Short overcurrent of 4W resistor

[eTobey]

I wonder how long I could run a 100A current trough a 10mR resitor. Is there any data for this or some rule of thumb?

[Siwastaja]

Transient power capability is basic parameter. Nearly all power resistors have the graph in the datasheet. Just look it up.

Rules of thumb are not very useful because resistor designs vary a lot on their pulse capability. Given same continuous (RMS) power rating (e.g. 4W in this case), two different products could differ in maximum pulse rating by 10x or even more. So just look at the data.

[TUMEMBER]

I wonder how long I could run a 100A current trough a 10mR resitor. Is there any data for this or some rule of thumb?

P=I@2*R---->100A*0,01Ohm---->10 000*0,01=100[W]
U=I*R---->100A*0,01Ohm---->1[V]
I'd say it won't last long. If you can manage the voltage drop to 0.1V, you can experiment with a 0.0001ohm resistor. In your version, you'll have a single-use cap. 1/3 of the load power in "air," in liquid (oil), power multiplier times 3 max. Well, unless the measurement lasts 1 picosecond, the "happy evaporation" effect won't have time to occur.

[Simmed]

I wonder how long I could run a 100A current trough a 10mR resitor. Is there any data for this or some rule of thumb?

ask claude ?

[UnijunctionTransistor]

Rules of thumb are not very useful because resistor designs vary a lot on their pulse capability. Given same continuous (RMS) power rating (e.g. 4W in this case), two different products could differ in maximum pulse rating by 10x or even more. So just look at the data.

This exactly. The surge ratings will vary significantly from similarly-rated thick films vs solid carbon composition vs wire resistors.

[shapirus]

FWIW, I once did some experimenting with passing short (~100-200 ns) infrequent (1/s) pulses of high current (~ 50 A) through small resistors like 0.1 Ohm 0603. Average power dissipation was negligible, but the resistors consistently experienced measurable degradation over not so long time (hours), which manifested itself in increasing resistance. What was the exact mechanism, I don't know. Electromigration, probably?

[Terry Bites]

It would strongly depend on not just the duration of the transient overload but the repetition rate too.
Murphy says with will get jammed full on until is smokes and thats the best value to design around.

[calzap]

Ohmite makes 100 W, 0.01 ohm resistors.  They do require proper heat-sinking.  Main use is for current sensing.
Mike

[IanB]

I wonder how long I could run a 100A current trough a 10mR resitor. Is there any data for this or some rule of thumb?

Apart from looking at the datasheet (you should do that), the power dissipation would be 100 W. If you put 100 W into any small object, it is going to get very hot, very quickly. So "how long?" I suspect is going to be "not very long".

[eTobey]

Not many datasheets say anything about this "basic" parameter. I found one though, now: Its a ohmite that says 5 times the rated wattage for 5 seconds.

I tested 500us pulses every 2 seconds (90A), and you cant see any temperature with a thermal camera on that resistor.

I want to test further, but my IDE is driving me crazy....

[shapirus]

Not many datasheets say anything about this "basic" parameter. I found one though, now: Its a ohmite that says 5 times the rated wattage for 5 seconds.

I tested 500us pulses every 2 seconds (90A), and you cant see any temperature with a thermal camera on that resistor.

I want to test further, but my IDE is driving me crazy....

Good test. Measure the initial resistance precisely with a milliohmmeter (or a trusted ammeter and voltmeter). Then run the test for say 2-3 days. Afterwards, measure resistance again and see if it changed.

Overheating should not be a concern (as long as you calculate the average power and keep it well below the rated value). You'll want to see if there's any degradation of the conductive material.

[IanB]

You can do some basic physics. Can you weigh the resistor? Do you know what it is made of? A ceramic material?

Transient temperature rise is given by:

   E = M c ΔT

Where E is energy (joules), M is mass (kg), c is specific heat capacity (J/kg/K), ΔT is change in temperature (K).

If you apply 100 W for 500 μs, that is 50 mJ. You can do the math.

[floobydust]

There is no rule of thumb possible, it's not such a simple issue. Resistor construction varies widely.
What duration is this "100A" overload? (or 25x power for the 4W part). Through-hole or SMT?
Not a problem for the open-air through-hole shunt resistors. Over 500A for 50msec is reasonable for a 3W part, from a few different manufacturers.
In a multimeter application, they (current shunt) has to be tough enough to clear the fuse, and not get damaged causing a resistance shift.

[mtwieg]

Yeah there are no rules of thumb for this.

Many resistors are "pulse rated" or "pulse withstanding", and will provide a curve of allowable pulse duration vs pulse power (for a single pulse, sometimes also repetitive pulses).

If you don't have this info, then you may as well just guess, or try and discover it empirically (not advised).

If you do find the pulse ratings for your specific resistors, then great, that's your answer. But don't assume it applies to other resistors in a similar form factor.

FWIW, I once did some experimenting with passing short (~100-200 ns) infrequent (1/s) pulses of high current (~ 50 A) through small resistors like 0.1 Ohm 0603. Average power dissipation was negligible, but the resistors consistently experienced measurable degradation over not so long time (hours), which manifested itself in increasing resistance. What was the exact mechanism, I don't know. Electromigration, probably?

It's likely that a hot spot forms at the thinnest part of the conductive film (often where laser-trimming was done) and the film simply burns away little by little. Not electromigration, that only becomes relevant for submicron geometries, far smaller than what you find in any discrete resistor. But overall has a similar behavior.

[MariuszD]

In the case of a wire or tape resistor, it can be weighed and the temperature increase calculated. For other designs, you need to find the manufacturer's data; this data is extremely rarely provided in the datasheet.

Transient power capability is basic parameter. Nearly all power resistors have the graph in the datasheet. Just look it up.

I have a different experience; this is very rarely provided information.

[floobydust]

Even the resistor's ohmic value (within one construction type) causes the pulse power to vary. Graph taken from Yageo Vitrohm KNS series.

Vishay advised me with pulse hits, there is an inflection point on the graph at the time constant where the heat can be transferred from the element, to the substrate. Longer duration pulse power uses this.
Really short duration pulses are only heating the element and its welds/connections.

[Siwastaja]

Can you weigh the resistor?

You would need to weigh the resistive material itself and even then you would  be assuming it doesn't internally hotspot.

No, methods like this are completely unrealistic. Only the manufacturer has proper thermal modeling (or testing) of the properties where limiting factors are the survival of the resistive material itself (e.g., metalization, not the bulk ceramic), its thermal coupling to the bulk, and any hotspotting that might occur (e.g., in metal film due to zig-zag patterns or laser trimming, creating tiny spots that heat up faster). Mass and surface area of the full component (including the bulk body) are more important for average / continuous power rating, but not that meaningful for transient pulse rating, which the OP is interested about.

But really I don't see why OP is struggling so much finding the data. Nearly every decent power resistor datasheet has this curve set. If not, then, well, that resistor is not probably very suitable for pulse application. Using parametric search of Digikey et al, pulse withstand type can be found directly, narrowing down the search.

[eTobey]

You can do some basic physics. Can you weigh the resistor? Do you know what it is made of? A ceramic material?

I doubt, that this formula is that much accurate. The temperatur rises quickly, so I believe thermal conductivity is just "getting its shoes on", when this happens.

What duration is this "100A" overload? (or 25x power for the 4W part). Through-hole or SMT?

The duration is what I would like to know. To prevent blowing the shunt.

Even the resistor's ohmic value (within one construction type) causes the pulse power to vary. Graph taken from Yageo Vitrohm KNS series.

This is some good information.

But really I don't see why OP is struggling so much finding the data.

Because its just not there! Go and look for yourself, and you will see.

[IanB]

You would need to weigh the resistive material itself and even then you would  be assuming it doesn't internally hotspot.

No, methods like this are completely unrealistic. Only the manufacturer has proper thermal modeling (or testing) of the properties where limiting factors are the survival of the resistive material itself (e.g., metalization, not the bulk ceramic), its thermal coupling to the bulk, and any hotspotting that might occur (e.g., in metal film due to zig-zag patterns or laser trimming, creating tiny spots that heat up faster). Mass and surface area of the full component (including the bulk body) are more important for average / continuous power rating, but not that meaningful for transient pulse rating, which the OP is interested about.

I agree with all of this, such analysis as I hinted at is overly simplistic and is insufficient.

That said, for "small" overloads, and "long" durations, where bulk temperatures might have time to equalize through thermal conduction through the substrate, it might give a first order estimate.

[eTobey]

My circuit just blew up...

But the resistor seems to be fine. Just the mosfet got desoldered and detached with a spark. Now its always conducting it seems.

Who would have thought. 

I just had it soldered on old fets leads, as it would otherwise have been a pain to desolder big blobs of solder.

Oh well...

The scope captured the current (kind of). It exeeded 130A, but I dont know the length anyway. I didnt hit stop. 
Maybe it was just the voltage spike that influenced the measurement.

[shapirus]

My circuit just blew up...

But the resistor seems to be fine. Just the mosfet got desoldered and detached with a spark. Now its always conducting it seems.

Ah, been there.

Whacking a 10k resistor between gate and source might help. In my case it would have prevented the blow-up when the pulse signal became floating upon some wire getting disconnected, or maybe when I disabled the signal generator's output.

[thm_w]

FWIW, I once did some experimenting with passing short (~100-200 ns) infrequent (1/s) pulses of high current (~ 50 A) through small resistors like 0.1 Ohm 0603. Average power dissipation was negligible, but the resistors consistently experienced measurable degradation over not so long time (hours), which manifested itself in increasing resistance. What was the exact mechanism, I don't know. Electromigration, probably?

I've seen plenty of "0 ohm" resistors blow up from ~50A pulses, as you say its not instant it degrades over time. Just looks like vaporization of the material at the interface between end cap and resistance material (it necks down so thats the thinnest point with the highest resistance).

[mjs]

Depends really on the resistor, use the manufacturer data for that specific resistor. As an example, 2W 100Ω Ohmite OY-series bulk ceramic resistor is happy with 1 kW short pulses. A 3W resistor with ceramic body and film construction can spew 50mm flames with the same load.

If you did not find data, change to resistor for which you can find data. Here's and example calculator (although I prefer having pulse data in PDF datasheet I can store myself for future reference): https://www.vishay.com/en/resistors/power-metal-strip-calculator/

[M0HZH]

If the purpose is current sensing, it's usually better to use Hall sensors at that level. For example, Allegro ACS758.

[Psi]

You can also make your own high amp / low resistance isense resistor out of pcb tracks with kelvin connection.
You can put all the layers in parallel to share the heat. Just need to figure out the resistance based on copper thickness, width, length etc.. And keep in mind it will have some error as the copper thickness will vary a little from one batch to another.

Edit: As well as temperature compensation, Thanks shapirus