Replacing the sensor arm LED is typically the final step in converting a Beogram 4002 to use all LEDs. After restoring the DC motor, the next task was to install a recently designed flex-PCB-based LED assembly to replace the incandescent bulb. The image below shows the sensor arm with the detector assembly removed:

Upon zooming in on the full-resolution image, a small black spot is visible on the glass bulb. This spot indicates significant tungsten evaporation from the filament, suggesting the bulb is nearing the end of its life and could soon fail, rendering the Beogram inoperable. Installing an LED offers improved long-term stability, which is the primary motivation for LED upgrades, apart from the difficulty in finding suitable small light bulbs.

The limited space within the sensor arm and the need to match the power consumption of the original light bulb to avoid disrupting the record detection circuit presented challenges for this LED upgrade.

The custom LED replacement incorporates a resistor to compensate for the significantly lower power draw of the high-brightness, low-color-temperature LED, which operates at a very low emission level. The assembly is shown below:

It’s constructed on a flexible PCB that folds neatly into the bulb compartment:

A 3D-printed red wedge glued to the assembly ensures the LED is positioned similarly to the filament in the original bulb.

Here’s the assembly in operation:

The use of a low-color-temperature LED, which has a high red component, results in an accurate color representation of the B&O logo.

When replacing the sensor arm’s light source, it’s crucial to measure the sensor’s response over a spinning empty platter. This verification ensures the signal strength is sufficient to activate the protection circuit, which prevents the arm from lowering onto the platter when no record is present. Below is the trace measured on the collector of TR3 after installing the LED:

At first glance, it appears normal, with the dips reaching 0V. However, a closer look reveals that the signal amplitude is only around 3.7V, significantly lower than the approximately 6V specified in the circuit diagram. This reduced amplitude is insufficient to reliably activate the protection circuit.

The relevant section of the circuit diagram is shown below:

Here’s how it functions: The voltage divider formed by R26 and the transistor’s BE diode biases the base of TR3. The signal from the photocell (OPH1) in the sensor arm is coupled through C12. In the absence of a signal, such as when over a record, the base voltage sets the collector voltage of TR3 to 4V, putting the transistor in a slightly ON state and creating a voltage divider with R27 (and R30). When OPH1 produces its AC signal with an amplitude of approximately 20mV, the voltage at the collector of TR3 oscillates between 6V and 0V, effectively turning TR3 into an amplifier. This 6Vpp signal then passes through C16, which removes the DC component, and feeds into the base of TR4. Since R32 pulls up the base of TR4 to 21V, and D15 protects it from negative voltage (relative to its emitter), the resulting signal at the base of TR4 oscillates between 21.6V and 20.4V. When the voltage is at 20.4V, TR4 turns on, allowing C18 to charge, which in turn biases the voltage toward 21V. This action pulls up the base of TR6 through R34 (acting as a current limiter), causing the collector of TR6 to be pulled to ground. This signal then disables the arm lowering circuit, preventing the arm from lowering.

The base of TR6 can also be pulled up through the D18/R37 connection, which is linked to the collector of TR5 through the Zener diode D17. This circuitry generates the ’no record present’ signal at the collector of TR6 if the light bulb fails. An open circuit in the light bulb prevents current flow through R36, causing the voltage at the collector of TR5 to rise. This, in turn, pulls up the base of TR6, causing the collector voltage of TR6 to drop and preventing the arm from lowering under any circumstances. Therefore, the LED replacement must be designed to maintain a current flow through TR5 similar to that produced by the original light bulb (approximately 50-60mA).

Returning to the signal measured at the collector of TR3, which was less than 6Vpp, it was observed that some Beograms experience issues with their detector circuits. These units may fail to detect the absence of a record and lower the arm onto the platter despite appearing to function correctly. In such cases, the stylus’s lifespan depends on the arm lowering limit being properly adjusted to prevent it from striking the platter ribs.

This problem was previously addressed by soldering a 10M resistor between the base of TR4 and ground. This modification slightly biased TR4 on, enabling even a weaker signal from TR3 to stop the arm lowering circuit. While the root cause was unknown at the time, the solution was discovered accidentally. It was observed that the circuit tended to function correctly when an oscilloscope was connected to the base of TR4. The oscilloscope’s internal resistance was resolving the issue. Through trial and error, it was determined that a 10M resistor to ground provided sufficient pull on the base of TR4 to maintain functionality after disconnecting the oscilloscope.

However, there’s a more effective solution: doubling the value of R26 to 2M. This change reduces the current flowing through the BE diode of TR3, making the transistor less ON. As a result, the voltage at the collector rises to the intended 4V. Implementing this modification in the Beogram resolved the issue:

With the oscillation amplitude now close to 6V, TR4 can be triggered reliably.

This raises an intriguing question: Why did B&O allow Beograms with 1M resistors and a low TR3 collector voltage to leave the factory?

One design flaw concerning TR3 is that its collector voltage is influenced by the DC current gain of the specific transistor used. Current gain is the most variable parameter of transistors, even within the same production batch, and can fluctuate by a factor of two or three. A more robust design would have used a stiffer voltage divider to ground to bias the base and incorporated feedback to regulate the amplification gain. The current design means that each Beogram effectively has a different TR3 operating point due to variations in the soldered-in BC182 transistor. When the sensor arm was equipped with a new light bulb, this variation might not have posed a problem because the bulb’s high initial light intensity produced a sufficiently strong signal, even with a non-ideal TR3 operating point. However, as the bulb aged, tungsten deposition on the glass gradually reduced its intensity, eventually weakening the signal to the point where the mechanism would malfunction. It’s possible that this problem went undetected during the manufacturing process.

An interesting observation for Beogram enthusiasts.