[time-nuts] New member, new old 4040A with PSU failure

[paul swed]

Patrick
Great job on the pix. Very helpful.
Interesting in many ways the 4040 appears newer in design then the 4050.
What is different is that in the 4050 the Cs tube multipkliers and such are
in a magnetic shield.
But the RF sections appear very similar.
Noted the ugly cap and you may have more evil ones to deal with. Or damage
to the board.
In the 4050s the supply was a big 28V linear lambda. Here I see its a
switcher. Again seems newer in design.
Note on the eevblog that the 28V supply generates a 32V supply. Thats
critical to operation. In the 4050 its a small multi transistor converter.
DC/DC boost on top of the 28V supply. That may be hiding under the bottom
cover on the module with some 8 screws. The caps inside made a mess of the
inverter as they leaked their contents on the board and it dissolved all
traces of the traces. Even under a microscope and various lighting it was
difficult to reconstruct the design. Unfortunately on that same board were
two other control systems and they were damaged.
However cleaning all the damage out and rebuilding the traces did allow the
control circuits to work. The converter was not recoverable. Solution
install, a 32V supply. Especially since I have no intent to run it on
batteries.
When I had most of it figured out, like you I simply used external supplies
and that showed the unit was operable.
Will be good to see what more you find.
Regards
Paul
WB8TSL


On Mon, Sep 28, 2020 at 4:25 PM Patrick Tanner <tannerpa at gmail.com> wrote:

> Have made some progress.
> Looks like bad caps. Capacitor "CA" is especially naughty.
>
> https://photos.app.goo.gl/Ye91Fv4Nmuj4xb7n8
>
> Pics and measured supply data:
> https://photos.app.goo.gl/dNWKyysC1QykCLWc8
>
>
> https://drive.google.com/file/d/1t1xjyNTbdwY_j7pg5rt_94GTY68R3ZbS/view?usp=drivesdk
>
> Thanks to Ed Palmer for the excellent thread over at eevblog detailing
> this.
>
> Patrick
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to
> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
>