OT - Transformer coupling in concentric coils

This one's aimed at all you gurus busily adding to the "Motor/Inverter" Thread.

We're busy commissioning a system at the moment which is designed to produce very high magnetic fields - up to 100 T for about 10ms. The basic design principle is that a capacitor bank of up to 15 2.88MJ modules dumps its charge into an outer magnetic coil, at a guess about 1 kA into a ~500 mH coil for about a second or two. This will develop a base field of 20 - 30 T. When that field is established and stable (which I guess will be identified by trial and error, or at least trial and measurement) further capacitor banks will dump their charge into a smaller concentric inner coil - ~30 kA into 3 mH - which will produce an additive field of 60 - 70 T on top of the base field. The inner coil is about 12" - 15" long and about 6" diameter and has so far produce a field of a little over 60T before self-destructing. The outer coil will obviously be bigger than that but how much bigger is difficult to guess, it doesn't exist yet and I haven't seen any design info.

What we're wondering about is what is going to happen to the current/voltage in the outer coil when the inner coil pulses. Obviously the coupling is going to be nothing like as efficient as in a transformer - a colleague believes he has heard a value of 30% being bandied about - but we're still left wondering whether there will be a possible significant reversal of current flow back into the capacitor bank modules.

Strangely enough we've never heard any discussion about this point, although it seems pretty fundamental to the whole design concept - especially if the changes in current flow in the outer coil are sufficient to cause a significant change in the base magnetic field.

I be grateful for any thoughts or comments that you people used to playing with magnetic fields in motors and transformers have to offer.

Wont you also also get volts induced in the inner coil when you pulse the outer to get the base field established. (NB: why have two fields ? why not one big coil + one big dump of current ?)

It sounds to me a lot like the coil in an auto ignition system - you know, where 12 VDC is magically turned into a few thousand volts DC to make the plugs spark.

Tom Jenkins said:

It sounds to me a lot like the coil in an auto ignition system - you know, where 12 VDC is magically turned into a few thousand volts DC to make the plugs spark.

Click to expand...

oh yea. otherwise known to some of us as "a curling iron". =]

SimonGoldsworthy said:

Wont you also also get volts induced in the inner coil when you pulse the outer to get the base field established. (NB: why have two fields ? why not one big coil + one big dump of current ?)

Click to expand...

This is true, but on the assumption (a big one, since it's far too long since I sat through these classes!), that the induced EMF is in the opposite direction then since the Thyristor has not yet turned on it should block the induced voltage. If the voltage is in the other direction then we may have a problem because current could then flow through the freewheel diode parallel to the SCR.

Two coils are needed because fields of > ~70 T are not possible with available materials. The forces generated at fields > 70 T are apparently great enough to disrupt the molecular structure of steel, i.e. it would tear unsupported steel structures apart. The windings of the coil are 4 x 2 mm flat copper wire with a woven outer covering of a Kevlar-like material. Each layer of the coil is then also bound in this material and soaked in some sort of epoxy resin. The first experimental version of the coil (which took them three weeks to wind) lasted 30 pulses before tearing itself apart with a pulse that produced a field of just over 63 T - not far off the current world record of 63.6 T.

Our worry is that the induced currents will interfere with the pulse form, or in the worst case if current is pumped back into a module, cause damage to the module. We're already seeing unwanted changes in the pulse form caused by the slightly different charging voltages (up to ~ 200V difference on a 24 kV charge) when discharging 3 or 4 modules in parallel into a short circuit and this is what started us off wondering about other effects. Obviously the short circuit is the worst case and it may be that when we have a coil on the end things may look a bit better - time will tell!

RMA said:

On the assumption (a big one, since it's far too long since I sat through these classes!), that the induced EMF is in the opposite direction then since the Thyristor has not yet turned on it should block the induced voltage. If the voltage is in the other direction then we may have a problem because current could then flow through the freewheel diode parallel to the SCR.

Click to expand...

I was just concerned that the thyristor could handle the voltages considering the large rates of change involved.

The Thyristor stack is rated for 35 kV (I believe), so I would hope that should be enough. It's a funny thing, I obviously wasn't involved in the purely electrical discussions, but it seems odd to me that this aspect never seems to have been discussed. I hope we're not going to be in for any unpleasant surprises the first time we discharge into two concentric coils!

There again, I'll probably a long way away before it gets to that stage!

Any coils that have mutual inductance will have induced voltages any time a current flows in any one of the coils. The potential voltage is definately measureable. NO current should flow in the windings where the thyristor is turned off.

When the additional coils are fired, you should be able to see the induced voltage on the other coils.

As you have a 35KV stack, do you have the appropiate voltage rated scope probe? Or are there low voltage test points for looking at the voltage waveforms? I am curious as to how you are looking at the signals.

As you have a 35KV stack, do you have the appropiate voltage rated scope probe? Or are there low voltage test points for looking at the voltage waveforms? I am curious as to how you are looking at the signals.

Click to expand...

We can measure the voltage on the capacitors (or later on the coils when they are installed) using a 1000:1 voltage divider from HiLo. Since this is DC coupled it's pretty accurate. The current we measure using a Rogowski coil which, of course, is de facto AC coupled. The cut-off frequency is about 1.5 Hz so it's not too bad at the moment while we're testing into a short circuit, but it'll lose a bit more when the coils are in place.

Thanks for the reply.

The voltage dividers are usually the most accurate for measuring voltage. I was amazed at how well the actual reduced waveform resembles the full voltage ones. I have looked at high voltage signals with the correct probe and the low voltage signals simultaneously.

I am not familiar with the current measuring device you mentioned. I presume it is a type of CT. I worked for a mfgr of medium voltage VFD's 20 years ago and was assisting in proving the low voltage signals were accurate. As it has been a while, I was curious if some new method has come about. Seems like some methods just get the job done.

As it has been a while, I was curious if some new method has come about.

Click to expand...

Strangely enough, this one belongs in the "Oldies but Goodies" section. So far nobody seems to have come up with anything better when it comes to measuring 100s of kA.

We looked at the possibilities of using a mOhm shunt to get rid of the AC-coupling, but the more we looked at it the more problems came up, so we're still using the Rogowski coil, even if the principle is over 100 years old!

As usual, Wikipedia has all the answers! Read all about it here!

I know that type of coil by the term "dog collar". Ya just gotta be careful putting it on the dog.

The ones I have used are more of an HED than a CT. Some I have used will also do DC current.

Are you aware that most of the technology for electrical apparatus we use today was in use before world war 2? The methods of employing it may have changed but the basic technology is the same.

One of the best motors manuals I have ever read was a 1942 US ARMY/NAVY manual. One company I worked at issued them to the customers that came in for training on the control systems we built for them.

I never worked on such devices in order to produce specific magnetic fields but spent a few years working on power supplies/weld heads used in electrical resistance welding. We used to dump a whole heck of alot of energy in through these pulse transformers in order to gain extremely high current levels but really small voltages, able to fusion weld different metals with it. It was such a cool job, modifying power supplies, re-designing weld heads, measuring and recording results and solving problems. What you guy's are working on is just so interesting I can't stand it. Best of luck to you and your team.