How much power can a resistor take for a short duration?

Thread Starter

scubasteve_911

Joined Dec 27, 2007
1,203
Hi everyone,

I don't do much high power stuff, so please excuse my ignorance.

I am trying to design an inrush circuit for a fractional horsepower dc motor circuit. The circuit is attached along with a simulation waveform.

Basically, I am getting on Channel A the current in amperes RED (not volts) through the resistor, and on channel B is the voltage drop across a power resistor. The resistor is 15 ohms.

Is there some sort of other limitation than average power for a resistor? Clearly, this is a startup routine and does not repeat often, so we can assume the resistor will have time to cool.

How can I find out if this thing is going to blow up, other than building it?

I estimated about 12-15J of energy need to be dissipated. Or, about 180W in 80mS. How can I choose a resistor? Do I have to do the whole volume and specific heat thing to calculate a heat rise?

Steve
 

steveb

Joined Jul 3, 2008
2,436
How can I find out if this thing is going to blow up, other than building it?

....

Do I have to do the whole volume and specific heat thing to calculate a heat rise?
This sounds tricky to me, but I'll just throw a couple of ideas out in case they might be helpful.

1. Some power resistors specify maximum operating temperatures and even give derating curves to show power capability versus ambient or case temperature. I remember using a small power resistor (3W) that could run at 270 deg. F, (I think) before derating kicked in. This was much better that other 3 W resitors I found. So, power rating alone might be misleading in your application.

2. If you do a calcualtion, you have to wonder if your model of the material, construction and dimensions of the resistor are realistic. But, that is one estimate you can do, but you would need to leave a large safety margin.

3. If you don't mind doing a little experiment, you could measure the surface temperature versus time for the resistor with 180 W. You can even burn a few up just to see what the limits are. This data should help you estimate if it will survive for 80 ms. Test multiple manufacturers/types and then don't change once you decide.

4. If you want to be rigorous. Rig up a cycle test putting the expected worst case power spike in every 10 sec (actual time depends on resistor average power rating) for a few weeks. Ideally, you'd do this in a thermal chamber under worst case conditions. Check the resistance before and after to see if aging occured. There are probably ways to do accellerated aging tests, but I dont' know the guidelines for that.
 
Last edited:

KL7AJ

Joined Nov 4, 2008
2,229
This sounds tricky to me, but I'll just throw a couple of ideas out in case they might be helpful.

1. Some power resistors specify maximum operating temperatures and even give derating curves to show power capability versus ambient or case temperature. I remember using a small power resistor (3W) that could run at 270 deg. F, (I think) before derating kicked in. This was much better that other 3 W resitors I found. So, power rating alone might be misleading in your application.

2. If you do a calcualtion, you have to wonder if your model of the material, construction and dimensions of the resistor are realistic. But, that is one estimate you can do, but you would need to leave a large safety margin.

3. If you don't mind doing a little experiment, you could measure the surface temperature versus time for the resistor with 180 W. You can even burn a few up just to see what the limits are. This data should help you estimate if it will survive for 80 ms. Test multiple manufacturers/types and then don't change once you decide.

4. If you want to be rigorous. Rig up a cycle test putting the expected worst case power spike in every 10 sec (actual time depends on resistor average power rating) for a few weeks. Ideally, you'd do this in a thermal chamber under worst case conditions. Check the resistance before and after to see if aging occured. There are probably ways to do accellerated aging tests, but I dont' know the guidelines for that.

A wirewound resistor can handle severe overloads for short times. In fact, as they start to glow incandescent, they add further current limiiting. :)

eric
 

Thread Starter

scubasteve_911

Joined Dec 27, 2007
1,203
Thank you all for your responses!

Bertus,

I wonder if I can extrapolate the 5 X figure for 5 seconds? Maybe be a bit conservative and assume 0.5 seconds at 180W? 5W resistor X 5 X 10 = 250W for 0.5 seconds?

Steveb,

Thanks for the comprehensive post. I am actually doing this as a side project right now, on a personal basis, do I really don't need to qualify the parts or what-not. I would rather overkill the requirements a bit in order to save my time, because it would be the difference between a 2$ resistor and an 8$ resistor.

I really just want to see if it would explode or vapourize some part of it. I know there is that 'infinite' heatsink limitation for semiconductors, but there isn't that I have seen for resistors. I will do some thermal mass calculations, maybe it won't be too big of a deal because of the extremely short duration.

Eric,

Very interesting, I will look to see if anyone has specs on these that consider high overloads.

Thanks again for your help guys,

Steve
 

SgtWookie

Joined Jul 17, 2007
22,230
I'm inclined to suggest that you could go with a 50W resistor, as long as the motor wouldn't be in stall condition for very long.

OTOH, you could use PWM to ramp up the speed of the motor, and monitor its current to ensure that no parameters were exceeded. Then you might be able to use a resistor of much less wattage.

It depends upon the load placed on the motor.

Were it me, I'd try to do away with the 15 Ohm resistor altogether, and perhaps use something like a synchronous buck converter w/PWM to keep the motor under control. Resistors basically waste power.

If you can control the current through the motor by using an inductor with a synchronous half H-bridge, you'll practically eliminate that loss in the resistor.
 

Thread Starter

scubasteve_911

Joined Dec 27, 2007
1,203
Hey guys,

Eric! Very nice article, so, carbon composition is the way to go, which they sort of elude near the end.

Wookie,

This is actually just for the inrush circuit, the motor part will be later. The resistor basically limits the current to the capacitor during inrush, then, it gets bypassed by two parallel mosfets (only showed one) in order to rid of high power losses. The FETs will have about 30-40mOhms, which will be in series with the filter capacitor.

This method eliminates the need for a NTC, since these run very hot. I tried to use a MOSFET and a low-pass filter on the gate, then have a 3.3V signal switch it over slowly, but it didn't work well it the AC. It worked beautifully with a pure DC signal though.

From charts and stuff I have been looking at, I am planning to use two 5W 30ohm resistors in parallel.

Steve
 

bertus

Joined Apr 5, 2008
22,270
Hello,

For my work as service engineer I have to measure RF power from NMR transmitters.
The solid state NMR transmitters are 1 kW.
I am using a Radial 20 Watt 20 dB attenuator.
I give pulses of 10 μSec of 1kW and a repetition time of 1 Sec.
Nothing will happen to my attenuator (there is enough time to cool down).
The liquid NMR transmitters are 300 Watt.
These I can measure with 10-20 μSec and .5 Sec repetition time easily.

Greetings,
Bertus
 

studiot

Joined Nov 9, 2007
4,998
Wirewound resistors can quite happily glow red hot, as indeed they do in electric fire bars.

The main problem with this is that their resistance changes with temperature so a large temperature change moots a large resistance change.

If you can live with this about 1.5 metres of electric fire wire will do you nicely.
 

steveb

Joined Jul 3, 2008
2,436
Wirewound resistors can quite happily glow red hot, as indeed they do in electric fire bars.

The main problem with this is that their resistance changes with temperature so a large temperature change moots a large resistance change.

If you can live with this about 1.5 metres of electric fire wire will do you nicely.
Also, you have to be able to live with the higher stray inductance with 1.5 meters of wire. Probably OK in this application, but might be important in some cases.
 

SgtWookie

Joined Jul 17, 2007
22,230
If one went back-and-fourth in "U" shapes longitudinally on a ceramic cylinder, additional inductance would be minimal. Where one gets into inductive difficulties is when the turns are wound concentrically on a cylinder.
 

Thread Starter

scubasteve_911

Joined Dec 27, 2007
1,203
Thanks a lot guys! I really appreciate your input :D

I will post my circuit when I am finished, and hopefully make a project contribution within a month's time.

Steve
 

steveb

Joined Jul 3, 2008
2,436
If one went back-and-fourth in "U" shapes longitudinally on a ceramic cylinder, additional inductance would be minimal. Where one gets into inductive difficulties is when the turns are wound concentrically on a cylinder.
Well, yes I agree. I wouldn't expect someone to wind a coil, but even a long straight wire at 1.5 m has some inductance which would depend on the wire diameter. I would expect about 0.1-1 uH inductance which is perhaps small in this case, but could cause problems in some applications.
 

studiot

Joined Nov 9, 2007
4,998
To lower the inductance use the 'bifilar' winding technique. I have posted this somewhere before in this forum.

In this case simply double the wire back on itself and wind the doubled wire any way you like. The flux in each half of the doubled wire opposes and cancels out.

Post 7 of this thread shows a photo of something I did many years ago with an old fire bar.

http://forum.allaboutcircuits.com/showthread.php?t=12372&highlight=electric+fire
 
Last edited:

steveb

Joined Jul 3, 2008
2,436
To lower the inductance use the 'bifilar' winding technique. I have posted this somewhere before in this forum.

In this case simply double the wire back on itself and wind the doubled wire any way you like. The flux in each half of the doubled wire opposes and cancels out.
Ah, of course! SgtWookie apparently meant the same thing with the back and forth which would have a similar cancelation effect.
 

floomdoggle

Joined Sep 1, 2008
217
Try a ten ohm ten watt power resistor. Works really good in tube amps. Rectangular, about 2 inches long, 1/4 inch square. Carbon resistor. High heat resistance. Never seen one blow yet.
Dan
 
Top