Permanent magnet bias for choke.

Is this for a tube thing? A trick I remember reading about is running the AC tube filament current thru the other winding of a coupled inductor of appropriate turns ratio connected as a differential choke

Easiest way is to add a DC bias winding around the core with desired number of ampere turns in opposite sense as filter winding...

Then my choke becomes a transformer with a shorted secondary, unless a high impedance drive is used. Too difficult and I don't have power to spare.

Cheers

Shades of my 1950 Nash, field coil on speaker in the trunk also served at the power supply (vibrator) filter ;-) ...Jim Thompson

It could

Can you not use a rare earth magnet (Neodymium) inside the core to gain your negative force?

There are a few papers on the technique.

As far as I know an iron powder core can not be made into a permanent magnet - which is why it is used. I don't have the theory to support this assumption though.

John :-#)#

The bias magnet has to go in the magnetic path, and its permeability is very low, so it looks like a huge gap. You'd gain more in current handling capability at a lower cost in inductance if you just used a gapped core.

Alternatively, how about a cap multiplier? With a low-sat transistor and an auxiliary supply, you can get the voltage drop down to a couple of hundred millivolts.

Cheers

Phil Hobbs

I just magnetised an iron powder toroid core using a NeFeB magnet. Not much, but enough to attract/repel a compass needle.

Cheers

I'm filtering several volts of ripple at 8A and typically hundreds of Hz.

Lotsa turns makes a step-up transformer, so that's lotsa volts to accommodate.

A gyrator which can cope with that sort of headroom at 8A is one hot thing. I don't have room to make it large.

Cheers

Somebody used to sell inductors that had PM biasing. But I think the magnet acts like a shorted turn and makes problems.

Just add a choke in series with the bias winding.

Oh, wait...

Okay, the "several volts" thing would have been good to know at the start.

Where are you getting a variable frequency signal that large?

Cheers

Phil Hobbs

I would agree with tricky. The first problem is whether the 4 amp choke wire is adequate for 8 amps. Next is how do you determine that you have a bias equivalent to 4 amps. And trying to bias a toroid seems to be a major project.

Could you add an active circuit to eliminate the ripple? A LM317 would not work at that current, but it could drive a pass transistor. But you probably do not have the room to do that.

But you might be able to use a voltage regulator to reduce the ripple for the circuitry that is ripple sensitive and accept a couple volts of ripple in the rest of the circuit.

Dan

Dan

Yes, sorry. 1kV line too.

Motor drives, pre-existing customer property, can't be changed. Some newer types are very good, some older ones are absolute s**te but were never designed with power line signalling in mind.

Cheers

ome

Windmill generator housings limiting work space?

John ;-#)#

Yes, any metallic magnet (NdFeB, SmCo, AlNiCo..) would be bad.

Ferrite, has low conductivity, but also low incremental permeability (it looks like an air gap), so you gain a lot of air gap distance in the process, probably more than you wanted. Which is good for energy storage, but past a certain point, bad for copper loss because you need that much more for a certain inductance.

Tim

These were common TV sweep generator components, usually a ferrite solenoid with affixed magnet (one end painted bright red). The magnet was a ceramic type, probably not high conductivity. The rare-earth magnets with nickel plating (to keep 'em from crumbling) wouldn't be as suitable.

Not unworkable; the current in the correction winding could be anything convenient, like

The LM317 doesn't current-regulate well at MHz frequencies, though, so an inductor that filters the HF would keep its current-source impedance high, and would only have to tolerate the 200 mA.

Any such method is equivalent to a linear regulator, once you pull everything through the transformer. There are some shortcuts, where if you take the turns ratio, you can use less power, at the expense of handling less ripple; but this is equivalent to the hum-bucker circuit where a coupling capacitor and a lower-voltage amplifier is used to supply ripple to the higher-voltage supply.

Tim

with 4 volts of ripple accross the inductor at 8A, there's 160V compliance needed maintain drive that 200mA current.

Maybe some sort ot LED lighting power source couls be used, if it's fast enough.

Yeah, it looks like a hard-to-fill niche. The current source (LM317) could be low element of a cascode to get that 160V compliance, but energy loss is considerable. Still, there's places the energy would be useful, like an incandescent bulb (howzabout a 20 watt power indicator lamp?).