Beogram 4002 DC Motor Restoration: A New Tool on the Bench and the Importance of the Orientation of the Top Bearing

I recently received another DC platter motor from a Beogram 4002 for restoration. This shows the motor as received:

After removing the outer housing it was immediately clear that this motor had already received some 'attention'. The scratches and marks on top and on the tabs of the bearing assembly indicated that the top bearing had been removed once already:

This posed an issue since I know from painful past experience that if one bends these tabs twice they can easily break off, and this usually relegates the motor to the spare parts bin. After I opened the motor up I also found mangled brushes. Luckily the feedback pickup coils were not damaged.

So I proceeded and extracted the bottom bearing and submerged it in oil and pulled a vacuum. Immediately bubbling started indicating that the bearings most likely had not been oil infused during the previous repair attempt.

So I added the entire top part of the motor housing into the oil in the hope that the top bearing would also infuse while still installed. A lot of bubbles came out indicating that there are some nooks and crannies behind the magnets etc...:

After about 3 days no more bubbling could be seen and I removed the parts from the oil:

\

I waited a day to let as much oil as possible drain from the motor housing. Then I ultrasonically cleaned the housing a few times in water with detergent, followed by a course in isopropyl alcohol to remove the excess oil from the housing. These motors cannot run well if there is oil 'sloshing' around on the inside. The only good place for oil is inside the porous bearing material from which it can slowly leach out onto the shaft as it rotates.

The next step was to straighten out and clean the brushes:

After this I re-assembled the motor for the first time and tried to run it. No cigar, though. All that happened was that my bench supply went straight into current limiting mode indicating a short circuit. This is usually indicative of one or more shortened spark snubbers on the rotor coils.

I opened the motor up again. This shows the rotor with the original spark snubbers. They are the three wired 'blobs' on top of the windings (one is invisible behind the commutator, one has a faint red dot on it)

These can be replaced with modern bi-polar surface mount TVS devices:

One can bend their contact tabs out a bit, and then they fit neatly between the rotor poles:

If you try this at home, note that the TVS packages need to be below the commutator assembly. Otherwise they can interfere with the brushes when the motor rotates (an ugly noise!). After this surgery, I put the motor back together and installed it in one of my Beogram 4002s for an RPM stability test with my BeoloverRPM device:

It allows logging the RPM in 10s intervals for long periods of time. The blue curve in the graph below is the curve that I measured. While basically stable, there were ugly spikes indicating some occasional spontaneous RPM increases. Not very beolovely!

And this is when invention happened! I knew from past experience that maintaining the original rotational orientation of the top bearing can help reducing the number of problematic RPM stability measurements. This is why I started marking the orientation of the top bearing relative to the housing with a sharp needle that I can install in the exact same orientation after the oil infusion. This practice essentially eliminated issues with RPM stability after the restoration process. 

My hypothesis about the necessity to keep the bearing orientation is that these bearings wear unevenly over time due to the pull of the platter belt. This lateral torque on the shaft probably results in the formation of a 'more polished' area towards the direction of the platter center.

I think, one can actually sometimes weakly observe this polished area in some bearings when shining light through them.

If the bearing is oriented differently after the re-assembly a more rough area is faced by the shaft in the direction of the belt-pull, and this seems to cause some weak inconsistent friction changes that can occasionally throw off the motor control circuit, inducing an overcompensation that causes the observed spikes. In my experience these spikes are always to higher RPM. I also observed in the past that these spikes get slowly weaker after running the motor for many days, supporting the theory of the worn-in region on the bearing. I am actually wondering if these Beograms had a similar burn-in period when they were used for the first time when new. The spikes are probably too weak to be audible, and 1970s RPM measurements involved using a test record with a frequency analyzer, i.e. the measurement integrated over time to some extent and may have missed such fast variations. Theories...;-)

Anyway, the issue at hand was that the original orientation of the top bearing of this motor was unknown due to the previous repair attempt. And I did not want to run it for 2 weeks in the hope to get rid of the spikes.

So I settled for the design of a 'burn-in tool', which would enable an accelerated wear-in process. This is what I came up with:

I had a 1 inch ball bearing laying around and so I designed this jig with a 3D printed pulley that fit on top of the bearing and a holder for the motor in an appropriate distance to allow for a decent pull exerted by the rubber band, which came from a bunch of organic green onions that my wife prefers to buy. I estimate the pull is a few times stronger than the platter belt, and the rotation (at 4V) is maybe another 3x compared to 33RPM in-situ. This estimate comes from the fact that the platter RPM is usually about 90-100 rpm if the pickup coils of the motor are damaged, which eliminates the RPM control and the motor runs as fast as it can.

I ran the motor in this setup for about 4 days, and then I installed it back in the Beogram and measured another 24 hrs RPM stability curve. And what I measured is the red curve in the graph above. 

So it seems this approach may work if the bearing orientation is unknown prior to restoration. This motor is ready for duty again!