snubber resistor power rating?

I just simulated a triac circuit with a 100mH inductor. The current through the 100 ohm resistor (using a 100nF capacitor as the other 1/2 of the snubber) was 1.4A, giving a burst of power that peaks at about

200W. Now, the average power through that resistor is only about 84mW, so the average dissipation is small.

The question is whether a resistor will be able to absorb that power spike without damage. There is anecdotal evidence given elsewhere in the thread that it won't be able to in some cases.

--
Regards,
   Robert Monsen

"Your Highness, I have no need of this hypothesis."
     - Pierre Laplace (1749-1827), to Napoleon,
        on why his works on celestial mechanics make no mention of God.

In article , Arie de Muynck wrote: [....]

I haven't thought this through but:

It seems to me that the problem could be helped by adding a bit of complexity to the snubber circuit. This added stuff needs to do this:

When the triac is switching on, the snubber circuit is basically open circuited.

When the triac is switching off, the snubber initially is just the capacitor. As time passes the snubber starts to look resistive.

I have something like this in mind:

! --- C1 --- ! ! -------+------- ! ! V --- --- C2 ^ ! !! ! +----!!---- S1 ! ! !! ! / ! ! +/ o-+ ! X1 ! ! +--????---- ! ! ! --- V ^ --- ! ! -------+-------- !

C1 is the capacitor that is normally part of the snubber.

The bridge and C2 makes a way to soak up the energy without having to turn it into heat right at that instant. X1 is something that gets rid of the energy. S1 closes when the voltage hits zero.

Yes, I know this actually won't work! It does seem that something with a low parts count should be possible.

--
--
kensmith@rahul.net   forging knowledge

But that 1.5 factor is allowed for 5 seconds. I was assuming that from an I^2*t fusing effect that the dissipation capability would go up quite a bit if the time was in milliseconds.

--
John Popelish

In article , Terry Given wrote: [...]

IRC person: Hello, how may I help you?

Ken: I need some pulse rated resistors.

IRC person: Good how many box cars full, and where the siding.

Ken: I only need about 100

IRC: 100 Box cars full is a small order but I think we can do it.

Ken: No, I need 100 resistors

IRC: Ha-ha-ha (Craching sound as he falls out of chair)

Phone: Click Buzzzzzzzzzzzzzzzz

Also IIRC, their spec was weirdly written. It was some time ago but I remember talking to the technical help from some resistor company and pointing out that according to the specs, the part could handle less power for 16mS than it could handle forever.

--
--
kensmith@rahul.net   forging knowledge

Caddock is a perfect example of a poor surge (Pulse) power resistor. They are thin film and X1.5 rated peak power. A good surge rated resistor is X5000 or maybe 5 Joules. Call Richard Caddock and see if they speak in Joules.

Cheers, Harry

A-huh, assuming. Harry

No, X1.5 applies to 0.1 attoseconds to 1.5 seconds. 1.500001 applies below

0.1 attoseconds.

(or something like that)

Thermally the situation looks a lt like this:

Power in = I in R1 R2 R3 R4 R5 -> ---///------///----///----//----//--- GND ! ! ! ! ! --- --- --- --- --- --- C1 ---C2 --- C3 --- C4 --- C5 ! ! ! ! ! GND GND GND GND GND

C1 is the thermal mass of the actual resistive material.

R1 is the thermal conductivity of the resistive material

C2 is the thermal mass of the substrate the resistive material is on.

R2, R3 with C3, C4 are a lumped constant representation of the substrates bulk.

C5 and R5 are for the outer package to the air.

Unless you have some feel for R1 and C1 you really can't extrapolate from

1.5 seconds down. I know this because I did it and ended up with 1206 open circuits after a few months of service. The power spike was only 300uS long.

--
--
kensmith@rahul.net   forging knowledge

IRC (the company, I do Recall Correctly) make some great smt resistors. And have peak pulse power curves.

Cheers Terry

Yep. A good data sheet is much better than an assumption:

--
John Popelish

Philips (or whatever the hell they are called now) have nice peak pulse power curves in their discretes book. an 0603 RC21 5% resistor has an average power rating of 63mW, and can take a repetitive 1ns 2W pulse every 1us.

It would be nice if they just gave us the damn RC thermal model, instead of maybe giving us peak pulse power curves. I got a nice surprise the other day when looking at an Infineon FET - a complete thermal model, as part of the datasheet. yay.

Cheers Terry

Now, find me someone who stocks them.

Wow, thats GREAT! pisses all over the CHP series!

I'll remember that part, thanks John.

Cheers Terry

doh.

Cheers Terry