Beogram 4004: Restoration of PCBs, DC Platter Motor and RPM Panel

I recently received the PCBs, the DC platter motor and the RPM panel from a Beogram 4004 for restoration. As usual I started with the platter motor since it can take a few days until the oil infusion of the bearings is complete. This shows the motor as received:

I disassembled the motor to extract the bearings:

The bearings are the two small donuts on the black pad upfront. I immersed them into synthetic oil and pulled a vacuum. Immediately strong bubbling started:

This bubbling represents air that is drawn from the empty pores of the Oilite bearing material, making room for oil to diffuse into the material.

While this process was going on, I focused on the other tasks of this project. First I restored the main PCB. This shows it in its original condition:

Here a close up shot of the RPM section consisting of a National brand RPM relay and the two RPM trimmers to its left:

I replaced all electrolytic capacitors, power transistors of the H-bridge and platter motor control and the RPM relay and trimmers:

This shows the rebuilt RPM section with a new Beolover National Relay Replacement for Beogram 4002 and 4004 and two encapsulated 25-turn 5kOhm RPM trimmers:

Next came the output PCB:

In the Beogram 4004 this PCB carries both the output relay delay circuit as well as the remote control circuitry that allows a Beomaster 2400 controlling start and stop of the Beogram via its remote control.

This shows the output relay circuit:

I replaced all the electrolytic capacitors and the output relay:

Here a detail shot of the output relay:

I also installed a (red) switch that allows connecting system and signal grounds in case there is a hum in the output signal.

Finally, I updated the RPM panel with LEDs. This shows the panel on its back which reveals the two bulb covers:

I removed the covers

and replaced the bulbs with Beolover RPM Panel LED Backlights for Beogram 4002 and 4004 (Types 551x/552x):

These little PCBs solder directly to the terminals where the original bulbs connected. They essentially become extensions of the original PCB:

The boards do not interfere with the bulb covers, which can be installed like before:

Now it was time to implant all the components into my bench Beogram. After bolting in the main PCB I replaced the two power Darlingtons mounted on the solder side of this board. It is best to replace these while the board is bolted in. This makes positioning a snap. This shows the original IC1, a TIP120, which is responsible for regulating the 21V rail:

I usually replace these ICs with stronger types. In this case a TIP102 is the perfect replacement:

In this circuit configuration the modern TIP packages need a 100nF capacitor soldered between the Emitter (output) and ground. This quenches a high-frequency oscillation that can occur after the replacement. I also replaced IC4, the solenoid transistor, with a TIP107 Darlington:

The oil infusion of the motor bearings had come to an end and I extracted them from the vacuum chamber:

I reassembled the motor and implanted it along with the PCBs and RPM panel in my bench unit. Then it was time for a RPM stability test with these components. This shows the BeoloverRPM device in action:

In its 'slow' mode it outputs the RPM in 10s intervals. Using a terminal program on a PC or Mac the RPM can be logged for extended periods of time for generation of a RPM stability plot. This shows the 24 hrs RPM stability graph I measured for this motor:

It is slightly choppy. This is a result of this motor having been opened before without noting the original orientation of the top bearing. If the bearing is not installed in the same orientation there may be a period where the shaft polishes a new segment of the bearing as it is pulled towards the platter by the belt. In my experience these small variations slowly go away after playing the deck for a while. At any rate these fluctuations are much smaller (~0.3%) than what humans can typically discern when listening to music (>0.7%). So this motor is ready for duty again!

I also measured 'wow and flutter' (the short term RPM fluctuations introduced by the feedback speed control system) using the 'fast' mode of the BeoloverRPM device. In the fast mode it logs the RPM after every passing of a platter rib. This reveals a pattern that repeats every 24 measurements, i.e. after each turn of the platter. It is a result of minute spacing variations between the ribs due to manufacturing tolerances. The wow and flutter RPM changes are superimposed to this pattern as a sine-wave like feature, which is normal for analog feedback-based control systems:

Evaluation of this pattern yields a wow and flutter number that is smaller than 0.1%. This is slightly larger than what the manual specifies (<0.05%), but this may well be a result of the different way they measured wow and flutter in the 1970s, where they used a test record and an analog filter based 'frequency analyzer'. The joys of analog audio!...;-).

In summary, this was a successful restoration project and these components are ready for duty again!

Beogram 4002: Failed Restoration of DC Platter Motor and Substitution with Beolover SyncDrive

I recently received the DC motor and the RPM panel of a Beogram 4002 for restoration. The unit displayed the usual RPM variations associated with dry running motor bearings.

In the end I was not able to restore this DC platter motor with a satisfying result. The reason may be that the motor seemed to have been opened before. The pulley was loose and the screws only lightly tightened. As shown below the motor still has RPM jumps around 1%, even after oil infusing the bearings and replacing the spark snubbers. My hypothesis is that the reason when I get such jumps is an mis-orientation of the top bearing after the oil infusion. When I open a motor I usually mark the bearing position and then put it back in exact the same orientation. This seems to increase the chances for a successful result.

My theory for the occurrence of the observed RPM variations is that over time the bearing gets polished in a spot facing the center of the platter due to the pull of the platter belt. If the bearing is re-installed in a different orientation the polishing process needs to start anew and this causes the speed variations. So when motors come in that were worked on previously it is possible that the bearing is already not in the original position anymore when I extract it. But I guess we will never know exactly with this motor.

Luckily, there is the Beolover SyncDrive DC motor replacement. It upgrades any DC motor Beogram to AC motor RPM stability with a simple plug-and-play installation (no soldering required):

I usually offer the SyncDrive at a discount whenever I am not able to restore a DC platter motor.

This summarizes the work I performed for this project: 

The motor as received:

I took it apart to remove the bearings:

The bearings are the two small donuts on the black pad upfront. I inserted the bearings into synthetic oil and pulled a vacuum. These bearings were very thirsty and the air bubbles drawn from them by the vacuum were so many that they caused the oil to foam before I could take a nice picture of the process starting:

After a couple days the bubbling stopped and I extracted the re-filled bearings from the vacuum:

Then I reassembled the motor and hooked it up to my bench supply. I measured a fluctuating current around 0.7-0.9 Amps at a voltage of about 1.4 V:

Way too high current at such a low voltage! The motor itself made strange pulsating noises and did anything but run smoothly. Sounded a bit like a one-cylinder farm tractor from 1938....;-). This was a clear indication of one or more spark snubbers having gone bad. They seem to occasionally short-circuit and with that one of the coils on the rotor is dead making a short circuit. Hence the high current and erratic behavior.

I set up to replace the spark snubbers with modern uni-directional TVS diodes. This shows the original spark snubbers still in place. They are the three silvery devices soldered to the poles of the rotor:

Next to the rotor on the black pad are the TVS diodes prepared for soldering in. I removed the spark snubbers:

And then soldered the TVS devices in-between the poles:

If you try this at home, the TVS packages cannot protrude into the plane occupied by the commuter terminals. Otherwise they can interfere with the brushes and/or the iron pieces of the rotor. Then I put the motor back together and tested it again:

Now I got the proper 0.2-0.3 Amps at 5 V. So I implanted the motor in one of my 4002s and measured the RPM stability over 24 hrs. This shows the BeoloverRPM in action:

In its slow mode it logs the RPM in 10s intervals, perfect for detecting RPM inconsistencies. The blue curve is the graph I measured after 24 hrs:

Sadly, the motor did not run stable enough to declare victory. The blue curve shows RPM jumps of the order of 1%, about 20x larger than what is tolerable according to specifications (<0.05%). These fluctuations may still be too small to be detected by the average human, but when listening to piano music or similar it may well be noticeable. 

So my customer decided to let me replace the motor with a Beolover SyncDrive. I added a typical SyncDrive RPM graph measured on the same Beogram (red) for comparison. This curve is as good as the curves I usually measure on AC platter motor Beogram 4002 and 4000. This will fix my customers RPM issues!

On to the RPM panel. This shows it as received:

And flipped around, revealing the bulb covers:

I removed the covers:

And replaced the bulbs with Beolover RPM Panel LED Backlights for Beogram 4002 and 4004 (Types 551x/552x). They solder directly to the solder pads used for the bulbs, i.e. are drop in replacements:

The LED boards do not interfere with the bulb covers, which can be installed the same way as before:

Then I tested the RPM panel in my 4002:

The scales are backlit in a nice warm sheen, exactly like an old fashioned incandescent bulb would do it!

All good again with this RPM panel!