Chip operating temperatures

[tom cloud]

somebody wrote:
>>Just think of the power dissipation from four coil drivers at say
>>40% dutycycle at 6000 rpm (dwell = 4mS/Wasted-Spark) and say eight
>>injector drivers at 75% dutycycle at 6000 rpm (Fuel p/w = 15mS/SEFI)
>>Typically, a 4 cylinder ECU can be sucking up to 10 Amps when the
>>engine's really revving, I don't have figures to hand for 8 cylinder
>>systems and of course all these figures depend on coil/injector type
>>and so on, but we're still talking about passing a lot of current
>>through the system. 120 watts can have pretty serious heating effects
>>on the box especially if it's mounted in an area of static airflow.

Krister Wikstrom <kwi at mamma.icl.fi> replied:
>I think the 120W is not dissipated at the driver location, most goes to the
injectors. If the driver sinks 10A
>and drops maybe 0.6V, the "warming" power is 6W - right? Using some low
RDSon MOSFETs etc. may have even lower voltage drop and power dissipation.
>
An' then I says:

I've not been reading this thread real thoroughly so I missed the
original post.  A switch dissipates *NO* power, but a transistor
is not a perfect switch.  An injector may take an ampere pulse, but
I think their holding current is less (don't know about this).
Like Kris said, assuming 1 amp: any good switching power transistor
will have a VCE(sat) (that's the voltage across it when it's ON) of
less than 1 volt, maybe .3 to .6 volts.  Switching FET might be less.
Anyway: the power dissipated at 1 amp would be 1 amp times .6 volts
or 600 milliwatts.  If one driver is pulsing several injectors,
the power would be proportionally higher.  The *good* thing about
this is that those transistors can be bolted to the case and the
heat will mostly go outside.  It's the VLSI and TTL chips that
get hot and have no direct path to the outside for that heat.
They probably do more to raise the internal temp than anything else.