Multi-conductor transmission line impedance

This is exactly what I was looking for -- thanks. It bears out my suspicion that the reduction will be greater than just z/N, even if the reduction isn't as much as I thought it would be.

Before you plunk down a lot of NRE check out places like this:

A few 10ohm coaxes in parallel might do the trick.

Joerg:

Whoa. If it's low enough cost, that may be the winnah.

Thanks.

You can get double-sided flexible PCB material. 2 pieces of copper foil with thin insulation between them. Solder many parallel wires to the ends to terminate to whatever they are connected to. That ought to do it, assuming this is not also a high voltage application.

Jon

It's been mid-80's since I wound my own transformers but there used to be "foil windings" that were pretty thick for high current outputs, like 5v/200A. ...Jim Thompson

But have you ever seen how they make those transformers? They often use a wooden mallet to make sure the copper material drapes around the core in an orderly fashion. At one old transformer winding place the guys who did this sort of work had biceps the size of legs.

Tim's 700A are only pulsed but at that current level every milliohm is going to hurt. Either there have to be lots of parallel cables or it has to be a cable the size of a garden hose, at least.

Actually, if my calculations aren't wrong it can be done with fairly modest-sized conductors. (They're currently using a pair of #10 wires). There's loss, but even if the available energy is fully discharged in one shot (which doesn't always happen) the temperature rise is modest -- and there's a long recharge interval for things to cool down.

Yikes! Did anyone ever measure how much of the pulse energy actually arrives at the destination, and in what shape?

Yeah, plus several large fire trucks, hoses, the local TV news team, a couple of helicopters ... :-)

dy in

maybe just stack something standard like this?:

fold each end over itself and bolt and washer through the stack for each conductor? fold them left and and right if you want alternating polarity

-Lasse

Lessee -- per the ARRL: #10 wire, 1 milliohm/foot, 1-1/3 foot, 700A -- that's less than a volt of drop, resistive. It's the inductance that's killing us.

Oh, details.

I do not know where the energy for this 700 Amps is stored, but if it is in capacitors is it possible that it is ESR performance of the caps and not inductance at the output?

Nope -- we checked. It must be a good thought though, 'cause it occurred to us!!

(Oh - and, we're modest).

Tim Wescott a écrit :

When I designed the Raptor ATE tester DUT PSU we looked at low Zo flat cables Tlines. I can't remember who made these but these were of the shelf, with the specially designed connectors too (we ended not using them because we needed heavy duty fast ZIF connectors). We had Gore making custom low Zo coax cables (in the 10ohm range IIRC) with a massive center wire and thin Kapton dielectric and we used a bunch in parallel.

I just found this but it was not them...

Also AMP have this

ed

ttdesign.com

Do you need low Q inductance, or simply low inductance low resistance?

Again a connection is a connection if it is wider than long. Thus, a sheet of metal 50% wider at each edge, or twice the length, such as 30 inch wide sheet of metal for 16 inches will give you a 'gnd plane'- like connection. To get closer to an approximation of a GND plane, the GND plane has to extend beyond the connections, too. Envision the

Then back to a bundle of wires with interdigitated conductors. ABABABAB If these are 18 Awg wires they'll make up to better than 10 Awg, plus be much lower inductance. How much lower? How low you need to be? So, can tell if done with the design. How low an inductance do you need?

One minor point, how high is the maximum frequency content?

afaik 2 metal layers is the limit. Once you start stacking them they get stiff. stiffness is proportional to thickness squaed.

The golden ears audio people like cables that might be useful to you. Some of them are expensive, but several have construction that leads to low impedance.

For reviews see:

This cable looks like it would be good for your purpose:

You could buy some of the MWS "litz" wire, which is loosely twisted bundles of magnet wire. Use half the strands as each conductor. Maybe they would make you some with thermal-strip insulation, half of the strands a different color from the other half.

Insulation is thin, so Z should be low.

John

Oh, that's short. I was thinking many feet. Looks like an application similar to laser pump diode driving. Besides the inductance the other issue is usually that the inductance "talks" and the radios in the area go rat-tat-tat.