Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

MOSFET running at high temperature

Status
Not open for further replies.
T

treez

Guest
Hello,
In our 180W LED driver, the SPB07N60C3 mosfet was running with 170 degree measured on its tab.
We wonder how long it could survive at this temperature and whether much internal damage will be done?

SPB07N60C3 DATASHEET
https://www.mouser.co.uk/datasheet/2/196/SPB07N60C3_Rev.2.5-72639.pdf

The datasheet says it can operate at up to 150degC…so presumably there is no wearout mechanism up to 150degC?
 

Hi,

Not only the Mosfet is the problem, but the solder joints, the pcb material, delamination...
And what happens whe the ambient temperature is higher, reduced air flow, ....

I don't like the idea to go that close to the limit.

Klaus
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Also consider what happens when you turn your driver off, there will be additional contraction and expansion to worry about, inside the transistor and joints and components within it's heating influence.

Brian.
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Thanks, i know its an impossible question, but what would be the upper limit on the temperature at which this mosfet could be run at without havign a reduced lifetime of the fet.
The datasheet does say 150degC.
 

treez said:
The datasheet says it can operate at up to 150degC…so presumably there is no wearout mechanism up to 150degC?
Click to expand...

In the datasheet the 150C° is the junction temperature, not the Tcase. So if you measure 170C° on the case the temperature could be bigger inside, like 190C° or more, depends on dissipated power.
I am not sure extreme temperature causes hard damage, maybe it just degrades threshold voltage and changes its guaranteed electrical performance by the datasheet. But hard to believe it was 170C° on the case for me.
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Sorry i meant "junction".

May i add, We didnt measure the temperature on the case of the dpak, we measured it on the metal tab.....the DPAK was dissipating about 20W altogether. The datasheet doesnt give a thermal resistance from junction to metal tab.
 

Hi,

I'd say every tenperature rise will reduce lifetime. I think temperature vs reduced_lifetime is an exponential function, thus only the higher temperatures will have significant effect.

There will be true absolute temperature effects: aging, oxidation, migration...
And there will be temperature_cycle effects: mechanical stress on glue, traces and solder....
I assume there are technical papers dicussing these effects.

I design industrial equippment, usually running 7/24.
I try to keep temporary hot spots less than 120°C and long time temperature below 70°C, both worst case.
But these are not cost critical designs, because one malfunction will cause a complete production lot in a chemical process to be lost.
This often means thousand times the extra cost of the equippment design.

Thus I don't have experience with your extreme temperatures.

Klaus
 
  • Like
Reactions: asdf44 and treez

    T

    Points: 2
    Helpful Answer Positive Rating
    A

    asdf44

    Points: 2
    Helpful Answer Positive Rating
Metal tab and the case is the same I think...almost, I don't know. But with 20W you can calculate about 200C° at the junction.
Measuring equipment can decrease thermal resistance to ambient too, so Tj is at least 200C° I think.
 
Last edited:

Hi,

I agree: metal tab = case.
R_th is given with 1.5 /3.9K/W.
Makes 30K / 78K at 20W

Klaus
 

Some Semiconductor Manufacturers give Lifetime for their components depending on ambient temperature.
So your MOS that works in 170 degree ambient temperature will be dead in a short period regarding to them.
At least, its life will be shorter than expected.
 

Thanks, the case is plastic, so i was thinking it would be a higher thermal resistance from the junction than the metal tab?...i mean surely the metal tab must be closely thermally coupled to the junction?
 
Last edited by a moderator:

Hi,

don´t refer to the plastic case.

"case" = metal tab.

--> Rth(jc) = thermal resistance junction to metal tab.

Klaus
 

So your MOS that works in 170 degree ambient temperature will be dead in a short period regarding to them.
Click to expand...
Thanks, this is presumably due to repeated severe expansion and contraction of the finely engineered silicon bits and pieces inside the fet, as well as the bonding wires?

Presumbaly the bonding wires just become detached after a long period of months of 170 degc operation?...or they snap?.......or the silicon infused bits just get deformed?
 

KlausST said:
Hi,

don´t refer to the plastic case.

"case" = metal tab.

--> Rth(jc) = thermal resistance junction to metal tab.

Klaus
Click to expand...

For some reason I had to hash this out with a confused thermal engineer recently:
If a part has a thermal pad or tab -> Tc is that thermal pad or tab
If a part doesn't have a thermal pad or tab -> Tc is the plastic case

So anyway, yes, in this case it's the tab.


Temperature ratings usually are a somewhat arbitrary point on a lifetime curve. Almost everything lasts longer if kept to lower temperatures. And that's before considering temperature cycling issues.

Some parts are better than others at exposing these curves. Shopping for electrolytic caps recently some that are quite good at giving you lifetime vs ripple vs voltage vs temperature to let you trade those things off. I haven't seen such things for mosfets however.
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
measuring the tab temp on a device with switching on the tab is problematic - and often error prone

at low switch currents the device may well last at high temps - but not forever - the solder joints will deteriorate - and the die heat/shrink cooling cycles will affect the fet eventually...
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Thanks,
We put sone kapton tape over a bit of the metal tab, and then glued the thermopcouple to that, so that we could read the temperature while the product was powered.
 

Ah huh - it pays to check by switching the input power off and see that the temp decays in an expected fashion - if it jumps down 20C or more in 1-2 sec - then likely reading higher than true...

Those IR temp measuring guns are pretty cheap these days too, laser guided ...
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Long term wearout mechanisms are one thing, you can
run hot if you run less current than the reliability current
limit is based on.

But high temp also increases parasitic NPN gain while also
increasing the B-E shunt resistance (body access resistance)
bringing the device closer to onset of breakdown and burnout.
You do not know how close you are, how close you will be
on the next tube of parts, or what "just a bit more over the
top" stress might be applied by you, the power company or
the customer to put you on the wrong side of a pile o' stanky
field returns and/or plaintiff's bar.
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Usually this sort of academic discussion is appropriate when operating near, but still below, the rated Tjmax. But if you're >20C above that, then IMO you have no right to expect anything to be reliable. Your time is better spent trying to reduce dissipation, or improve thermal conductivity.
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Presumbaly the bonding wires just become detached after a long period of months of 170 degc operation?...or they snap?.......or the silicon infused bits just get deformed?
Click to expand...

Neither. They are welded nicely (I presume) and except for random failures, they will stay put.

There are elements diffused in the Si substrate (also overcoating for connection) and they will now diffuse faster at higher temperature and cause device failure (in general).

During manufacture, the diffusion profiles of several dopants are carefully controlled and operation at high temp over extended periods will cause a change in the device characteristics.

Consider a simple p-n junction diode; operating this at high temp for extended period will change the characteristic of the junction appreciably and can cause the diode to fail (not bonding wires or wafer deformation)
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top