Beogram 4004 DC Platter Motor Restoration Using Dksoundparts Bearings

I recently received the DC platter motor of a Beogram 4004 from a customer in the Netherlands. This customer wanted me to install new reproduction bearings he had purchased at the dksoundparts store.

Normally, I re-infuse the original bearings with oil under vacuum. So it was an interesting experiment for me to see if there is a difference in outcome. 

This shows the motor as received:

I took it apart as usual:

The original bearings are on the black pad up front.

The new dksoundparts bearings came in a small metal can:

I took them out of the can. They are oily since they were also infused with oil under vacuum:

This shows them installed in the top and bottom parts of the enclosure:

If you try this at home, make sure the larger bearing goes into the top part and the smaller into the bottom.

After closing the motor up it was time for a 24 hrs RPM stability test with the BeoloverRPM device:

In its 'slow' mode the BeoloverRPM sends an RPM measurement every 10 sec to a computer serial port.

This is the curve I measured after about 24 hrs:

This curve looks a bit 'wilder' than the curves that I get for most of my restorations using the original bearings (see for example this restoration project for a comparison). This does not really come as a surprise to me since over time I came to realize that it is important to re-install the original bearings in the same orientation as they were installed originally. This significantly shortens the time it takes to break in the bearing. It seems the bearing surface gets polished around the main contact area defined by the pull of the platter belt, which over time lessens the amount of RPM fluctuations. 

When installing new bearings there is no already polished area, so they always need to get broken in. I expect this motor will probably clam down after playing it for a few hundred hours. I suspect that new Beograms back in the 1970s had this issue, too, but no one noticed it since these variations are well below the threshold humans can detect.

I thought it would be interesting also measuring wow and flutter (short term RPM variations introduced (mainly) by the feedback loop that keeps the RPM constant in DC motor Beograms).

This can be done in 'fast' mode with the BeoloverRPM.

This is the curve I measured during about 60 platter rotations:

In fast mode the BeoloverRPM logs a RPM measurement every time a platter rib passes under the sensor. In other words one gets 24 measurements per rotation. The above graph looks 'complicated' since the actual RPM variation measurement is superimposed by a measurement artifact caused by minute variations in spacing between the platter ribs around the platter. Since the time between ribs passing is measured and then converted into an RPM value, changes in the spacing results in proportional changes of the calculated RPM. This causes the observed zig-zag pattern (the 'platter pattern'), which is added to the actual RPM variation. The true RPM variation in turn corresponds to the slower wave pattern that is superimposed onto the platter pattern. An evaluation of the graph seen here suggests an RPM variation of maybe 0.2%. 

When using the original bearings I usually get something around 0.1%, i.e. a slightly better performance. This probably also has to do with the rougher surface inside the new bearing that challenges the feedback loop a bit more causing more variation.

It would be very interesting to measure this motor again in a couple years after this Beogram played a few hundred records. I would expect to see a better performance.

After this experience, I think one can say that it may be better to use the original bearings if possible. But the new bearings available at the dksoundparts store are a viable alternative if a vacuum setup is not at hand. It would be interesting to know how the results with new bearings compare statistically, i.e. after doing a larger number of motors with them. What I know from my restorations with re-infused original bearings is that in a small number of cases (<5%) the restoration results can be less than satisfactory.

In general, this discussion is pretty academic since even 0.2% variations are much smaller than the 0.7% threshold where humans begin noticing pitch variations.

In summary it can be concluded that this motor is ready for duty again! Time to send it back to the Netherlands!