transistor package

[anton]

hi everybody, what's the difference between the packages used for transistor, why some transitor use TO-3P or TO-220 etc.. what is the meaning? i really appreciate an answer.

[poobah]

Semiconductors come in different packages for a number of reasons.

TO-3s were once the standard for medium power devices. Now, To-220s and TO-247s have largely taken over.

The newer devices are generally easier to mount and integrate into a circuit board.

When you see the same device offered in different packages, the thing to look for is the thermal resistance from junction to case... R_theta_j-c. This is usually expressed is Degrees C per Watt (C/W).

In themal calculations, mutliply the power by resistance to yield the temperature difference (as you would multiply current times resistance to find the volatge difference.

As a general rule, packages with a lower R_theta will keep the silicon junction cooler.

Designers usually choose the smallest package to save money. But this is often limited by the need for cooling... so R_theta is ONE of the factors used in choosing a device package style. As always... a comprimise is involved.

At the end of every device data sheet, you will generally find the package dimensions. You will usually find R_theta toward the front. Occasionally you must look for package dimensions elsewhere.

[jackinnj]

Originally posted by poobah
Semiconductors come in different packages for a number of reasons.

TO-3s were once the standard for medium power devices. Now, To-220s and TO-247s have largely taken over.

A TO-3P is more akin to a TO-247.

[poobah]

yep... and ON-Semi tried some years ago to do a TO-264 device... bigger than the 247s. I haven't followed it though... don't know if they succeeded or not...

[Nisbeth]

Originally posted by poobah
yep... and ON-Semi tried some years ago to do a TO-264 device... bigger than the 247s. I haven't followed it though... don't know if they succeeded or not...

I don't know if it's a success but they still have it Going from memory, MJxxxx is TO3, MJLxxxx is TO-264 and MJWxxxx is TO-247.


/U.

[MikeB]

Originally posted by poobah
yep... and ON-Semi tried some years ago to do a TO-264 device... bigger than the 247s. I haven't followed it though... don't know if they succeeded or not...

They definitely succeeded, the TO-264 package is the standard package for OnSemi power-bjts. All MJLs are TO-264, the MJWs (TO247) are seen rarely.

Mike

[cliffforrest]

In the early days of transistor manufacturer, the 60's (I worked for UK Plessey, then Fairchild) a hermetic metal can was the ony way of ensuring a reliable device.

From the intraduction of planar devices, the chip die was bonded (welded) to the inside top of the can (collector) and wire bonds added between the two isolated lead posts and th B&E pads on the die. Finally a "top-hat" can was welded over the whole thing. Expensive! Rows of girls doing fiddly things.

Epoxy with the right tempco so that lead seals didn't break came later.

Thus TO-18 (small can, eg 2N2222) and large can (TO-5) were standard. And expensive.

They came originally with long leads, c. 40mm, which could be wired directly between terminal strips (look at a Tek scope of the period. These were almost impossible to handle automatically and the leads of a bag of devices became all tangled up!

Power Devices were again metal can, with TO-3 (OC-28 (!) and 2N3055) with TO-66 as a miniature variety, much used in Mil equipment. Both these had large surface areas for cooling and could fit into sockets because of the 1mm dia B/E leads.

The drive for autohandling, pcb insertion and cost first produced small can devices with shorter leads then a variety of different "plastic" devices following the TO-18 and TO-5 footprint.

In-line leads were produced to allow taping from production through autotest to auto pcb insertion.

Industry being what it is there was little standardisation and every conceivable combination of EBC leads is to be found. Pity.

As stated above, plastic power devices are now the norm. Cost and handlng again.

[john curl]

T0-220 is very bad for heat dissipation. It has 1/4 the effective surface to transfer heat as a T0-3.
Plastic TO-3's are OK.

[AndrewT]

Hi,
not just ONsemi doing To264.
Toshiba's 2sa1943/c5200 is also To264.
Quite a few of the FETs out there are in that bigger case style.

It is an industry standard, unfortunately many call up To3P when referring to either the To264 or the smaller To247.

The major difference between the plastic packaged styles and the metal cans seems to be the lower Tjmax of 150degC for the majority of plastic types.