Recovering WWII magnetic amplifier technology

[wvphysicist]

I have a continuing interest in recovering the schematic of a 2 stage DC magnetic amplifier that was used in my graduate school. I was in the physics lab at the University of New Orleans working with superconductors. The magnetic amplifier was used to provide feedback from a rotating gaussmeter coil to the big magnet current source in our lab. It had a gain of about 1000. I have procured toroids and tested their characteristics. I also made a one stage magnetic amplifier with a saturable reactor. I am looking for help with the two stage circuit because my designs are very crude and use more parts than the schematic which I saw. Is there a similar thread still open?
The circuit which I saw and used was powered by audio frequency AC. It was a DC amplifier connected in a standard operational amplifier configuration. It was not a chopper type amplifier and I do not remember how the output ac was converted to DC, possibly without using diodes.

 

[Baluncore]

The use of positive feedback can give current gains of 200 to 300 and power gain of 10k to 100k. The circuit employed is very similar to the basic two core magnetic amplifier except that duplicated secondary windings are required on the cores, along with a bridge rectifier.

Arrange the two cores with their primary windings driven opposed in series. Each core has two secondary windings. The load, AC supply and one secondary on each core are all connected in a series loop through the AC terminals of a bridge rectifier. The two DC terminals of the bridge rectifier are arranged to feed in series the duplicated secondaries on the two cores so as to work in the same direction as the input DC current.

The more gain you have the more unstable will be the transfer function. Use a regulated AC supply for the most stable results.

Simple, isn't it.

 

[wvphysicist]

I want to build this thing and I had better get the circuit right. I want to upload a schematic but I don't see a way to do it from my computer. The schematic will be wrong any way. It is not complete.
I have some bare cores. I have a terrible time putting turns on the cores.

In the schematic that I saw, I don't remember seeing many diodes. I don't know how they got DC out. I don't think they used a chopper relay. The gain was controlled by a feedback resistor like with an op-amp.
Maybe you can help me.

 

[davenn]

I want to build this thing and I had better get the circuit right. I want to upload a schematic but I don't see a way to do it from my computer. The schematic will be wrong any way. It is not complete.
.....

when creating or editing your post, click on the Go Advanced button just below the text box. When in advanced mode you will see a section below the text box called additional options. See Attached files and click on Manage Attachments button. It will open an upload manager window from which you can browse you computer and get the file you want to upload

cheers
Dave

 

[Baluncore]

Positive feedback circuit is on page 177 of the Electronics Pocket Book, attached.
Thanks Davenn for the upload advice.

 

[Baluncore]

To wind a toroidal core, get a circular spool that will fit like a chain link through the core. The spool can be a slightly larger diameter than the toroid but it's section must fit through the wound toroid. As an example a plastic fishing line spool could be trimmed to fit. Cut the plastic spool once (or twice if very rigid) and link it with the toroid. Tape over the spool's joint(s). Wind a length of wire onto the spool, allowing the toroid to slide ahead of the applied wire. Then wind the toroid by slipping the wire off the side of the spool as you place the wire in final position on the toroid.
Remove any unused wire, the tape and then the empty spool. Done.

If winding many toroids you will arrange powered pinch rollers to hold and drive the spool while spinning the wire onto the spool.

Place windings sequentially, or if they have the same number of turns you may pre-twist bifilar or trifilar before spinning it onto the spool.

 

[Baluncore]

Examining the positive feedback magnetic amplifier in figure 4, you will notice that there is only one current path through the secondary of the circuit. The bridge separates this into two parts. An AC current flows in one part, while DC flows in the other. If you move the load from the AC part of the circuit to the DC part, then the load will have a rectified DC current.

You can change the gain by placing a shunt resistor across the DC feedback windings.

 

[wvphysicist]

I need more tools and parts

Thanks for the help.
I started a previous thread some years back:

https://www.physicsforums.com/showthread.php?p=4171433#post4171433

I don't know if I will have enough room for a feedback winding. My power might be the AC mains. Then I am afraid to handle the bridge while operating. I need to modify figure 4 to get a polarity sensitive output.

I once watched a torid winding machine in Denver. I am looking hard for the right spool. I have used a stick with a notch at each end like a schuttle. I don't see how I can get enough turns on a core. I take at least four times the space that the industry uses.

It turns out that I have Edition 6 of Sams Reference Data for Radio Engineers and the magnetic amplifier chapter is 15. I looked and there was a bookmark right at the page which is 14-4 in the Acrobat file. So I have seen this before. I fear that care must be taken when using diodes to preserve AC flow in the inductor winding or the control winding will not have a good AC null.

I am interested in the characteristic of Sams 14-4 figure 6D. As dc is increased, does the output rise or fall?

I want to understand how these saturating devices work. If the usual equation is delta M= X delta H. Then I presume saturation makes X smaller so the effective inductance is smaller.

 

[Baluncore]

@ wvphysicist.
I think your fear regarding the maintenance of AC symmetry with diodes is based on lack of recent experimental experience in the field. When compared with the metal oxide rectifiers of 50 years ago, silicon diodes today are close to perfection. Getting a good AC null is achieved by using matched cores with ideal saturation characteristics.

Winding small toroids topologically indicates a spool rather than a shuttle. A circular spool can be much more compact passing through the core than a shuttle. You need to wind a couple of general purpose saturable core transformers so you can progress your chain of experiments. Attempting to engineer a final circuit without experimentation is unlikely to provide education or a functional first circuit.

I would experiment at lower voltages, say 12VAC. That is safer because it is both low voltage and isolated by a transformer that will limit power transfer. Lower voltages also imply higher currents and so proportionally less turns need to be wound on your cores, but with thicker wire.

DC output of one polarity is possible from a single magnetic amplifier, but not a bipolar output without a class A bias offset. You can get a bipolar output by using two magnetic amplifiers, one above the other with a split supply plus two diodes that decide which is driven. That could be doubled for use on a single supply as a bridge amplifier providing a bipolar output.