Magnetic DC operational amplifiers 1

[wvphysicist]

I am trying to learn the ancient art of magnetic amplifiers. Not power supplies but real DC operational totally magnetic amplifiers.

I saw a schematic of a magnetic amplifier once at the University of New Orleans Physics Lab. It was used to amplify the millivolt output of our laboratory gaussmeter up to a volt or so to control the current in the big magnet. It's gain was set by a feedback resistor just like a real op-amp. The circuit was powered by an audio frequency AC source. There were at least two inductors and at least four windings. I cannot understand how the two gain stages were connected or how the output was rectified to DC without phase sensitive detector diodes. Maybe it was a chopper but the frequency, I think, was about 2KHz. I saw no semiconductor devices at all in the schematic, only transformers and resistors and few capacitors. The DC gain was around 1000.

I have some tape wound toroids waiting to be wound with wire, to make a home made two stage amplifier if I can get some good ideas.
My understanding is that saturation will reduce the inductance of a winding on a piece of iron. And it is easy to saturate a big piece of iron with a low voltage DC using plenty of turns.
I built an isolated DC ammeter that used two highly saturable toroids driven by AC, a phase sensitive detector and a feedback winding. All that stuff was easy compared to the magnetic amplifier mystery.

 

[Skogsgurra]

Now, I feel real old!

Magnetic amplifiers were the hottest thing when I entered the world of industrial electronics in the late fifties.

There are several terms that describe the magnetic amplifier. If you google "magamp", "transductor", "saturable choke", you will find lots of articles on fertilizer, translator systems and other things. But you will also find a few articles telling you about the technology.

Intuitively, it is about an iron core whose inductance is influenced by a DC field so that inductance decreases when the DC field is increased. More analytically, it is about driving a core into saturation so that the "working winding" switches between "OFF" and "ON" - very much like a thyristor does. The delay is simply controlled by how much "presaturation" the DC field already has achieved in the core.

Transductors were actually the precursor to thyristor drives and there are still a few remaining supplying large DC motors with a variable voltage to run extruders for plastic and I think a few smoke stack fans as well. Those are in the 400 kW range.

Lately, transductors have been used as post regulators in multiple output switching regulators. The big advantage is that there are no thermal losses. The "bucking action" is entirely reactive and does not produce any more heat than a tiny portion oc winding resistance and core eddy current plus hysterisis loss. But, as said, very small.

Good luck with your work in magnetics. Not many understand these things today. But it is a very interesting technology. What you need to do is to get the Volt-second unit into your system. Think Vs whenever you apply a signal to an inductive circuit. It helps a lot.



Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[waross]

Hi Gunnar;
The OP mentions two inductors and assumes that it is a two stage amplifier. Is it more likely that the two inductors run "back to back" to amplify positive half cycles and negative half cycles?


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Skogsgurra]

Yes, they always were two. But more often in parallel. All control windings were series connected, though, so that the induced back-voltages cancelled over one period and could be filtered out in a low-pass filter. Which usually was a series inductor and not an RC filter. The reason for that is that a capacitor would short the induced voltage and counteract the "transducting" property of the cores. Putting a large resistor in series would have helped, but would also consume a large chunk of the available control energy.

The so-called "spar-selbst-magnetisierte" (sorry, can't remember the English name any more) magnet amplifier used two transductor cores within a rectifier bridge so that the DC produced by the rectifying action produced a nearly full saturation of the cores.

A bias winding was then used to take the cores back to unsaturated state so that the output was close to zero. Ten, there were several sets of windings for the input, the feed-back, the stabilizations and so on.

In control applications, it was very convenienjt to have all control circuits isolated from each other.One winding for set value and one for the actual value. Adding signals was easy - just put another winding and apply the signal there.

I have searched the attic for litterature. And actually found a book from 1958. It is by Siemens and the name is "Transduktorbauelemente" which translates to "Components for transductors".

I have scanned a few pictures and also made a rough translation of some of the text.



Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[VEBill]

http://en.wikipedia.org/wiki/Magnetic_amplifier

 

[wvphysicist]

It is likely there were four toroids.
The schematic is only my memory. Did the paint file get attached?
Likely some dots are wrong.

I looked at the Gegentaktvorstufe but cannot tell the resistors from the transformers, or input or output. I believe the top transformer is just for distributing power.

 

[Skogsgurra]

Windings are filled, resistors are not filled.

I will try and find an opamp application. I know there was one in the ASEA (now ABB) ROMA and ROMB systems.

Stand by.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[xnuke]

I've seen the basics of mag-amps taught as recently as 14 years ago in U.S. Navy technical schools. Some old automatic voltage regulators used on turbine generators had them in the control circuits. Some may still be in use in the fleet on older vessels.

xnuke
"Live and act within the limit of your knowledge and keep expanding it to the limit of your life." Ayn Rand, Atlas Shrugged.
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[DHambley]

Last time I saw a mag-amp was just a year ago when a guy brought a design into a review. This was for a new product. We talked him out of it.

They are cool to analyze. Fun too (if you're into math like me), but modern SMPS's with switching transistors are far more practicle.

 

[wvphysicist]

"Last time I saw a mag-amp was just a year ago when a guy brought a design into a review. This was for a new product. We talked him out of it.

They are cool to analyze. Fun too (if you're into math like me), "

Is there any chance for a schematic or detailed description?

 

[wvbht]

Perhaps these sources will help, that is if you can find them:

AMPLIFIER HANDBOOK, Shea, McGraw Hill, 1966. Chapter 8 is "Magnetic Amplifier Devices" by Harold W. Lord. (I have this book.)

H.M. Gale and P.D. Atkinson, A Theoretical and Experimental Study of Series-connected Magnetic Amplifier, Proc. Ints. Elec. Engrs. (London), vol 96, pt. I. pp. 99-124, May 1949.

I remember reading a book on magnetic amplifiers when I was in high school and this may have been the one.
H.F. Storm, "Magnetic Amplifiers," John Wiley & Sons, Inc. 1955.

D.L. Lafuze, "Magnetic Amplifier Analysis," John Wiley & Sons, 1962.

I don't know just how much power you want to control, but here's one that is listed as 160,000-Ampere.
A.B. Rosenstein, "160,000-Ampere High-Speed Magnetic-Amplifier Design," AIEE Transactions, Vol. 74, Part I, pp. 90-97; 1955.

I have other references, but these should get you pointed in the right direction. You will notice, too, that most references are in the late 40's and 50's.
I could scan the reference listing page and get it to you, should you require more listings.

Let me know at wvbtay@hotmail.com

I remember reading about magamps that would amplify audio frequenches. It may be in the book by H.F. Storm.

 

[wvphysicist]

I am not wanting to control any power, maybe milliwatts, but want DC gain of 1000 or better, and linear. Once the millivolt signal becomes some volts it is easy to read a meter.
Is there any chance that you could scan a schematic?

 

[wvbht]

I have scanned some information from an old copy of REFERENCE DATA FOR RADIO ENGINEERS, Fifth Edition.
There are only four drawings and not much to go on, but there are enough design equations to get you started.
I don't think that magamps were ever purely dc, that is to say that the small signal (ac or dc) was used to vary the saturation of a transformer core in a way to control the inductance of the winding(s) to control an ac source. A changing inductance would not have any effect on any dc that was flowing in the controlled winding.
The hewlett-Packard 328(?) DC MILLIAMMETER used an arrangement in the dc-clipon probe whereby the dc effected a change in a 40kHz signal that was amplified and rectified for the metering circuit.

 

[wvphysicist]

I have already made the amplifier in this drawing and it works. One can imagine a DC bias winding at the input and a similar DC source added to the output so that the signal input and output can swing through zero. Or alternately a pair of these can be used in opposition so that like a push pull amplifier the output is symmetrical. To make a two stage amplifier two of those can be put in series making four altogether, and I am confident that I can get a gain of about 30 with a power out of a watt.
The problem is that this circuit has too great a parts count and makes use of many diodes. There is a simpler circuit that has a gain of 1000 or better.