5

I have a diode. It has a forward current rating of 40 mA. It has a surge current rating of 100 mA for t <= 10 usec. There is no mention of a duty cycle in the datasheet.

Is there a rule of thumb for determining what kind of off period I need in order to not kill the diode? The naive assumption is that if I do 100 mA for 10 usec, and I need 40 mA forward current as an average, then I would need to use 10 usec on, 15 usec off.

If that is the case, then my intuition would then suggest pushing 80 mA for 9 usec, and then double the off period to 30 usec minimum. This would allow for overhead like temperature, lifespan, etc.

ajs410
  • 8,381
  • 5
  • 35
  • 42

2 Answers2

6

Your assumptions are basically correct. 100 mA for 10 uS, then off for 15 uS would provide for a 40 mA average and not exceed the specs.

It's always good to "derate" your stuff. A 1/4 watt resistor would normally not carry more than 1/8th watt in normal operation, and your 40 mA diode would carry maybe 20-30 mA on average. There is no hard and fast rule to derating components, it's all a balance of cost, lifespan, availability of parts, PCB area, etc. In the end, it is usually just a guess by the engineer.

Resistors are often derated 50% on power, electrlytic caps are derated 30-50% on voltage, CPU's are almost never derated on clock frequency. Diodes are derated in the neighborhood of 25-50% on power or voltage. Of course, if this is a one-off project where failure isn't terrible then don't derate at all! :)

  • I've never seen a diode rated in power. Voltage drop is very much proportional to current, so they are just rated in amps. – Thomas O May 26 '11 at 20:40
  • Am I right in assuming that derating can be some combination of lower max current, shorter max current duration, and/or longer off period? – ajs410 May 26 '11 at 20:41
  • 1
    @ajs410 Derating is basically intentionally not stressing the part to the amount that's technically allowed in the datasheet. The specifics on that is completely up to you. Max current and average power dissipation are two good areas to consider. @ThomasO You're mostly right. Power is the other side of the Current coin. Sometimes amps makes more sense, sometimes it's watts-- but as you said they are directly related. I just went to the diodes.com advanced parametric search and one of the columns you can sort on is power rating, so I'm not completely alone in this thinking. –  May 26 '11 at 22:35
0

Your assumptions are correct--adjust the duty cycle so that the average current is below the maximum average forward current to not exceed the diode's specs. You should also make sure the peak surge current is not exceeded.

The exact definition for the Maximum Average Forward Current IF(AV) and Non-Repetitive Forward Surge Current IFSM will depend on the manufacturer. Look for an application note from the diode manufacturer.

Advanced Power Technology has a good application note explaining the datasheet parameters in detail for their products. It has this to say about IF(AV) and IFSM:

IF(AV) is the maximum allowable average forward current when the case temperature is maintained at a specified value between 25ºC and 75ºC below the maximum allowable junction temperature.

...

Obviously it has to be an average of something, and, in the APT diode data sheets, the basis for the average is a train of square pulses with 50% duty cycle. In such a train of pulses, the average current is half of the peak current. Therefore, a 15 A rating implies a train of square, 30 A pulses with a 50% duty cycle. Historically, the basis for the average was a train of half sine wave pulses such as would be the output of a half-wave bridge. In other words, the time between the pulses was equal to the duration of the pulses. Because fast power diodes are called on today to handle square or trapezoidal pulses, the basis for the average is no longer the train of half-sine wave pulses.

...

IFSMis the maximum allowable nonrepetitive half-sine wave surge current under the following conditions: TJ = 45°C and the base-width of the half-sine wave surge pulse is 8.3 ms. A sample of diodes is selected and one-by-one the diodes are tested to destruction. This is done by hitting the DUT with a single surge pulse and checking to see whether the diode was destroyed. If so, the peak value of the surge is recorded as that diode’s pulse-height capability, and the next diode is tested. If not, the junction temperature is allowed to return to 45°C, the peak value of the surge is increased, and the DUT is hit again. This process is repeated until all of the diodes in the sample have been destroyed. Then the pulse-height capabilities are averaged and IFSM is set equal to half of the average.

Source: http://www.ohm.com.tr/doc/Microsemi---Defining-Diode-Data-Sheet-Parameters.pdf

jjz
  • 103
  • 4