The thermal mismatch could be a problem, but I think a far bigger
problem is the thermal mass you need to heat up in order to reflow
the solder. To reflow the part, you must feed it enough energy to
bring it up to the melting point of the solder. A larger thermal
mass needs more heat to reach the same temperature as a smaller part. 
Therefore, you will need to either: 1. Slow down your conveyor, 
possibly keeping the part in other stages of the reflow process too 
long; or 2. Apply a higher temperature in the reflow section of the 
oven. If you have other parts on the board, they will get damaged. 
The board itself may burn.

If you really have to reflow the body of this part  (and I wouldn't
recommend it), try putting the one part on, reflow it at a high 
temperature, then place the componentsand reflow it at the regular 
temperature. Keep in mind this thermal mass may draw heat away from 
nearby components.

Alan Hilton
Emulation Technology

------- Forwarded Message Follows -------
Date:          Thu, 27 Jun 1996 07:23:42 -0700
From:          Eli Dallal <[log in to unmask]>
Organization:  Tadiran Telecommunications Ltd.
To:            [log in to unmask]
Subject:       TO220 reflow soldering

Dear Techneters
For years we have been reflow soldering TO 220 transistors to ceramic 
substrates to provide good thermal dissipation . now we want to use FR4 
boards . Our intention was to bend the TO220
leads and insert them in pths and reflow solder the body  of the 
transistor to the board on a suitable land and later wave solder the 
leads . But I was told that in pcbs it is customary to use screws and 
nuts or eylets instead of soldering .One explanation was that there is a 
big thermal mismatch but no one seemed to know for sure .
What I would like to know if anyone has any experience in reflow 
soldering TO220 s to pcbs and if anyone knows why  it should not be 
done.
 thanking you 
Eli Dallal
process engineer 
Tadiran telecommunications

===========================================================
Alan Hilton, P.Eng.         Phone: (408) 982-0660 X213
R&D Engineer                Fax:   (408) 492-9170
Emulation Technology, Inc.  eMail: [log in to unmask]
2344 Walsh Avenue
Santa Clara, USA 95051-1301