerely!

[Sch3mat1c]

That'll cure initial permeability, but won't effect hysteresis. Hysteresis has to do with any change, not necessarily around B/H = 0.

Tim

[percy]

Gentlemen, thank you for the responses.

I think we are headed in the right direction in the sense that initial permeability is in focus.

There are various materials/alloys including ferrites that exhibit very high initial permeabilities but I havent explored what drawbacks or catches - if any - are there in using those materials.

Sch3mat1c, what is mu metal ?

Any other implications about using high permeability cores ?

[Sch3mat1c]

Mu metal is an alloy which has very high permeability, over 100,000. There may be others up in the 200k range. These aren't used normally because they are expensive and have low values for saturation; inductance is also a small concern when it comes to power transformers. Mu metal is just what you need here, however.

Tim

[astouffer]

Ever price mu metal in bar stock? We buy it a few feet at a time at work, half inch diameter is a few hundred bucks a foot. Might as well use gold at that price

[smoking-amp]

I checked in the book "Modern Ferrite Technology" and it looks like 18,000 is about tops for permeablility of available ferrites. 5000 is more typical. Usual transformer lams are only about 1800 Mu (initial Mu). Some amorphous metal alloys are around 10,000 Mu. The 80% nickel alloys go up to 100,000 Mu. Mu metal is in this group. The 100,000 Mu type materials are usually quite sensitive to mechanical shock. Drop them on the floor, and they lose the high Mu. But they are the preferred material for microphone xfmrs. (better hope they are shock mounted in a microphone)

The ferrites are readily available, can get samples for free often. The 80 % nickel alloy tape wound cores are expensive, but maybe can get a sample. Try calling "Magnetics", ask for a Round Permalloy 80 (Mu = 20,000 to 50,000) tape wound core sample or, if you are feeling lucky, ask for a Supermalloy (=MuMetal Mu= 60,000 to 100,000) sample. I was able to get a spool remnant sample of Permalloy 80 tape once. This is whats left over on a spool of the metal tape when its not enough to do a full big core. Dangerous stuff to handle though (for 0.5 mil thickness for high frequency), like a roll of razor blade. Has to be wound into a toroid config to use it.

Don