3

I'm an EE, not a physicist or chemist, so please forgive if this question is dumb.

I need a resistor that can dissipate many kW for 10 us, followed by > 10 ms off time.

Since the duty cycle is ~0.1%, I could use a wirewound rated for > 100 W, but the large inductance those things have makes them a bad option. While the bulk is not a problem, the hundreds of dollars they cost is.

I figure if I mix clay from a pottery store with aluminum powder or some other conductor finely ground up the resulting slop could be made to have almost arbitrary resistance. The desired resistance is 50 kohms although it does not need to be very precise. Anything within a factor of 2 would be nice.

I'm also willing to dissolve salt in water and allow the water to vaporize as a cost of doing business. It's not ideal but it's not out of the question either.

Is this in the realm of feasibility or is this all ridiculous? I don't know a lot of lattice structure crystaline chemistry stuff, and sometimes you've go to care about that sort of thing when it comes to conductors.

JRE
  • 67,678
  • 8
  • 104
  • 179
James Strieter
  • 373
  • 1
  • 6
  • 2
    Last time I did something similar, we ended up with screen printing conductive paste on ceramic substrates and water cooling them. Controlling the thickness of the printing was the hardest part. If it’s not mass production, there are some regular 1/2 W THT resistors with spectacular high peak power rating, in the order of 100 kW for 1 us. At 1 kW average, you would need a couple of thousands of them though. Give Kanthal or Isabellenhutte a call? – winny Mar 20 '23 at 17:47
  • 11
    1 MW and a 50 kohm resistor means a voltage of 224 kV (no matter how briefly it is applied). Are you sure you know what you are doing? – Andy aka Mar 20 '23 at 17:50
  • 5
    What are you actually doing? – Tim Williams Mar 20 '23 at 18:04
  • How low inductance do you need? How about 500 regular 3-5 W film resistors (500 V working voltage on each) in series and probably two such strings in parallel on a very long (set of) PCB(s)? – winny Mar 20 '23 at 18:27
  • 1
    Wound resistors need not have a large inductance - see bifiliar - but high voltage may mandate prohibitive dimensions. There have been "impulse power resistor" operated submersed - I wish I remembered in what medium. – greybeard Mar 20 '23 at 18:44
  • @Andyaka I should have thought of that. The resistors will see several kV, though not quite that high. – James Strieter Mar 20 '23 at 18:59
  • @TimWilliams I want to put this resistor in series with a diode in order to passively spin down an inductor. The purpose of the resistor is to decrease the time constant of the LR circuit while inductor is de-energizing. – James Strieter Mar 20 '23 at 19:00
  • @greybeard you just gave me an idea. I can wind a coil so that for every turn conducting in 1 direction, there is another turn conducting in the opposite direction. Then the magnetic fields cancel and resulting inductance is near 0. – James Strieter Mar 20 '23 at 19:03
  • @winny how did you make conducting paste? – James Strieter Mar 20 '23 at 19:04
  • 1
    @JamesStrieter The way they do it in non-inductive wire wound resistors is a bifilar winding, basically take the resistance wire, fold it back on itself at the center, and wind it on a form. – GodJihyo Mar 20 '23 at 19:10
  • 1
    We bought it. Came with datasheet and instructions for how to print it for a given resistance and power rating. – winny Mar 20 '23 at 19:14
  • 2
    @JamesStrieter more basic than that: application. What value inductor (L, DCR)? What waveforms? What diode? What switch? Can you even do what you are proposing to do (are there other component limits of which you aren't aware)? Not being an EE or physicist, asking for a specific resistor (especially of such ratings) sounds either very ambitious -- or you might be missing something. – Tim Williams Mar 20 '23 at 19:34
  • 1
    I think you could make this with an oven, 10 micron silica and some graphite or carbon black and a couple other easily-sourced materials. PPE is required to avoid pulmonary fibrosis. Look at how composition resistors were made. – Spehro Pefhany Mar 20 '23 at 19:53
  • 2
    I would look into liquid resistors. Much easier to adjust the resistance by changing electrolyte concentration and/or changing the submerged area of the electrodes. 10kW for 10us is 100mJ which is negligible in the context of e.g. a 5-gallon bucket of electrolyte. – vir Mar 20 '23 at 20:02

2 Answers2

2

There's a type of resistor that's made exactly for applications like yours (pulsed, low inductance, high power): Metal film resistors in transistor packages.

Here's an example of a 20W type that can take extreme pulses such as the one you expect to dissipate in the resistors: https://www.vishay.com/docs/50005/rto20.pdf

Depending on the actual voltage you expect to encounter, you could put 10 to 20 of them in series, giving you a total voltage rating of 5kV to 10kV and a power handling capability of 200W to 400W. The momentary overload capability of this resistor string would approach 100kW for 10µs, according to the datasheet. (Each resistor can handle 20W continuously, >5kW in 10µs pulses, and 500 Volts)

Keep in mind that you have to cool the resistors appropriately. Additionally, if the voltage spikes exceed 2kV, you have to put insulating pads between the resistors and the heatsinks to avoid dielectric breakdown of the resistor's package. Alternatively, you can tie the heatsinks to one leg of each resistor (or group of resistors) to lower the voltage between the resistor and heatsink. Note that this puts the heatsinks at high voltage.

Jonathan S.
  • 14,865
  • 28
  • 50
  • Try reading the question again maybe. Then read my comment under the question. I might have misread what the op said but, so might you. – Andy aka Mar 20 '23 at 18:03
2

I figure if I mix clay from a pottery store with aluminum powder or some other conductor finely ground up the resulting slop could be made to have almost arbitrary resistance.

This is actually pretty much what ceramic composition resistors are, which are commercially available. They have a very high pulsed power rating, comparable to the old fashioned carbon composition resistors (which for some reason aren't made anymore--I'm not sure why); they can typically handle many hundreds or even thousands of times their continuous power rating in pulsed applications. (They may be rated in pulse energy rather than power, as that's what the fundamental limitation is.)

If it's also a high voltage application (which I assume it must be for several kW into several kΩ), you may need to spread the voltage out over multiple resistors. You could make your desired 50 kΩ by putting fifty 1 kΩ resistors in series, for instance, and provided the PCB layout is appropriate, you could get fifty times the working voltage compared to a single resistor.

Hearth
  • 27,177
  • 3
  • 51
  • 115