Littelfuse about to acquire IXYS

[Eva]

Finally some Mr. Slim-Product acquires Mr. Fat-Silicon. Some of us have been anticipating this for a long time.

Littelfuse will acquire IXYS for $750 million

IXYS has a tradition of producing fat silicon dies for >=600V applications, despite this has never been an optimum solution. The TO-247 cases or bigger result in limited switching speed, which translates into reduced high load efficiency. The huge die capacitances result in reduced low load efficiency in most cases.

The transaction is probably related to the fact that an important portion of current IXYS product portfolio, comprising silicon FET rated >=600V, is imminently being rendered obsolete by SiC, GaN and GaAs technologies.

IXYS parts have a tradition of being available, in first place, in semi-custom way for direct sale to big/exotic equipment manufacturers, and only in second place, through component stores, for mid/small manufacturers and for technicians repairing equipment.

I hope this transaction does not result in a massive drop of old parts... But anyway, many of these old fat IXYS parts can be replaced by something from other brands performing better in some field or costing less.

 

[kASD]

Their GEN-X3 class power 250V mosfets in TO-220 package are very attractive for class-d hard switched application. My sources told me that the new generation parts will remain in production and old generation parts are bound to vanish for sure.

 

[Eva]

250V parts:
- IPP220N25NFD: released 2 years ago, halfway between IXFP60N25X3 and IXFP80N25X3, similar cost, around $3/1k pcs.
- IRFB4332: released 9 years ago, 20% worse body diode, 50~100% worse Rds-on, optimum cost, around $1.8/1k pcs.

In comparison, 200V part:
- IRFB4227: released 11 years ago, 20% better body diode than new 250V parts (still best 200V), 15~50% worse Rds-on than newer 250V parts, optimum cost, around $1.5/1k pcs.

A part designed to be optimum remains optimum as long as universe constants do not change.

A TO-220 can take 25W without resorting to laborious mounting and heatsinking techniques. 2 pairs per half-bridge are OK for <200V, bridgeable down to 2 ohms. For 250V with new devices 2 pairs are also OK, but at the cost of the 3 pairs that would be needed with older FET. In other words, the 25% increase in voltage translates into both 50% increase in cost and output power into low impedance. It is still reasonable to make "hi-volt" and "hi-curr" versions, at same cost 😀

 

[davidsrsb]

Littlefuse are more widely distributed than IXYS

 

[Reactance]

Finally some Mr. Slim-Product acquires Mr. Fat-Silicon. Some of us have been anticipating this for a long time.

Littelfuse will acquire IXYS for $750 million

IXYS has a tradition of producing fat silicon dies for >=600V applications, despite this has never been an optimum solution. The TO-247 cases or bigger result in limited switching speed, which translates into reduced high load efficiency. The huge die capacitances result in reduced low load efficiency in most cases.

The transaction is probably related to the fact that an important portion of current IXYS product portfolio, comprising silicon FET rated >=600V, is imminently being rendered obsolete by SiC, GaN and GaAs technologies.

IXYS parts have a tradition of being available, in first place, in semi-custom way for direct sale to big/exotic equipment manufacturers, and only in second place, through component stores, for mid/small manufacturers and for technicians repairing equipment.

I hope this transaction does not result in a massive drop of old parts... But anyway, many of these old fat IXYS parts can be replaced by something from other brands performing better in some field or costing less.

Thanks for the heads up eva. IXYS parts are very expensive as well!