Beogram 4002 (5503): Replacement of a Loudly Humming Transformer with a Modern Universal Power Supply

Oh well, after playing the lovely Beogram 4002 (Type 5503) that I recently restored I had to realize that the transformer made a pretty loud humming noise, which was quite noticeable during lower volume music passages. Especially, when placing the unit on a wooden sideboard the resonance of the wood amplified the hum to quite annoying levels.

I decided this definitely needed fixing before I could possibly sell this unit. A welcome project since I thought the initial crop of AC motor Beograms also needed a universal power supply upgrade similar to what I recently developed for the later DC motor models. 

Modern monolithic universal power supplies have several advantages over the original transformer/rectifier based setup: They are usually short circuit proof, they do not hum in any audible way and they mostly are designed to run on grid voltages ranging from 100V to 240V, i.e. can be plugged in anywhere on this planet. This feature also makes them brown-out resistant and they will also reliably protect the Beogram circuitry from voltage spikes etc...

The reason why my solution for the DC motor Beograms cannot be used in the AC motor models is these have a more than 3x higher power rating. DC motor Beograms are rated 15W, while the AC motor models are rated 50W. This difference is mainly caused by the AC platter motor, which is a power hog. This explains why the older Beograms get much warmer than the later ones.

It means that the 25W Meanwell supply I utilized for the DC motor Beograms is not powerful enough to run an AC motor model. Further investigation yielded that similar supplies rated for higher output had a too large footprint and would not fit into the space vacated by the transformer.

After a bit of poking around I found a more advanced design made by Traco that combined a small footprint with a 30W output: The TPP 30-D Series, which is a high-quality medical power supply. It is fully encapsulated and the output voltage can be adjusted (a rarity among this type of power block). The specific TPP 30-124-D (24V nominal) type can be adjusted to a maximum output of 26.4V by connecting an external resistor. This was great, but 30W is still not enough for the standard setup of an AC motor Beogram.

However, when combined with the Beolover Efficient 22.8V Power Supply and Main Capacitors for Beogram 4002 (Types 550x) 30W is enough. The Beolover 22.8V supply reduces the power consumption of the AC motor models significantly since it replaces the other power hog of this 1970s design, the linear voltage regulator, with a buck converter that has a high efficiency.

This shows the Transformer Replacement Kit for Beogram 4002 (Types 550x):

It includes the TPP30 supply mounted on the breakout board, the adapter necessary for bolting it to the enclosure bottom, the jumper for connecting it to the fuses assembly, the mounting bolts and a set of fresh fuses.

Let's see how I implanted the Traco TPP 30-124-D supply in this Beogram!

*********************************************************************************************************

Safety first: Potentially lethal voltages can be present during the installation of this part. Unplug your Beogram before installation. Ensure that the supply is bolted firmly to the adapter plate, and that the fuse box cover is in place before powering the unit up after installation.

*********************************************************************************************************

This shows the original setup:

The first step is unbolting the transformer and the (fake) voltage selector/fuse box housing. This shows the four bolts that need to be removed:

After removing them

the transformer and fuse box assemblies can be 'flipped' out of the enclosure:

The next step is disconnecting the transformer from the fuse box assembly and the rectifier. This shows the fuse box after removal of the fake voltage selector dial:

The grey wires come in from the power plug and the red/yellow wires go to the transformer primary windings:

I unsoldered the red and yellow wires and pulled them out of the fuse box assembly:

Be careful when you try this at home to not melt the plastic fuse box housing with the soldering iron. The next step was unsoldering the wiring from the rectifier (mounted beneath the floating chassis leaf spring next to the transformer):

This shows the unsoldered wires:

and the extracted transformer:

Now it was time to bolt the 3D printed adapter plate to the enclosure. It uses the same bolt holes that were used to hold the transformer in place:

This shows the adapter bolted in:

The next step is soldering the provided wire jumper to the fuse terminals previously connected to the red/yellow wires towards the transformer. This shows the wire jumper inserted from the bottom into the fuse housing:

And on the top side soldered to the fuse terminals. Left

and right:

It is a good idea to leave the fuses in place while doing the soldering. It will keep the fuse terminals in the proper orientation while they get hot.

Once the wires are soldered in, they need to be fed out of the fuse assembly through the 'exit channel' along with the grey power input cable:

Then the fake voltage selector can be stuck back onto the assembly,

which then can be bolted back to the enclosure:

This shows the fuse holder assembly back in place:

The next step is soldering the leads to the in- and outputs of the TPP30 supply's breakout board. First come the leads to the fuses housing assembly. They solder to the terminals "Input 100-240V AC~". Polarity does not matter since these wires carry AC:

Then the red and black wires originally connected to the rectifier output are soldered to the terminals labeled "Output 26.5V DC=". Make sure the red wire is soldered to the "+" labeled terminal and the black to the "-" labeled one:

Now the TPP30 assembly can be turned around and bolted to the adapter with the three provided screws:

Make sure that the in- and output wires are properly fitted into the cutouts on the adapter.

The final step is replacing the original fuses with the new 1.6A fuses provided with the TPP supply. It is easy to remove the original fuses by pushing them out with a suitable screwdriver:

 The new fuses provided with the kit can easily be pushed into the holders with a finger:

The final step is replacing the fuses cover. If it is not in place there will be potentially lethal voltages on the fuse terminals once the Beogram is plugged in. This shows the completed setup in place:

I plugged the Beogram in for a function test, and everything worked very nicely! Absolutely no noise from the supply! Beolovely!

Next I measured how much current the supply draws under the various operational modes of the Beogram.  As expected, the maximum current was drawn under 'play condition', i.e. platter motor running, solenoid engaged and the carriage moving towards the center of the platter. This shows the measurement:

The multimeter is connected between the fuse terminals on one side and set to its 10A range. In this setting it essentially acts as a small current sensing resistor in-between the fuse terminals, and the voltage drop across this resistor is converted into a proportional current reading.

The display shows 0.2 Amps RMS. This corresponds to about 24W power at 120V. In other words, the supply runs at about 80% of its maximum capacity of 30W. 80% is the usual engineering safe margin for reliable long term operation. All good in the power supply department! I should point out again that this new power supply only works in tandem with the Efficient 22.8V Power Supply and Main Capacitors for Beogram 4002 (Types 550x), which significantly reduces the power intake of the Beogram compared to the original setup. Therefore, please do not use the TPP30 supply if your Beogram runs from its original linear voltage regulator based 22.8V rail. It would not be safe and the fairly expensive TPP30 might suffer premature retirement...;-)

Beogram 4002 (5503): Full Functional and Cosmetic Restoration of a Unit Rescued from Ebay

***********************************************************

Note added in proof: During testing, this unit revealed a loudly humming mains transformer. I replaced it with a new universal power supply that runs very quietly. It also has the added benefit that this unit is now capable of running on any line voltage between 100V and 240V. This also means it is much better brown-out and over-voltage protected than the original setup. Read more about this upgrade here.

After extensive testing, I realized that the unit sometimes would not trigger auto-return of the carriage after playing a record. This is usually related to damaged carriage position sensor housing plastic parts, an out-of-spec photo-resistor, a broken bulb, or a misaligned plexiglass 'ruler'. Since it can be time-consuming to detect the actual root cause, I decided to develop a modern monolithic replacement of the carriage position sensor and implement it in this unit. Read here how this new part, the Beolover Carriage Position Sensor for Beogram 4002 (Types 550x), was installed.

***********************************************************

This post describes the work done during the restoration of a Beogram 4002 (Type 5503 with AC platter motor) that I was able to purchase recently. My initial assessment of this unit is posted here.

This shows the final result of my labors:

Looks nearly new and also plays like new! Beolovely! This Beogram is available for purchase at the DKaudiolover store.

Let's see what it took to get there:

This shows the unit after removal of the aluminum panels and platter:

While the AC platter motors are less troublesome than the later DC motors due to their brushless stacked can design, they also have issues with drying out Oilite shaft bearings. Dry bearings often cause them to 'knock' while driving the platter. Therefore, I also re-infuse their motor bearings with oil under vacuum when I restore this Type. This shows the AC platter motor together with the original big capacitor cans that serve as motor coupling and phase shift capacitors, as well as reservoirs for the main circuit:

I removed it all leaving behind a bunch of orange transport lock bushing fragments:

This shows the extracted motor:

It has four leads since it is a two-phase synchronous motor. I took it apart to prepare the motor enclosure halves for oil infusion:

It is impossible to remove the motor bearings without doing significant damage, so I simply submerge the entire motor housing in synthetic oil making sure that both bearings are completely in the oil:

After pulling a vacuum one can usually see air bubbles coming from the bearings in the center of the pieces. This indicates that air is drawn from their porous bearing material making room for oil to diffuse into the material. When the bubbling stops the process is complete and the bearings are ready for another tour of duty.

While this process was underway, I focused on the other restoration tasks. First came the restoration of the carriage. The arm lowering and carriage motion mechanisms usually suffer from hardened lubricants. Therefore it is a good idea taking it all apart for a good cleaning and re-lubricating. This shows the original encrusted setup:

A detail shot of the as usual cracked black plastic carriage pulley:

I removed all moving parts:

It is a good idea to rest the carriage on a soft pad to protect the delicate wiring at its bottom.

This shows the removed parts ready for the ultrasonic cleaner:

Aside from cleaning and re-lubricating it is also a good idea to de-magnetize the solenoid. If the plunger is magnetized it can lead to delayed arm raising, which can cause the needle dragging across the platter during auto return after playing a record. Not a fun moment when a newly rebuilt $800 cartridge is mounted on the arm! This shows the extracted solenoid:

The removed plunger was indeed magnetized. It was able to attract a small M3 iron set screw:

After treating it a bit with my tape head demagnetizer the setscrew was able to stay put even at very close distances:

In the meantime the parts came back from the ultrasonic nice and shiny:

I started re-assembling everything. It is a good idea to install a new damper gasket:

The old ones are often hardened and/or deformed, which often causes inconsistent arm lowering speeds. Another 'fun moment' when the nice new cartridge occasionally crashes on the platter without damping!

An 'Achilles heel' of the AC motor Beogram designs is the solenoid arm. It has a plastic extension that is riveted to the metal arm. Like most Beograms of this Type this unit had also already developed a crack in the plastic around the rivet:

This is a dangerous issue since it often leads to a burnt solenoid coil if the problem is not addressed in time. When the extension finally breaks off, the solenoid switch is not actuated anymore after the solenoid lowers the arm. This causes the solenoid current to flow full strength while the record is playing. After a few minutes the solenoid then gets hot enough to melt its spool and the polyurethane wire insulation. The process ends with a massive short circuit that blows the fuses.

I always install a 3D printed Beolover Solenoid Arm Extension for Beogram 4002 (Types 5503, 5504). This shows the new and the removed original cracked part in comparison:

I bolted the new part in place and used a good amount of epoxy to fix everything in place:

The epoxy is an important part of this installation. The plastic extension needs to be attached very well and securely.

The next step was updating the solenoid switch. I now always replace these switches as part of my standard restoration 'package' since I had an unnecessary warranty call last year when a 5503 that I sold earlier in the year came back with a broken solenoid switch. These switches are a bit under-dimensioned for the task since they need to break currents of 3 to 4 amps every time the arm lowers, while they are only rated for  2 amps. I guess B&O accepted a bit of 'planned obsolescence' here assuming the switches would not see enough switch cycles over the expected 10-15 year product life-time to cause significant trouble.

This shows the original solenoid switch board sill in place (together with the rebuilt solenoid arm):

I unsoldered the three leads and removed the board:

This shows the board with the new (white) switch installed:

There was another cracked original plastic part, the spindle nut holder. It usually breaks around the bolt hole due to over-tightening. This shows the partially cracked original part next to the new Beolover Spindle Nut Holder for Beogram 4000, 4002, and 4004:

This shows the new part installed:

It bolts in with the original screw.

This shows the entire setup back together:

Here a picture of the new Beolover aluminum carriage pulley:

Beoshiny! There was one more pivot point left in need of cleaning and re-lubrication: The damper to arm linkage! It is mounted in-between the tone- and sensor arm assemblies. You can see it stick out from the v-cut in the small metal linkage that is bolted to the back of the counterweight:

In order to get to it the sensor arm needs to be removed. This shows the sensor arm lying on its side with the linkage already removed:

Of course the small copper plate that helps the arm move laterally when up came loose with only a light tug with my tweezers. I epoxied it back into place after re-assembling the arm:

The next step was aligning the sensor arm base and the sensor arm itself orthogonally relative to the rods the carriage travels on. I did this using my recently developed Beolover Arm Alignment Tool for Beogram 4000, 4002 and 4004:

It streamlines the process considerably compared to eyeballing the alignment, which I did before I came up with the tool.

Next I replaced the original incandescent bulb of the tracking sensor with a Beolover Tracking Sensor LED Light Source (Beogram 4002 and 4004). This shows the original setup with the black bulb housing still in place:

I removed the bulb housing, which revealed the tracking aperture:

It permits more or less light onto a photo resistor, which drives the carriage motor to keep the light intensity constant. This shows the original and new parts next to each other: 

The small SMD LED of the Beolover part is in the same spot like the filament of the bulb enabling a smooth drop-in replacement process:

This completed my work on the carriage and I moved on to restoring the main PCB. It usually has two 'after-thought' capacitors installed on its solder side:

I replaced these two tantalum capacitors with modern units of like value

and then I unbolted the board and flipped it over:

I replaced all electrolytic capacitors and power transistors, as well as the RPM trimmers and the sensor arm transistor (which is usually out of spec and does not have enough gain anymore). This shows the rebuilt board together with the extracted old components:

While the board is up it is a perfect moment for replacing the two TIP31/32 transistors of the AC platter motor push-pull amplifier stage. They are bolted to the bottom of the enclosure for efficient heat dissipation:

I replaced them with new and stronger TIP 41/42 packages:

Next I removed the keypad assembly and took the output board out:

It carries the output grounding relay and its time delay circuit:

I replaced the relay with a new Beolover Siemens Relay Replacement for Beogram 4000, 4002, and 4004 and a new capacitor for the delay circuit:

I also installed a (red) switch that allows connecting signal and system grounds if there is a hum in the audio signal. Connecting the grounds usually fixes such issues.

At this point pretty much everything had been removed from the enclosure except the floating chassis itself. This made it the perfect moment for fixing the decayed transport lock bushings. This unit had lost them completely. This shows one of the transport locks partially dismantled:

While the floating chassis is out it is a great moment for replacing the solenoid transistor, which is bolted to the bottom of the enclosure:

This shows the floating chassis removed for cleaning out the enclosure. It had a lot of transport lock fragment strewn around:

I gave it a thorough vacuuming and then implemented new Beolover Transport Lock Bushing Set for Beogram 4000, 4002, and 4004:

They install very easily since they are designed in two parts. Simply insert one half from the bottom and the other from the top:

Note that some Beograms have slightly bigger transport lock orifices than others and it can be a good idea to glue the lock bushing halves in with a small amount of super glue gel on the vertical parts. This shows one of the bushings with the transport lock bolt in place:

I also installed a new solenoid transistor:

I usually replace the original TIP41 with its higher voltage cousin TIP41C hoping for some more longevity. 

The next step was the installation of the Beolover Efficient 22.8V Power Supply and Main Capacitors for Beogram 4002 (Types 550x) board:

It elegantly replaces the original big capacitor mess and also gives the Beogram a modern precisely controlled 22.8V power rail that is much less wasteful than the hot original linear voltage regulator setup. This means the Beogram will consume around 30% less energy than before when this board is installed. The board fits into the space previously occupied by the big capacitor cans and uses the same bolt holes:

I usually solder all the non-motor leads into place first:

followed by the motor wiring:

This shows everything in place:

I usually bolt the original mounting strip that held the capacitors in place back in. It nicely keeps the wiring down so it does not scrape on the platter later.

There were still a few light bulbs that needed replacement. First I replaced the one in the sensor arm. This shows the small compartment pulled out with the original bulb still in place:

Next to it is the Beolover Sensor Arm LED Light Source (Beogram 4002, 4004 and 4000) together with its alignment aid. When I removed the bulb it became immediately clear why the sensor did not work anymore. The bulb had lost one of its leads to corrosion:

I implemented the LED board:

and then tested it:

It is based on a warm-white LED which emits enough red photons to illuminate the B&O logo in a nice warm incandescent sheen.

Two more bulbs that need replacement are in the RPM adjustment panel above the keypad. This shows the panel flipped on its front, which reveals the two bulb covers:

I removed the covers:

As usual the 33 RPM cover had heat damage, as is evident from the brownish discoloration on its inside. I replaced the bulbs with Beolover RPM Panel LED Backlights for Beogram 4002 (Types 550x). They solder directly to the points where the bulb leads went:

Here a detail shot of one of them:

The little boards essentially act as extensions of the original circuit board.

I tested the LEDs and as expected the white background behind the 33 RPM trimmer scale was wavy: 

The background foils also suffer heat damage from the hot incandescent bulbs. I removed the plastic assembly from the front plate, which revealed the foils. Indeed the 33 RPM side was wiggly:

I replaced the foils with 3M white insulating tape cut to size:

A second test after reassembly showed a nicely homogeneous background:

Beolove is in the details!...;-).

The remaining incandescent bulb in the 5503 is on the carriage position PCB under the plexiglas 'ruler' that is bolted to the carriage. This shows it after removing the ruler assembly:

First I replaced the solenoid resistor and the decoupling capacitor of the position sensor photoresistor. This shows the two new components in place together with a rear view of the black position sensor components:

The black housing on the right conceals the light bulb in the sensor. This photo shows the bulb after removal of its housing:

I usually replace the bulb with an amber LED and a current limiting resistor:

The next step was adjusting the bias of the sensor transistor to yield the service manual prescribed 3.0V at its collector:

After the adjustment I moved the trimmer over to the component side of the board:

Then I checked the sensor response with my oscilloscope. This shows the collector voltage with an empty platter spinning underneath the sensor:

Each dip in the curve corresponds to a black platter rib passing under the sensor. The measured amplitude of 6.7V is an excellent result. The record detection circuit was working very well in this Beogram! 

While I had the oscilloscope fired up I also measured the motor signal. This shows how I connected my probe:

The measured signal was a perfect sine-wave. I adjusted the motor voltage trimmer to get 4.6V RMS and the 33 RPM trimmer to yield 43 Hz frequency:

I also adjusted the 45 RPM trimmer to get 57 Hz:

These frequencies are not precise but they closely correspond to 33.33 and 45 RPM.

This unit probably spent some time in an attic or basement. All the external wiring was gnawed on by some hungry rodents. This shows the output cable:

It also had a very grimy corroded DIN5 plug:

I had a good condition output cable without rodent bites in my stock. I installed it in lieu of the damged one and also put a new all-metal plug on it with gold plated contact terminals:

Beogolden! Next came the replacement of the dangerously damaged power cable which had some stretches with exposed copper:

I soldered in a spare cable that I had rescued from a parts unit that I was given at some point in the past:

I had to remove the fuses to be able to solder the new cable to the contacts in the fuse box. When I popped them out of their holders one came apart:

This is a common failure of the original fuses and it is generally a good idea to replace them with new ones whenever a Beogram is restored:

After fixing the wiring it was time to the mechanical adjustments. In order to do this the aluminum platers need to be installed, which made it the perfect moment for installing a refurbished keypad to replace the badly damaged original one. This shows the original keypad (left) in comparison with a newly restored keypad:

Please, see here if you are interested in fixing your keypad or let us rebuild it for you.

I also re-installed the nice condition original wood frame. I replaced the cracked original plinth guidance washers with a new Beolover Plinth Guide Washer Set for Beogram 4000, 4002, and 4004:

This shows one of them installed:

And here a shot of restored keypad together with installed plinth:

Beolovely! Now it was finally time to do all the adjustments. In these adjustments the arm travel across the platter needs to become parallel, the platter needs to be in the proper distance from the arms, and the platter needs to be flush with the surrounding aluminum panels. This can sometimes be a pretty iterative slog. Due to the little space the floating chassis has for movement there is not much room for 'improvisation' and everything needs to be perfect.

At this point it was a convenient moment for doing my usual 24 hrs RPM stability test. With the AC motors I usually do not expect much trouble in the RPM stability department, but it is always nice to see a very stable RPM curve!

I installed my BeoloverRPM device for the measurements:

The BeoloverRPM has two operational modes. In 'slow' mode (shown above) it measures the RPM in 10 sec intervals and sends the data to a serial port of a computer where it can be recorded with any generic terminal software. This allows plotting the RPM over long periods of time using Excel or similar software. This shows the result of my 24 hrs measurement:

As expected the curve is very flat. The AC platter motors are synchronous motors driven by a Wien oscillator, a great 'recipe' for a rock solid RPM behavior!

Next I collected some high resolution data in the 'fast' mode of the BeoloverRPM:

In this mode the device sends an RPM measurement every time a platter rib passes through underneath the sensor. This gives 24 RPM measurements per platter turn (there are 24 ribs). This allows visualizing short term RPM changes caused by the feedback-based motor control circuitry. This graph contains the data covering ~35 platter turns (~70 sec):

The prominent zig-zag pattern is a measurement artifact that comes from the fact that the spacing of the platter ribs around the platter is slightly irregular due to manufacturing imperfections. This generates a repeating pattern for every platter turn (sort of a 'platter fingerprint'), which is superimposed to the real RPM changes that are introduced by the platter motor. This real RPM change is essentially the weak sine-wave like pattern that modulates the zig-zag pattern. An evaluation of the wavy component yields a wow and flutter estimate of about 0.1%. Since the platter motor is a synchronous motor, such variations can only be introduced by frequency fluctuations of the Wien oscillator, or due to the elastic coupling between motor pulley and platter, i.e. oscillations that develop in the rubber belt due to motor torque changes during its rotation. We may never know exactly. 

It is interesting to note that 0.1% is double the 0.05% stated in the specs list in the service manual. This difference is most likely systematic due to the entirely different way wow and flutter was measured in the 1970s when these turntables were produced. I see it for all Beograms that I restore. Back then the measurement was carried out with a test tone on a test record. In these measurements deviations from the tone center were measured with an analog spectrum analyzer and then converted into a wow and flutter number.

It should be pointed out that this discussion is pretty academic in the first place since humans start typically recognizing frequency fluctuations above the 0.7% threshold, i.e. the RPM fluctuations of this Beogram are well below this threshold, whether the number is 0.05% or 0.1%. This motor is definitely ready for duty again!

After these measurements were completed I did the arm adjustments. But first I replaced the flimsy locking washer that holds the counterweight screw in place

with a square nut and a washer:

The nut allows locking the calibration in place so it survives shipping.

Then I put a cartridge on the arm and calibrated the counterweight to get 1.2g at the 1.2g setting of the little tracking force dial:

The next step was adjusting the arm lowering limit that the needle would stop dropping about 1 mm above the lower platter rib sections:

This is an important adjustment in case the photocell in the sensor arm fails or another circuit malfunction occurs. The Beogram circuit only detects sensor lamp failures, which cause the arm lowering mechanism to be disabled.

Then I adjusted the tracking feedback

after which it was finally time to give this lovely Beogram a first test spin. I selected one of my favorite Herbie Hancock records, "Secrets" which was released in 1976 (CBS 34280).  What a beautiful 1970s cover!...;-). A perfect contemporary to this Beogram, which may have been purchased in 1975 or 1976:

The Beogram played perfectly and so I started working on installing a new plexiglass dust cover. The first step was removing the badly scratched original cover from the metal hinge assembly. The mounting screws are easily accessed after lifting the aluminum trim strip on the sides with a razor blade:

I always epoxy small (grey) plastic bits into the strange channels they designed into the otherwise very nice replacement hoods from DKsoundparts in Denmark:

These channels make it impossible to bend the aluminum trip sharply down on the sides. Since there is a void under the strip in the channel area the bend invariable gets a bit rounded. The plastic bits make it easier to get a reasonably sharp bend across the entire width of the trim strip.

The next step was centering the new trim strip with my alignment tools

I usually use a piece of blue tape so I can make a pencil mark for the strip alignment:

Then I bolted the new hood to the hinge. I usually put a bit of white wood glue on the threads:

This helps keeping the screws in place. One should tighten them very much since this can easily lead to premature cracking of the plexiglass around the bolt holes.

The final step was glueing the new trim in place and bend it down at the sides: 

Beolovely! I offer hood replacements as a service in case the above process seems daunting.

I installed the hood on the Beogram.

Here are a few pictures of this nearly new looking Beogram. Enjoy!