Beogram 4002 (5513): Restoration of the DC Motor - Oil Infusion of Bearings, and Repair of a Broken Off Brush and a Disconnected Sensing Coil

I recently received a package from the Netherlands containing the DC motor from a Beogram 4002 (5513). This motor tested 'dead' after I took it out of the box, so I opened it up. What I found was a broken off commutator spring, which was taped on with 'electrical tape'

No surprise that it would not turn in this condition...Further inspection yielded that also one of the sensing coils had a broken off lead (when everything is o.k. the coils have about 22 Ohms resistance between the contact terminals):

This disabled the feedback mechanism of the motor, i.e. even with intact brushes it would have spun at about 100 RPM instead of 33.33. 

This shows the motor completely disassembled with bearings taken out:

I set out to repair the damage. Luckily I was able to solder the broken brush

back together:

Then I tacked the broken off part of the coil lead back on:

The next step was to do the oil infusion of the oilite bearings. This shows the bearings the moment I put them under vacuum:

The bubbles indicate escaping air from the porous brass material, which at that moment is also replenished with motor oil. After 24 hrs the bubbling had stopped

indicating that the bearings were full with oil again. Then I put the motor back together:

and gave it a 20 hrs test with my BeoloverRPM device. The picture shows my recently redesigned enclosure, which is more practical to use and also more easily 3D printed:

This is the 20 hrs RPM performance graph that I obtained:

Most of this curve is pretty much within spec (±0.05%), but a few spikes that go to about ±0.1% are visible. Unfortunately, this seems to be pretty standard for many of the DC motors in the 551x/552x type series. It may even be an issue with the analog motor control system itself (which comes in many component variations indicating that B&O themselves were experimenting with getting these motors to run more stable). I am still working on figuring this out, which is tricky due to the intermittent nature of the phenomenon. The good news is that these variations are very small. Research shows that most people can identify pitch variations starting at about 0.7%, i.e. this is well below the detectable threshold, and therefore this Beogram will be able to deliver a satisfying listening experience once this rebuilt motor is back on the job.