Beogram 4002 (5503): Intermittent Tracking, Installation of New Carriage Motor and Restored Keypad and Update to Latest Beolover State-of-the-Art

I recently received a Beogram 4002 (5503) that I had restored in early 2020. After running for a while it came back in August 2023 with a fried H-Bridge. The H-Bridge is responsible for driving the DC carriage motor. It allows reversing the current direction even if there is only a single positive voltage rail in the system. It does that with transistors that act as switches to reverse the polarity on the motor leads. These transistors need to be able handling the motor current. If everything goes well, the current capacity of the transistors is safely above the current drawn by the motor in all operational situations.

Electric motors draw more current when they run under load since they run slower at the same applied voltage. The slower RPM causes the motor back-EMF (the self-induced voltage in the motor, which is opposed to the driving voltage) to be lower, so more of the applied voltage arrives at the motor coils, which in turn increases the current in the motor windings. This is a great feature of electrical motors since it means maximum torque at zero RPM, great for burning some rubber at a green light to impress ICE vehicle drivers (whose torque-to-RPM ratio is reverse: Minimum torque when the car is at rest and maximum torque when it is at speed...;-).

But this also means that if a mechanical system that is driven by an electrical motor develops more friction over time due to hardened lubricants or dry motor bearings, the motor will draw more current than when the system was new.

I finally came to realize over the years that this may be the root cause for H-bridge failures in Beogram 400x. The H-bridge transistors are able to handle currents of 1 Amp max. And when everything is according to spec fast forward or reverse typically draws about 0.1-0.2 Amps. 5x is a decent safety margin. But when there is additional friction, the current can get dangerously close to 1 Amp or even exceed it. This is evident from the often found blackened PCB surfaces under failed H-bridge transistors.

The main reasons for this issue are the carriage motor itself and hardened lubricants in the carriage translation mechanism. The latter can easily be addressed by cleaning and re-lubricating, but in the case of the motor replacement is necessary. In difference to the later DC platter motors, these motors cannot be rebuilt easily. This was the reason I designed a replacement motor that draws less current than the original motors at a similar torque. Due to its modern design it also creates less vibrations and noise compared to the original motors. This post describes an evaluation of my design.

So I am not really surprised anymore that this Beogram came back again with carriage drive issues considering the earlier H-bridge issues. This time it sometimes stopped tracking. An indication that the carriage motor developed even more friction in its bearings since the last visit. 

Therefore, the first step was to replace the carriage motor with the new Beolover Carriage Motor for Beogram 4000, 4002, and 4004:

This shows the original motor in place:

I extracted it and opened the enclosure up:

Then I installed a new Beolover motor:

It is a bit shorter, which permits feeding the leads through the hole in the enclosure bottom for convenient routing. This shows the motor in place:

My customer also wanted me to update the main capacitor setup and the wasteful linear regulator based 22.8V power supply. This shows my work of 2020, when I still used big radially leaded capacitor cans similar to the original setup:

I removed this setup and installed a Beolover Efficient 22.8V Power Supply and Main Capacitors for Beogram 4002 (Types 550x) board:

This board solders directly to the existing leads that previously connected to the big capacitors. The board also replaces 0TR1 (to the right of the platter motor in the picture) that regulates the 22.8V rail in the original setup. My board uses a modern buck converter to create the system voltage from the rectified DC coming from the transformer, which is much less wasteful and results in a cooler running and less energy using Beogram.

Looking at my notes from 2020, I saw that I did not de-magnetize the solenoid plungers yet as a standard restoration item. These plungers often get magnetized to various degrees, which can result in sluggish arm lifting. During auto-return at the end of a record this can cause the tip to drag over the platter for some distance before it finally lifts up when the spring overcomes the magnetic attraction. For demagnetizing the plunger the solenoid has to be extracted and then the plunger unscrewed from the angled bolt that connects the solenoid lever. This shows the solenoid in place:

Indeed the extracted plunger was magnetic. I usually test this with a ferrous set screw. If the plunger attracts the screw it needs demagnetizing:

After using my tape head demagnetizer on it a bit the screw was not attracted to the plunger anymore, 

so I put everything back together.

I also found that the sensor arm LED replacement was still one of my early versions based on a home etched flex PCB that I folded into the small bulb compartment:

This shows the extracted old part in comparison with my current approach, based on a small PCB and a 3D printed alignment aid:

I installed the Beolover Sensor Arm LED Light Source (for Beogram 4002, 4004 and 4000):

This shows it in action. It uses a warm white LED that has enough red photons for lighting up the B&O logo in a realistic warm red:

Next came the replacement of my original early design transport lock bushings. The one on the left in the picture below shows my original design, which has a much wider wall thickness. Over time I came to realize that it makes the adjustment of the sub-chassis much easier if there is a bit more room around the lock bolts. So I re-designed the bushings with thinner walls (shown on the right): 

An added nice benefit of these new bushings is that the sub-chassis can move much more before it hits the lock bolts. This gives gives it a much more supple feel when the platter is touched and just freely swings. This shows one of the bushings installed:

Their two halves are simply pushed in from the top and bottom, which makes installation very easy. This shows the liberated chassis during the installation of the bushings:

The final update was to replace the smudged keypad with a new Beolover replacement. This shows the original keypad:

Most of the 4002 and 4004 keypads develop such use traces over time. First it starts with such 'smudges', basically polished areas caused by friction with the fingers when operating the Beogram. As time goes on the coating wears fully through. Luckily co-Beolover Beomazed recently succeeded after a long time of trial and error with reproducing the keypad plates. His plates are completely new and carry a modern resilient coating that promises to last for a while. This shows the restored keypad:

Absolutely stunning. They really look like the original pads! Read here how he does it!

If you are interested in getting your keypad restored, please visit here.

I installed the renewed pad in the Beogram. This is how this beauty looks now!:

And then it was finally time to enjoy this fully Beolover state-of-the-art restored Beogram with one of my favorite records by Chico Hamilton: "chic, chic, chico", which he recorded in 1965 on Impulse! I have the stereo version AS-82. He looks really chic on this one!...;-). A stunning record that has just the right amount of 'avant-garde edginess' as one would expect from an Impulse! record! A great match for this beautiful Beogram 4002!:

I will now play a few more records on this deck and then it will be time to send it back to its owner in California!

Beogram 8002: Full functional Restoration and a Test Spin with Cannonball Adderley

This post discusses the functional restoration of the Beogram 8002 that I received from a customer in California a while ago. This post discusses my initial assessment of the unit.

When I work on a Beogram 8002 or 8000 I usually extract the technical parts from the enclosure:

This makes it much easier to work on them. This unit had some mechanical issues with the carriage. The first step is always to extract the parts that guide and move the carriage for cleaning:

A 30 min run in my ultrasonic removed the original lubricants and the parts could be reinstalled. At this point I also adjusted the tone arm parallelism since it is necessary to have the carriage up for the vertical adjustment.

Next I focused on rebuilding the PCB. Often the electrolytic capacitors are out of spec or even short circuited. Another common issue are bad solder joints on the board headers. This shows the board in its original condition:

I removed the microprocessor can and replaced all the capacitors:

I also added small 'hooks' to pins 4 and 6 of P5:

They are useful for connecting a multimeter for doing the <</>> button adjustments later on.

The perhaps trickiest part of the board restoration is replacing the decoupling capacitor (C28) that stabilizes the power supply of the microcontroller. Such capacitors need to be placed close to the to be decoupled device and therefore it is located inside the can:

I opened it:

The capacitor is found after lifting up the small additional board that is inside:

This shows the solder points from the bottom:

The tricky part is that the ground lead of this capacitor is also used as a ground via, i.e. is soldered to pads on both sides of the board. 1980s technology is fun!...;-). Since there is not much space around this capacitor it needs to be destroyed for removal. I usually unsolder the 5V lead first and then pull it upright. The next step is rotating the capacitor until the ground lead breaks and it can be removed:

With a desolder pump the remaining fragment of the lead can be extracted and the pad cleaned on both sides. Installation is reverse. With a small tip solder in the ground lead on the component side and then on the solder side, followed by soldering the 5V lead:

After this operation I closed the can back up and re-installed it on the baord:

Then I reflowed all the board headers. Especially the one that connects to the transformer block is often affected by broken out solder joints, which causes intermittent power to the deck. I usually ad a bit of solder to those points to make them stronger.

The next step was replacing the output relay, which is often oxidized. It is mounted under a shield inside the DIN7 output connector assembly: 

The shield is just clamped to the relay and can be pulled up easily:

I replaced the relay with a new Omron relay with the same pinout:

Nowadays such relays are fully enclosed, which should make them last longer. The relay I used has the same lateral dimensions like the original relay, which makes it easy to put the shield back on:

In many cases there is a hum when connecting the 8002, espeically to RCA type inputs. This hum can be quenched in most cases by connecting system and signal grounds. I usually install a switch on the output assembly that allows choosing whether to connect the grounds or not. It is convenient to solder it directly to one of the big GND pads on the PCB and make the signal ground connection with a wire:

There are two more capacitors (OC1/2) that need replacement. They are directly soldered to the voltage regulators that use the floating chassis as heat sink:

This shows the new caps in place:

One more in the transformer block:

This capacitor is responsible for the phase shift necessary for the second motor phase. Since the motor runs on AC this one is a non-polar electrolytic capacitor. It can be replaced with two back to back polar units, like I did it in earlier days, or with the more modern Beolover Motor Capacitor for Beogram 8000 and 8002, which is designed based on modern SMD ceramic capacitors. Those do not care about polarity and so are the perfect choice for this application. The board is set up that it can be used in both 50 and 60 Hz configured Beogram 8000 and 8002, which need different capacitances for the proper phase shift. In this case the unit is a US 110V/60Hz unit and so I used the smaller 27u capacitance:

After closing up the transformer box I focused on replacing the light bulb that enables the pressure sensitive gradually adjustable << and >> carriage speeds with an LED. LEDs emit light in a more consistent way over time, which is important in this case since the light intensity permitted onto the photo resistors mounted on the sides of this assembly governs the carriage speed. The bulb is inside the black box on the board:

This shows the board from the bottom. From this perspective one can see the seesaw metal aperture component that is pushed down on either side with the <</>> buttons to change the amount of light on the photo resistors:

Removal of the top cover of the box reveals the light bulb and the two photo resistors:

I usually replace the bulb with a 5mm white LED that I sand a bit to diffuse its light emission:

Since LEDs run at a lower voltage than the bulb it needs a current limiting resistor, which can be conveniently added on the bottom of the assembly:

Then I plugged everything back together and adjusted the <</>> voltages by using the two hooks I soldered to P5 when I rebuilt the main PCB:

The two screws on the <</>> assembly need to be turned in or out until 620mV are achieved on the respective P5 pins.

After adjusting the tracking feedback to yield carriage movement after about 2 turns of the platter it was time to play a first record. I selected one of my all time favorites, "Somethin' Else" recorded by Cannonball Adderley together with an illustrious group of contemporaries including Miles Davis, Hank Jones, Sam Jones and Art Blakey in 1958. I currently only have the Blue Note 75th Anniversary Reissue (2014), which was (sadly) cut from digital hi-res transfers of the original analog tapes. Still sounds pretty good on this Beogram 8002 in 'service position'..;-):

The scratched hood is on its way to Beomazed, who figured out a way to transfer the original 8002 trim to new hoods from dksoundparts. Stay tuned!