R Power Derating for C Inrush Current Limiting Applications

[SeattleChris]

Hello folks,

I want to use a ceramic cap + series resistor in place of a single Tantalum cap for the load capacitance on a USB VBUS power input (end-point). I'm doing this for both cost and cap size (height) profile advantages. I'm planning on using a 1 ohm resistor in series with a 4.7uF ceramic cap, but I'm not sure how to quantitatively determine the power rating for the resistor.

If I assume that the resistor will limit the rail overshoot to 5.5V on hot plug, the initial (max) instantaneous current through the resistor will still be V/R = 5.5/1 = 5.5A and the power requirement is therefore 30.25W!

I know that I probably only need a 1/4W resistor but I'm not sure how to use temperature derating curves for resistors to do the math.

I think I'm missing the formula for temperature rise per Watt per unit time through the resistor. It's probably something obvious, that I'll kick myself for later, but having an ego doesn't help me get my job done here! :O)

Thank you for your help!

Chris

 

[electricuwe]

Temperature derating curves will not be very helpful in this case.

The basic approach is like this:

- calculate the energy stored in the capacitor
- during charging the same amount will be dissipated in the resistor
- search for resistor datasheets that specify "single pulse rating" of resistors (that is an energy rating)

 

[itsmoked]

You can always look at the resistor specs to see what their instantaneous current limit is. The limits are often pretty high before they flash-bulb. Otherwise you need to figure out the "area under the curve" of how much power is going to be dumped during the charge up.

I would just cut to the chase. How hard is this to test? A cap , a resistor, a rigid power supply. Pick the resistor you want to use and pick the next size down. Try the one you want to use. Does it survive? Okay use the next size smaller. Did it survive? If yes, then the size you want to use has some power head room and should be fine.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[jimkirk]

Your energy storage in the cap is 71 microJoules.
I doubt that there is a 1 ohm resistor that can't tolerate that, even in an 0402 surface muont package.

Put another way, the 5.5 amps will decay with a time constant of 4.7 microseconds, hardly enough time to heat up the thermal mass of the resistor.

What you're really interested in is the "transient thermal impedance". Google that, see if you can find a resistor that has it specified.

 

[SeattleChris]

Thanks everyone. After poking around the KOA Speer website a bit more I did find an App Note on the subject:

http://www.koaspeer.com/pdfs/Thick_Film_Pulse_Limitation.pdf

However, the time resolution on the graph only goes down to 10us, so I can't get a definitive answer from it directly. I know what I'm looking for now so I can take it from here. I'll see if Panasonic or Vishay-Dale has a similar app note. Yeah, I could figure it out empirically, but I prefer understanding how to calculate the requirement so it becomes a design skill rather than a data point from an experiment... ;O).

I see what you mean jimkirk. After one RC (4.7us) the voltage across the resistor will be 0.368(5) ~ 1.84V which means the power will be down to 3.3W. Pretty much at the app note graph limit for a 0402 resistor. I'm going to see if I can find a better graph somewhere else. I would expect the curves to turn sharply upward at very short pulse durations. According to KOA I need a 1210 package! :O)

Thanks again everyone!

-Chris

 

[electricuwe]

You found exactly one of the appnotes I have used in the past. If you want to fulfill the pulse requirement with a smaller size resistor search at Welwyn/TTelectronics, e.g.:

http://www.welwyn-tt.com/pdf/datasheet/PWC.PDF

In your case the pulse waveform is exponential, this will be less demanding then the same energy in a rectangular pulse. You can either regard the limit for a rectangular pulse as a sfe side limit for your case or do a more detailed analysis.

I can only warn anyone to adress this topic empirically:

If such a rating is violated, the resistor will not fail immediately, but it will degrade quite fast in application.

You should consider that even for the so called thick-film tehcnology which is used for al large amount of small resistors, the active element has a quite small thermal mass comapred e.g. to wire-wound resistors of past decades.

 

[SeattleChris]

Thanks everyone for all the helpful replies. I'm going to go with the TT Electronics PWC series in a 1206 package for now. I'm more concerned with the H dimension than L & W so this will work fine for the current application.

-Chris

 

[itsmoked]

Thanks for the update. Good luck.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com