DEC 3000/600 Alphaserver problems

Peter Coghlan cctalk at
Wed Jan 12 17:39:52 CST 2022

On Wed, 12 Jan 2022 at 19:30:58 +0000, Peter Coghlan via cctalk wrote:
> On Wed, 12 Jan 2022 at 11:02:47 -0600 Jon Elson via cctalk wrote:
>> On 1/11/22 5:17 PM, Peter Coghlan via cctalk wrote:
>>> Thanks Jon.
>>> Here is a badly focused picture of the suspect components:
>> WOW, I don't know what those are!  They look like big 
>> ceramic caps to me, but if they don't have any polarity 
>> marking, then they can't be Tantalum.  Big multilayer 
>> ceramics are prone to cracking due to board flexing, and 
>> that can cause catastrophic failure, too.
> Thanks again Jon.
> They do have spikes of solder sticking out from the positive ends
> which can be seen in the picture.  There are no circuit references on
> the I/O board where the picture was taken.  However there are markings
> on the system board including little + signs at the same ends as the
> solder spikes on similar capacitors on that board.
> I managed to desolder one of them from the middle of the picture using
> two soldering irons.  I am not very good at this surface mount stuff :-(
> It came out ok though.  I must have got lucky and picked a good one
> because it measured 47uF on the capacitance range on my multimeter.

I looked around some more with the magnifying glass and found the tiniest
crack in the body of one of the smaller orange components, the vertical one
in the middle of the picture.  I desoldered it and measured it.  It was
9.73uF and had 35.7 kOhms leakage (the 47uF one had no measurable leakage).
It is not one of the three I suspected initially and it has 12V across it
in operation, possibly associated with the nearby AUI connector.  I presume
it is a 10uF unit and despite having released some magic smoke, it seems to
be continuing to function reasonably well.

I guess failure of these sort of capacitors is probably not anything to do
with the cache failures I am getting then.

Time to get back to the power supply I think.


>> Jon

More information about the cctech mailing list