TDK used to sell inductors that were pre-biased with a magnet to maximize the unidirectional flux swing. I have some small inductors in my junk box (from Electronic Goldmine?) that have little ceramic magnets for the same purpose.
The magnet would have to have a high energy product, be thin, and also non-conductive.
The magnet would have to have a high energy product, be thin, and also non-conductive.
Last edited:
I don't see that it would have to be thin or non-conductive. The path length would be extended by the length of the magnet, and the gap would be considered as it appears on either side of the magnet, which could be an insulator if you imagined it shorting the ends of the laminations. The flux modulation in a high energy magnet would be low for the densities used in audio output cores, so eddy loss would also be low. Well I guess it would have to be thin in that you would want something other than a saturated core at a useful gap length.
Last edited:
I wouldn't look so much for a reduced core volume., instead going for extended flux swing. This will probably result in higher core losses due to said wider flux swing, but then, nothing comes for free. This technique may be useful for extending the operating range for nickel-iron cores with their lower saturation flux density.
Another thought - I never really looked at the notion, but permanent magnets must have low permeability as part of their definition (think about it). The dipoles must resist reorientation in order for the material to be a permanent magnet - this means low permeability. That being the case, the bias magnet for a single ended transformer must needs be pretty thin, otherwise the primary inductance would suffer too much. This probably will limit you to a really high energy material like the neodymium/iron/boron magnets. This will be a thin, flat piece, with the field normal to the surface, meaning that there will be some eddy current losses, mostly at the lower frequencies, where the flux swing is greatest. I would try measuring the resistivity of the magnets I have around, but I think mine are all nickel plated, which tells me nothing about the volume resistivity of the material.
You could probably think about the magnet as a level translator. It has essentually no permeability, for the highest energy types, being just about fully saturated. Whatever field you stuff into one end comes out the other +/- the magnet flux density. So it wouldn't change the inductance of the wound core except for the permeability shift of the static B level that remained in the core with bias current after compensation. The air gap would still be the principal limiter of core flux, so in all likelihood an easily practical thickness of NdFeBr material would be required. This company has a neat online calculator that could help predict some gap flux values without doing any calculations: Australian Magnetic Solutions - Your best source of magnetic products in Australia and the World - The attraction is quite simple!
There was a thread about this very idea and related ones a few years ago. Someone tried a permanent magnet in the gap but got poor results I think. Caused to much loss of permeability by the time they got enough magnet strength. Later I suggested using a toroid core with a plated on hard magnetic film on the laminiation surfaces wound up to make the core. It could then be energized by a powerful pulse at turn on to set the magnet. This allowed a strong magnet due to the large surface area, but kept the gap width to a minimum.
http://www.diyaudio.com/forums/tube...sated-se-output-transformer-7.html#post629867
Easiest way is just to put a bucking winding on the core and power it by a CCS.
http://www.diyaudio.com/forums/tube...sated-se-output-transformer-7.html#post629867
Easiest way is just to put a bucking winding on the core and power it by a CCS.
Yeap !
Why not use the secondary to do that ?
Let's see :
Say our amplifier gives 8 watts on 8 ohms, this means we'll have 8 volts rms that is some 24 (rounded) volts p to p.
Say the primary impedance is 4K and that the DC current is 80 mA.
This calls for a Z ratio of 4000/8 = 500 and thus the turn ratio is 22.
So we need 22 times DC current in the secondary to balance the primary one, and that is 1.76 Amp.
Peek a 24 volts DC power supply able to give 1.76 A and tie it on the secondary thru a CCS wich will waste almost all the voltage and thus have to dissipate something like (24*1.76) 42 watts
Note that whatever is the impedance of the winding to wich DC is applied, we'll innevitably waste at least as many watts that the output tube dissipates
Another approach could be to feed the secondary current thru a coil so we'll waste less voltage accross it.
This coil must accept 1.76 Amp while maintaining an inductance hi enough versus 8 ohms to do not shunt lowest frequencies.
It will innevitably be a big gapped one
It looks like if we have moved the parafeed choke from the primary to the secondary
(edit)
Forgotten to say that, if the secondary resistance is 0.3 ohms (a typical value) the DC voltage accross the LSP will be 2 volts . . . unless we add a serie capacitor
Yves.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.