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

TO247 tab connected to drain...why?

Status
Not open for further replies.
C

cupoftea

Advanced Member level 5
Joined
Jun 13, 2021
Messages
2,638
Helped
54
Reputation
108
Reaction score
116
Trophy points
63
Activity points
13,845
Hi,
Why is it that all TO247/TO220 SMPS switching FETs have the metal tab connected internally to the drain? Surely the back tab should be connected to the Source, or not connected to anything electrically….this is because the heatsink would never be willingly connected to the drain, as it’s a switching node…and you woudlnt want your heatsink propagating the switching node?

So why cant you find FETs with tab connected to source?...or are there some?

Example of Drain connected to tab...
 

Sort by date Sort by votes
The drain terminal is the die backside, the drift region lies
between the body "neck" and the drain terminal. The
drain being the case / tab, comes from the mechanical
structure (source inside body inside drain).

You'd need an "advanced packaging" to flip-chip the die
to tab, yet snake the gate connection out of that thicket
(presumably a plethora of source bumps). But then you
would be choking the thermal path, down to those same
bumps with their thermal impedance contribution, and
probably these bumps sit on fired vias in an insulator
/ interposer (to get the gate, out) - more thetaJC.

And the heat is generated in the drift region, which
at high voltage encompasses most of the die thickness,
so you'd really rather have it close as practical to the
slab.
 
Upvote 0 Downvote
You can buy MOSFETs in a package (e.g. TO-220F) with the tab insulated from the die.
Otherwise you can use a thin mounting insulator to isolate the tab from the heat-sink, of course.
 
Upvote 0 Downvote
Hi,

there are isolated packages. Like TO-220F.
But the difference in thermal resistance is huge:
0.5K/W non isolated vs 3.4K/W isolated package. (From Fairchild FDP(F)39N20 datasheet. Randomly chosen)

Klaus
 
Upvote 0 Downvote
there are isolated packages. Like TO-220F.
But the difference in thermal resistance is huge:
0.5K/W non isolated vs 3.4K/W isolated package. (From Fairchild FDP(F)39N20 datasheet. Randomly chosen)
Click to expand...
You can buy MOSFETs in a package (e.g. TO-220F) with the tab insulated from the die.
Otherwise you can use a thin mounting insulator to isolate the tab from the heat-sink, of course.
Click to expand...

Thanks, that is very interesting....i actually wonder if the TO220-F package has lower RthJC than the TO220 when the TO220 has a thermal spacer between itself and the heatsink?

The pads dont seem to have much info in their datasheet

Page 14 of this gives thermal conductivity in W/mK, so i guess its a case of converting that to K/W?

The best thermal rubber pads seem to have a W/mK of some 0.9......thats the same as concrete......copper is 384 W/mK.

Wish to try and find the Rth(pad).?
 
Last edited:

Upvote 0 Downvote
Hi,
cupoftea said:
i actually wonder if the TO220-F package has lower RthJC than the TO220 when the TO220 has a thermal spacer between itself and the heatsink?
Click to expand...
I guess it depends on the spcaer. There are many different types. Silicone, filled or not, ceramics, mica...

Klaus
 
Upvote 0 Downvote
..thanks, im trying to find out what the RthJ-H woudl be like with one of these ?

..there must be some way of using the W/mK figure to get the RthJH value when its on the back of a TO247 FET ?

My "thermal sense" kind of tells me that a 0.152mm thermal rubber pad with W/mK of some 0.9 would be better than using a T0247-F type FET?...ie lower therm res to heatsink?

If i was offered to put my finger on one or the other...i reckon it'd get less burned if i touched the metal tab fet with the 0.15mm thermal rubber spacer........and that for any type of spacer on the market....than the F type FET with no spacer.

I believe F type FETs are made of ceramic...known for very poor thermal conductivity?
 
Last edited:

Upvote 0 Downvote
As you know, thermal conductivity of 0.9 means that for a 1m^2 area of 1 metre length of this thermal pad, then say if one end of that "bar" was 20 degc...then the other end would be 19 degc if the hottest end was having a dissipation of 0.9W
 

Upvote 0 Downvote
In audio circles - often the metal case of the TO-3 can was connected to the Vcc or Vss of the supply to limit interference to the heatsink and thence to other more sensitive parts of the circuit - this is still done in fact.

Ideally it would be better - in many apps - to have the source as the tab ( Power integrations does this on its parts ) - instead of having to isolate the drain - and have the capacitor this forms to the heatsink.

The problem is manufacture ( cost ) and the need to get good thermal properties, means the back of the die needs to be soldered to the main structure of the copper lead frame - the tab and hence the drain.

There are many power diodes ( TO-3P ) with insulated copper tabs that use Al2O3 isolation in the device to allow fitment direct to a common heatsink - but they are more rarely used these days ...
 
Upvote 0 Downvote
cupoftea said:
..there must be some way of using the W/mK figure to get the RthJH value when its on the back of a TO247 FET ?
Click to expand...
It is. You determine the effective metal tab area (the part that is directly connected to the chip), divide by insulator thickness and get the "m" value.
cupoftea said:
My "thermal sense" kind of tells me that a 0.152mm thermal rubber pad with W/mK of some 0.9 would be better than using a T0247-F type FET?...ie lower therm res to heatsink?
Click to expand...
Can you refer to a TO247-F MOSFET on the market? Good thermal interface materials have W/mk numbers of 6 up to 10, by the way. Only Al2O3 ceramic is much better.

Thermal spacers can have a considerable amount of capacitance. Using it on an isolated heatsink doesn't necessarily reduce EMI.
 
Upvote 0 Downvote
it is. You determine the effective metal tab area (the part that is directly connected to the chip), divide by insulator thickness and get the "m" value.
Click to expand...
Thanks, thats 1.5e-4/1.52e-4 = 0.987degC/W.

...yes indeed....Better to use metal tab + spacer than F type FET.
Can you refer to a TO247-F MOSFET on the market?
Click to expand...
....thanks, i cant find one...just looked.....no TO247 F type FETs exist...just looked on digikey over all 650-750V TO247's....that seems to answer the qu?.......F type must be worse than using spacers
 

Upvote 0 Downvote
To meet safety specs the F type must have at least 1mm of homogeneous insulation between tab and ( earthed) heatsink. And they don't like to spend money on thermally conductive plastics ....
 
Upvote 0 Downvote
cupoftea said:
Hi,
Why is it that all TO247/TO220 SMPS switching FETs have the metal tab connected internally to the drain? Surely the back tab should be connected to the Source, or not connected to anything electrically….this is because the heatsink would never be willingly connected to the drain, as it’s a switching node…and you woudlnt want your heatsink propagating the switching node?

So why cant you find FETs with tab connected to source?...or are there some?

Example of Drain connected to tab...
Click to expand...
Hello. Let me share how the TO247/TO220SMPS can be connected. It is a step-by-step guideline that might be very helpful to you.
 
Upvote 0 Downvote
Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top