Hi, folks:

I'd like to tap you folks' knowledge, for this is not something I do
every day.

I have a failed uPA1478H (four element silicon transistor array), but
only one-fourth of it is bad

I am repairing a Subaru ECM, that presently does not provide a ground
for the Check Engine Light (CEL). I have determined that the base drive
is present, but the failed junction does not turn on to complete the
lamp ground.

As I don't see an easy source for a replacement uPA1478 (tell me if you
find one), what I'd like to do is sub a discrete transistor for that
element, preferably one I already have in my junk bin -- I have a pretty
decent junk bin.

Datasheet picture (PNG format, with my scribbled-in circuit notes) he

http://www.ncplus.net/~asavage/Subaru/uPA1478_01_.png (77k)
http://www.ncplus.net/~asavage/Subaru/uPA1478_02.png (37k)

Or, if you prefer JPEGs:
http://www.ncplus.net/~asavage/Subaru/uPA1478_01_.jpg (147k)
http://www.ncplus.net/~asavage/Subaru/uPA1478_02.jpg (74k)

(Original NEC datasheet:
http://www.necel.com/nesdis/image/IC-3566.pdf (59k) )

As you can see, the purpose is for 5v at the base (pin 8) to complete
the lamp's ground (connect pin 9 collector to pin 10 emitter/ground) --
this is the OEM circuit. I don't think that I need the zener included
in the original design, there's no inductive load.

I do not know how much base drive current I have available beyond what
the uPA1478 requires.

Suggestions?

--
Regards,
Al S.

On Mon, 8 Sep 2003 02:22:07 UTC, "Al Savage" wrote:

Find the specs at http://www.nteinc.com/specs/2300to2399/pdf/nte2338.pdf

A cursory look through my junk bin yielded two D2018 parts, which cross
to NTE2338, so I'm all set. Soldered in, testing tomorrow.

The 2SD2018 is working perfectly. Again, I thank you.

--
Regards,
Al S.

This OS/2 system ("Tori", W4 FP15) uptime is 8 days 10:18 hours