1206 resistor that can dissipate 1W?

[treez]

Please can you confirm that the 1206 resistor of the "LRMA" series cannot really dissipate the 1W that it says it can, without being soldered to a PCB with multiple cooling copper planes, including a significant "same side " cooling copper plane?

https://www.welwyn-tt.com/pdf/datasheet/LRMA.pdf

 

[schmitt trigger]

You forgot to add, without being cooled with liquid nitrogen.

Now seriously, it is poor engineering practice to operate a resistor at its full rated power for extended periods of time. One must de-rate it. For commercial purposes, 80% derating is applicable, for hi-rel applications I've seen 50% derating.

 

[treez]

Thanks, but I expected you to go further, I mean, a 1206 is usually just 250mW absolute max....so why are Welwyn thinking they can get 1206's up to 1W, or even 50% of one watt?

 

[KlausST]

Hi,

You are right, with standard 1206 package and no additional layout effort it is not possible.

But one can rely on the datasheet, wit the described copper planes and thermal viasvit will be possible.
Did you recognize the max. temperature of 170°C?

And i agree, it wiil not become my favourite 1W resistor.


Klaus

 

[schmitt trigger]

I've seen other manufacturers using high temperature resistive inks, which allow very significant increases (double) in power dissipation for a given package.

Even though the resistor itself will not burn, the concentrated heat source could affect nearby heat-sensitive components. Judicious board layout would be required to fully use this resistor's power rating.

Have you seen SiC diodes and transistors? The high temperature they withstand is beyond amazing. But I've seen evidence that the thermal stresses they create affects the long term reliability of solder joints.
Again, a good understanding of the component is required. A well thought layout coupled with a FLIR analysis may be required to fully use its potential.

 

[mrinalmani]

Well... that is a metal alloy resistor and not a film version. It is absolutely common for metal body resistors to dissipate 1W in a 1206 package. I have used 2512 2W metal-alloy resistors. Although they ran extremely hot, but they didnt fry even at higher power than rated. See the attachment below, it will confirm that 1W in a 1206 package is not uncommon for metal alloy current sense resistors.

 

[treez]

The datasheet says it needs thermal vias to multiple layers, even though its metal alloy...this is what I suggest in my initial post...that it could do it, but would need multiple thermal via connections to multiple copper planes.
I think it would be unwise to use this part, as it clearly needs very special pcb layout treatment to make it workable......best t use a resistor that is not so over-sensitive to thermal pcb layout procedures.
Otherwise I could see this part de-soldering itself.

 

[FvM]

The board area required to dissipate 1 W is about the same for a 1206 or a 2512 resistor. It's much larger than the resistor size itself, e.g. 20x20 or even 30x30 mm, depending on the acceptable board overtemperature.

The smaller resistor will expose a higher local peak temperature, respectively show larger TC induced resistance variation, so you'll choose the larger resistor if the component density isn't critical.

 

[treez]

OK thanks, We actually have a 1206 resistor which is metal alloy on our PCB and it dissipates 400mW.......the PCB is in early prototype stage. Nobody willallow the pcb to be made bigger.
The 1206 resistor concerned here has three 0.35mm vias near one pad going to the bottom layer through the 1.6mm FR4 board. I cannot tell if these vias are filled. Also, I am not sure if its even possible to ask a pcb assembler to fill just three particular vias with solder....I thought if you ask for filled vias, then you have to have all vias on the pcb filled , or none filled?.

 

[FvM]

Instead of filled vias and similar exotic stuff, you should use a multiple standard vias to connect the resistor pads to inner and backside copper.

It's the sum of heat sources in combination with the power planes and copper pours that decide about the temperature distribution of a board. At best, the total PCB surface can equally dissipate the generated heat.

 

[schmitt trigger]

The 1206 resistor concerned here has three 0.35mm vias near one pad going to the bottom layer through the 1.6mm FR4 board. I cannot tell if these vias are filled. Also, I am not sure if its even possible to ask a pcb assembler to fill just three particular vias with solder....I thought if you ask for filled vias, then you have to have all vias on the pcb filled , or none filled?.

What I've seen, is that on the stencil one makes small openings over the vias to allow solder paste deposit. Then during the solder reflow, the solder will fill the vias.

But as FvM mentions, is best if you put multiple via holes in the first place.

 

[mrinalmani]

Individual vias are probably not explicitly filled with copper. Whether a via will be filled or not after the electroplating stage depends upon the diameter of the via. The standard wall thickness of plated through hole vias is 150um. Therefore any via that is less than 300um in diameter will automatically be filled. I also remember an application note of TI suggesting that thermal vias should be sized between 0.3 to 0.33mm so that they may be filled completely. And also, we may use a mesh-grid of .33mm via with a centre-centre distance of 0.5 to 0.6mm instead of using larger vias which will ultimately land up not being filled completely.
I had asked several PCB manufacturers about "copper filling of thermal vias", but they all refused saying that they cannot possibly fill individual vias. All what they could do is allow slightly higher plating time so that all the vias may have thicker walls. But then increasing via thickness beyond 150um required a minimum clearance of 10mil, as opposed to 8mil for a 2oz PCB
May be larger manufacturers allow for fill individual vias, but none that I have spoken to offer such a thing.

- - - Updated - - -

Also... I seriously doubt the fill-factor of a via filled with a re-balling solder paste.

 

[treez]

thanks that's useful info , though I was referring to filling vias of any size completely with solder.

..just seen your edit.,..i am of similar opinion. and suspect that filled vias probably aren't really filled with solder...there will be gaps?