pushing a resistor network to the thermal limit (heat sinking it)

[SArepairman]

So I have a resistor network that is quite nice but I need it to be able to dissipate more heat on the 100 ohm part.

http://www.vishay.com/docs/60044/100-267.pdf

This thing is rated to dissipate up to 0.1W, but I am not sure how hot it will get at this temperature.



If I fold this part down and use thermal paste to stick a heat sink on it (as big as is not ridiculous), how far can I push the dissipation.

Ideally I would like to put 250mW through it indefinably, instead of its 100mW specification, how far can a heat sink "extend" this limit? What if I stuck a little peltier cooler on it?
(I am shorting out various resistors to use this thing a decade gain adjust and it would be nice if the ratios always held)

I guess I just want someones educated guess as non of the thermal specifications are available. I am going to heat sink this thing regardless to improve its performance ever so slightly..

*i cannot reduce the currents, the only option is to use an external 100r resistor.

and does anyone know of a heat sink that would work for these types of resistor network, kind of like a clip heat sink?

[T3sl4co1l]

If you're using it as a voltage divider, you'll be hard pressed to get more than 0.09W into it, not 0.1W, and not even the 0.5W "package limit".  It's only rated 300V.

If you're using it slightly differently, like, I don't know, driving current into the bottom resistor, as... a current mirror, or something?  You'd be better off finding something that can handle a lot of power.  0.1% 100 ohm resistors aren't hard to find, maybe in higher wattages but there's probably something still.

Assuming this is a new circuit, not a mod, I find it hard to believe you can't "reduce the currents" somehow anyway.

Tim

[SArepairman]

ehh I guess you are right. How about a clip heat sink for it?

and the biggest problem is that it is 100 ohms rather then 90 ohms, I just realized this problem........ so I will need a 900 ohm in parallel with it.
I might as well use a external 100/900 network I guess, to keep things civilized

I can't reduce the currents because its acting as the last part of a current source.


But, do they have a decade divider thats 100/1000/10000/100000/1M?

Everything I found either is 1 to nine or 1 to 1.01 etc. I never found a strait decade box style resistor network.
If I could get a "normal" divider then I would stick to reduced currents, if I have to use a 900 ohm parallel resistor then I might as well use a external 100 ohm and blast the TCR tracking to hell for the extended current

[SeanB]

Disconnect the 100R section and add an external 100R 0.1% 0.4W metal film resistor in it's place. Will reduce the tracking tempco slightly but will handle the power nicely. Place the resistor in contact with the thick film unit and it will work.

[SArepairman]

Disconnect the 100R section and add an external 100R 0.1% 0.4W metal film resistor in it's place. Will reduce the tracking tempco slightly but will handle the power nicely. Place the resistor in contact with the thick film unit and it will work.

just thermal paste them? how do i connect the thick film to the other resistor?

[SeanB]

Connect it to pin 2, and just have the resistor next to the thin film, touching the white side of the substrate. That will work for most applications with very low loss of the TC of the whole assembly. Not as good as a high power thin film unit, but those are not really going to be good at high power anyway.