Replacement for QP1008 (HP9816 PSU)

Tony Duell ard at
Mon Apr 14 17:24:36 CDT 2008

> Rik wrote:
> > The PSU isn't build by HP but by RTE POWER/MATE a firm with doesn't exist
> > any more.
> That explains it.  If HP bought the PSU as an assembly, there wouldn't
> be HP part numbers for its components, as it would be considered a single
> FRU.

Is that _always_ the case? I think I've seen a Tandon TM100 drive where 
the logic board is identical in layout and circuitry to a standard one, 
but is gold-plated (like most HP boards of the time), has an HP part 
number in the etch and 1820-xxxx chips on it. I would be very suprised if 
HP made that board themselves.

> >From Tony's schematic, I'd guess that it's a semi-custom part.  Motorola
> probably binned one of their standard parts to a higher spec for RTE.

And alas we have no idea what it needs to be selected for, so there's no 
point in making measurements on a good PSU or its transistor.

> Perhaps you can look at schematics for other switching PSUs
> with similar ratings and determine what transistor they use.  The
> safety-critical parameters are breakdown voltage and maximum current.

Too low a current gain may mean the transistor operates in the linear 
part of its chracteristic, and promptly overheats.

> If the PSU is rated at less than 500W, I suspect you could use a BU208A,
> which costs less than $3 from Digikey, but don't take my word for it.

I am almost sure it's less than 500W. It's quite a small supply. 

A BU208A is a good suggestion. As is a 2SC1942. And there are many 
others. The problem is it gets expensive if one fials after another. And 
you don't know if they're failing becuase they're unsuitable or because 
the original fault, the one that caused the original transistor to fail, 
is still there.


More information about the cctalk mailing list