Superconductors?

HTS is being used with liquid nitrogen, nut to my knowledge, it very limited in the high power- high current applications. I think the NMRI community uses LTS. best research into this area, but again I am not sure.

There are several superconducting projects using liquid nitrogen.

For super conducting cables and current projects check out

Interesting! I first thought that the skin effect must be tremendous in HF applications. After a few seconds I realised that the opposite must me true. You can use very thin wires and that means less pronounced skin effect. Is that the way you think about superconducting?

Thanks for the info so far guys. Maybe I should look more closely at LTS as well. What method of cooling is normal?

Hmmm, not really skogsgurra. As I understand it the skin effect limits the depth of usefull conductor. A bundle of thin wires will allow only the outer ones to conduct. Superconductors might just enable me to carry large currents in the skin, without high impedance. I may very well be wrong.

I have considered having bundles with spaced out wires, to reduce the em field concentration towards the centre. This may not give any advantge with regards to current flow.

ElectLect,

I think that that is a waste of resources. Use just one thin wire. Since resistance is zero, there is nothing to gain "ampacitywise" using a bundle of wires. Zero resistance is hard to get used to.

Hi ElectLect,

Have a Google for Litz wire. This is basically a bundle of thin wires insulated from each other. Used in high-frequency medium-power applications where skin effect is a problem. Higher current applications tend to use copper laminated busbar for similar reasons.





------------------------------

If we learn from our mistakes,
I'm getting a great education!

ScottyUK
note that Litz wire is only effective below 1MHz and that in order to be effective the strands are woven to be first inside the bundle then outside.

All,
superconductors are not zero impedance, they are zero resistance at DC. The resistive loss increases with the square of frequency and can exceed ordinary copper above 100GHz (see Ramo and Whinnery … , Fields and Waves .., 3rd ed)

logbook,

I can understand that the impedance increases with frequency. But the losses? Are we talking about radiated energy (=loss) or does the resistance (R) really increase? That would not be a superconductor any more. Is that so?

Most superconductors are not ideal, hence too high a current density will usually detrimentally affect one or more of its other performance parameters. That's why there might still be a need to run a bundle to minimize the current density.

TTFN

OK. IR,

They are like people then, not perfect in all respects.

I live what I say ;-)

TTFN

I echo IRSTUFF's statements, based on early work of HTS superconducting antennas, With the caviot, that the magicians doing the chemistry were in fact having success in pushing the cutoff higher and higher in frequency.

Thanks Logbook,

To a power guy 100kHz is high frequency; 1Mhz is RF and not something we deal with very often. Strand transposition is something we use in the very heavy windings on generating plant. Same problem, except our problem occurs at 50Hz because the conductors are so large. Just shows how different the perspective is from different industries!




------------------------------

If we learn from our mistakes,
I'm getting a great education!

Again, thanks guys.

I am actually considering litz wire. I'm not sure what weave I should go for, for the frequencies of interest to me (~1GHz, maybe 2.45GHz). Likely I would need to have something made up especially, but couldn't find much design information on Google. It needs to have very high current handling. How much does weave effect signal group velocity?

Am I to understand that superconductor may not be appropriate at the frequencies I need? I must admit I'm not sure if I'm getting myself into an expensive dead-end with this. I will be contacting Oxford Instruments at some point, to learn a little more about their superconductor usage. They make NMR imaging machines, but the mag field is DC (as I understand it) so I'm not sure how much I can learn.

The other thing that worries me is how superconductors react in high EM stress environment. I gather current levels are limited, for this very reason. Anyone give me some (more) usefull pointers?

Skogsgurra,
it would seem from Ramo & Whinnery (op cit) that the loss is a surface resistance rather than an equivalent radiation resistance, but there is no reference to the method used to establish these pretty graphs.

ElecLect,
I strongly advise against Litz wire construction above 10MHz. From what I have read (and understand generally) it would be worse than useless. Current likes to flow on the outside of the wire because it is easier. By transposing the filament to the inside of the core you make that section hard to get through. The impedance and loss should be higher when Litz wire is used at too high a frequency. Do a web search on Litz wire and you should see recommended operation frequencies up to around 1MHz.

Guys,
In general for superconductors, don’t think that just isolating the strands is going to fix the "current density" problem. The big problem is magnetic field strength. You cannot take any current at all if you run the superconductor at its critical temperature. In order to allow current you have to be below the critical temperature. As you increase the current you have to run progressively lower than the critical temperature in order to remain superconducting. Above a critical field strength even running at absolute zero will not preserve the superconductivity. Thus a practical superconductor needs a high critical temperature, a high critical field strength and preferably ductility (something which these new high temperature ceramic superconductors are not good for!).

Thanks for the info on Litz, Logbook. I may now have a solution to the skin effect, but need to do some calcs/simulations. Any other useful ideas?

I'm also interested in material current density. I have always assumed that non-cooled copper can handle about 10A/mm2, is this about right? Likely it will be either Cu 101 or Cu 103, material grade. How does superconductor compare? I realise that high e-m stress will affect this.

10A/sqmm is rather a lot but that depends on the application. On pcb tracks you don’t want the copper to get too hot because the tracks will peel off. However, long before you reach that point the voltage drop would probably be unsatisfactory. I always consider the volt drop rather than the temperature rise, resulting in cooler tracks.

For wires you are again interested in melting the insulation or volt drops, depending on the run length. Heating effects then depend on the conductors being encased in nice insulating conduits or not. ‘Open-air’ copper could get stupidly hot without causing too much trouble, but then again the volt drop might be a problem.

The most efficient shape for conductors for low skin loss (at a given cross-sections area) is hollow cylinders (not strips). Running the coolant through the conductor is therefore a nice idea.

IIRC, for a reference point, certain commercial MRI scanners use 38-40 guage Niobium alloy wire with DC current in the 200 - 400 amp range.

The coil is kept inside a Liquid Helium dewar, high-temp stuff need not apply.

Thanks Logbook and Slagenthor. Extremely useful information, as I have come to expect on this forum.