Magnetic Amplifiers

They were and probably still are used for nuclear instrumentation in Naval Reactors. I never heard of one failing up until the time I left the service. At that time there were probably 60 to 80 reactors in ships. Far fewer now.

Transistors can't even begin to match their reliablity.

I have a variation of one that is used to control the heat in my heat treating furnace. I can control the saturation of the core using a rotary switch and therefore the heat produced in the furnace.

I first saw Mag Amps in the old Rad Lab series and little has changed in the design of them since. And you're correct in that its mainly used in the naval industry today.

Prior to WWII the US used Thyatrons and other means on their subs and Germany used th mag amp. This allowed them to stay down longer on less batterys than we could. The sub in the Museum of Science & Industry (Chicago) actually used Mag Amp technology.

There was a resurgence in interest in this when Atomic Subs were built as you point out, and is highly classified. A variant is used in modern electric power plants and AC distribution for Power Factor Correction. They use a synchronour motor instead of just a coil to get the capacitive reactance side. Neat stuff!!

I did an MTBF on a mag amp design once and it was in the 100 million hour range with max stress ratios applied. I figured I'd screwed something up and finally showed it to my boss who laughed and said, "yea, you goofed, it's about 1000 times higher". Blew my mind.

To answer your question as to its use today, yes, variable reluctance (reactance) control is still alive and well.

Bill S

This is all well and good, but for the lesser educated, what is a magnetic amplifier and how does it work??

I can remember an article in either 'Wireless World' or 'Electronic Engineering' in the forties 'build your own magnetic amplifier' but I no longer have it or a clear recollection of it.

Amazon.com have a reprint of a 1942 paperback on magnetic amps for $8.00 so I have sent away for one and will publish some drawings when I get it.

The wireless world crowd thought they were rather clever as they suggested using positive feedback to increase the gain.

.....my how CR4 brings back memories.....I have not seen a Mag amp since leaving the Royal navy in 1973.......and sadly, I have forgotten the details of how they work!!

Since a few seem interested, Mag Amps are basically a form of AC Valve that controls the Flux Density in a magnetic circuit. Picture a series connected AC Source with a variable Inductance and a Resistive load. If the Inductance is large (compared to the load), very little AC voltage appears across the Resistive load, and if the Inductance is small (Saturated is the term used) the Source voltage is effectively connected to the Load. Intermediate states can also be used for such things a Speed Control of a motor, and are very efficient (>90% min).

The neat part is how the Inductance is changed between these two states. They couple magnetically to a single core (usually an uncut tape wound) with a great deal of turns on the control winding and only a few in series with the load. Since N*I is proportional to B (Flux Density), and the core has a fixed B-H curve, the value of its Inductance can be varied with very little current in a magnetically isolated circuit, hence the "amplifier" term in the name. Actually pretty lousy in terms of fidelity.

The key is the magnetic material the core is made from. Folks like Mag Inc have been making these cores for years. Materials like Permalloy and others are still made and their website used to have some excellent articles on magnetics design. The B-H curve of these materials has a very narrow profile indicating low Hysteretic Loss, which used to be the "figure of merit" for these products.

I'd expect there are some better materials today but they were pretty good 20 years ago. I'd expect to see them used in space stations and mag-lev trains in the future. Perfect applications in terms of efficiency and reliability assuming they are correctly designed and herm sealed. The only thing I ever saw kill one was a massive overvoltage/current from a direct lightning strike. The FMECA actually said it had "endured inumerale " (sic) strikes before it actually failed.

If you do a "google" on "Mag Amp", and "Saturable Reactor" you'll find some good articles no matter what level you want to continue this on. I would also advise some remedial physics before you do. Understand how an inductive element behaves and the basic laws that govern them. It'll make the articles more enjoyable.

It might help to remember the correspondence between an electric circuit and a magnetic circuit, and in particular Ohms law:

• Electric circuit: Resistance = EMF / current (with dimensions: [Volts per Amp])
• Magnetic circuit: Reluctance = MMF / Flux (with dimensions: [AmpTurns per Weber])

MMF is the MagnetoMotiveForce [AmpTurns] which sets up the magnetic Flux [Webers]. Reluctance is very often more convenient to calculate directly as follows:

• Reluctance = Length / (u0 * uR * Area)
• where Length is the length of magnetic core
• where u0 is permeability of free space = 4 * PI / 10 000 000
• where uR is relative permeability of the material
• where Area is the cross-sectional area of magnetic core

Plug these into your favorite spreadsheet and have fun learning about the Magnetic equivalent of Ohms law. You can make up series and parallel circuits that are calculated just like resistors in series and parallel.

NeilJ

Just a few more ideas and comments following on from my previous post:

• EMF is ElectroMotiveForce which is commonly expressed as [Voltage]
• Flux density is easily calculated: B = Flux / Area [Webers per sq m]
• Reluctance of an air-gap is easily calculated as follows:
• Rgap = (Gap-Length) / (u0 * Area) [AmpTurns per Weber]
• u0 dimensions are: [Weber per (Amp.Turn.Meter)]
• On the same spread-sheet you could model electric circuit and magnetic circuit interactions, and hence start to model and understand Magnetic-Amplifiers
• Use MKS / SI units, and all should calculate out fine

NeilJ

We use magnetic amplifiers in SMPS more than 25 years with great success. In figure the example of use of a so-called "slow magnetic amplifier" is shown.

It works as a current limiter of a primary winding of transformer T1. The factor of amplification on capacity is equals to 100-1000.

Krass.

P.S. I can't load my circuit. If needed mailto: yikras@mail.ru

About twelve months ago I recommissioned 4 1976 model W120C Electric Drive trucks that employed "Mag Amps" or Saturable reactors. The technology has long been superceded in this field but they worked fine until the customer destroyed the chassis by overloading. The mag amps controlled the excitation to a pony driven exciter (via a PWM) which excited the main alternator which in turn supplied traction current to series wound traction motors. The advent of power electronics led to use of "Static excitation" of both the Alternator and the seperately excited shunt traction motors. I've also serviced them on reclaimers and draglines. The drive system is more prone to flashover than modern systems for several reasons including the slower reaction time. They are robust but tend to be fiddly to maintain as a system. The fault is probably not all directly due to the reactors, but constant calibration is a feature of keeping them running.