Beomaster 1900 Type 2904: Solving the Voltage Difference Between the Channels

I ended the previous post with a problem regarding the output of the Beomaster 1900 amplifier.
Specifically, the Left channel output signal was about 20 percent stronger than the Right channel.

At first I thought the problem might be with the output amplifier transistors...either the Left channel or the Right channel.  However, I also had to keep in mind that the amplifier sounded good and while there were measurable differences, the distortion and frequency response measurements were very similar to other Beomaster 1900 units I have seen.

Still, I felt that a 20 percent difference in output between the channels was a little much.
So I did some audio signal tracing and measurements of the test signal through the various stages of the Beomaster 1900.

At the point of the input signal selection, right before going into the volume control, both the Left and Right channels are identical.

After the volume control, the difference begins...but it isn't always 20 percent.

Studying the Beomaster 1900 volume control again, I realized that the light source controlled, LDR volume control assembly had to be the most likely culprit.

The Beomaster 1900 (and 2400) volume control is an interesting design and is worth looking at here.

Owners love the sleek design of the Beomaster 1900 and in the mid-seventies, the touch control buttons were quite a modern marvel.  Source selection and Standby mode selection were simple on/off state changes while the up/down volume control was a more complicated matter.

The volume control has two parts to the control.

There is the control from a human finger holding the volume button up (or down).  Then there is the actual attenuation control of the audio signal.

First, the detection of a human user wanting to change the volume level.

This circuit and description show how the Beomaster 1900 takes commands from the volume control up/down buttons and turns them into different voltage levels.

The selected voltage level is used by the actual volume control of the Left and Right audio signals.
The way that works is the selected volume level voltage controls the amount of light a small lamp emits on a set of four LDR (Light Dependent Resistor) devices.

The LDR volume control assembly containing the light source (lamp) and the four LDR devices looks like this.

Disassembled, the Volume Control LDR Assembly looks like this.

As you can see, the volume control assembly comes apart and the discrete components can be replaced.
Martin Olsen (Beoparts.com) supplies a rebuild kit for these that consists of the lamp and four LDR devices.

The Beomaster 1900 Service Manual describes how the light source and LDR volume control works.

Here is the Beomaster 1900 schematic section that shows the Left and Right audio signals passing through the volume control and to the output amplifier.
The 8002293 device shown in red is the LDR volume control assembly.

The schematic above shows four key measurement points I used in looking at the audio signal.
TP200 and TP300, before the volume control device.
TP201 and TP301, at the entry to the output amplifier.

An important thing that stood out from the Service Manual description of the volume control device was that full volume occurs with a minimum amount of voltage on the lamp (typically 1 VDC) while the lowest volume level is typically with about 5 VDC applied to the lamp.  

That is opposite of what I would have thought before diving in to the inner workings of the volume control.
What that means though, is that the high end power (maximum output) of the Beomaster 1900 that I want to measure the performance of is at the lowest light source level.

Combining that information with what I observed in the Beomaster 1900 performance...
The Left and Right channel outputs at the speaker load were much closer to each other at output levels below 5 Watts.  That is when the light source is emitting more light.  From 5 Watts and up (to the maximum 20 Watts), the Left channel output got further away from the Right channel. That is when the light source is emitting very little light.

That looked to me like a volume control calibration problem.

I ended up playing around with three different volume control assemblies (two shown here).

Eventually I settled on the one I could adjust the best.

I have to mention that rebuilding this volume control device is more involved than simply de-soldering the old components and soldering in the new ones from the kit.

The calibration technique I used was to connect the lamp terminals of the volume control device to a bench DC power supply.

From there, I applied voltages from 1 VDC to 5 VDC and compared the resistance of each LDR device.  I wanted the L&R sets (R200/R300) to be close to each other and (R201/R301) to be as close as possible to each other.

To adjust the LDR devices to accomplish that requires moving the LDR devices a bit towards (or away) from the light source.  Even tilting them a bit changes their value.

The adjustment procedure means applying a voltage to the lamp and measuring the LDR resistance with the metal cover on the volume control assembly.  To adjust an LDR means having to remove the metal cover, readjust the LDR, then replace the cover to measure the resistance again.

The procedure is quite tedious and I ended up decided on a source voltage of about 1.3 VDC as the target level I wanted to adjust the LDR devices to. 

That corresponds to a pretty high volume level and wasn't too difficult to measure with a DMM.

When I was finally happy with my result, I reinstalled the volume control device and retested the Beomaster 1900 amplifier.

The THD measurement immediately showed an improvement.
 

The Left and Right channel outputs at maximum rated audio power into 8 ohms are now within almost 6 percent instead of over 20 percent.

The THD measurement for both channels at that maximum output level is in the 0.03% range.
That is a very good value for this amplifier model.

The Frequency Response measurement for the two channels also shows the signal amplitudes being closer than before the rework on the volume control LDR device.
The Frequency Response measurement is within the limits I have previously seen on the Beomaster 1900 units although the Right channel does appear to have a better response trace below 1KHz than the Left channel.

I will move this Beomaster 1900 into a listening room for better real world listening tests (than out in the workshop).  Unless I don't like something I hear during those tests, I will consider the bench testing complete.

Cosmetically there is still the issue of peeling veneer from the plastic side panels.
I need to see if I can save those while the Beomaster 1900 is in the listening room.

Beomaster 1900 Type 2904: Reassembly and Testing

Continuing from the last post, I performed two of the Beomaster 1900 Service Manual adjustments.
The No-load Current Test (12mVdc across the output amplifier emitter resistors).

...and the P1 and P5 Tuning Voltage adjustments.

The Beomaster was now ready for reassembly and some audio testing.

The first step was to re-install the plastic masks for the Bass, Treble and Balance controls.

After that, the button panel is installed.

The button panel installation is a little tricky. There are 10 touch control buttons.  Each button has a metal spring connecting the panel button to a corresponding metal post soldered onto the main board.
When installing the button panel you have to make sure all of the button springs mate with their metal circuit board posts.  

My technique is to get most of the spring contacts to mate...8 or 9 out of 10 is good.

For the spring contacts that do not successfully mate, they can be corrected using a prong tool from underneath frame.

Here is an example of one of the springs not mating.

The picture on the right shows the corrected spring contact mated with the post.

I screwed down the button panel and powered the Beomaster 1900 on to check if, indeed, all ten buttons worked.

They do.

Now for the installation of the program panel and test play some music through the Tape input.

I used both speaker outputs of the Beomaster 1900 during this initial audio testing.
A pair of Beovox MC120.2 speakers on Speaker 1 and a pair of Beovox S-55 speakers on Speaker 2.

They sounded very good so I finished up the re-installation, installing the bottom cabinet plate.

I hooked up an FM antenna and did some listening tests on the tuner.  P1 through P5 all tuned successfully and also sound very good.

While a successful listening test carries a lot of weight in judging the success of a restoration, I like to do a basic harmonic distortion test at full power and a 1000 Hz sine wave input (on the Tape input).
I also like performing a 20 Hz to 20KHz frequency response test at full power (again, using the Tape input as the source).

For these audio measurements I use a QuantAsylum QA401 Audio Analyzer.  There is a newer QA403 available now but the QA401 is still very good and I continue to use it.

I always mention that my audio tests cannot be directly correlated with what is printed in the service manual (or B&O published specs) as the equipment in the mid-seventies was quite a bit different.
Also, many of the specs are a little vague regarding how, specifically, they were performed.

For that reason, I try to do my tests the same way each time using my audio analyzer, then compare my results to previous results I measured on other restoration projects.

My speaker loads for the audio tests are fixed resistors that provide an 8.08 ohm speaker load for each channel (Left and Right).

8.08 ohms because each 8 ohm load resistor assembly is made up of one 4 ohm load in series with a 0.08 ohm, sense load and two 2 ohm loads.

The reason for that is because the QA401 audio analyzer has a limit of 28 Vp-p (or 19.8 Vrms).
19.8 Vrms = 26 dBV which is the form of the input value QuantAsylum uses in stating the amount of input.

At full power of 20W output across 8 ohms, the Beomaster 1900 maximum output my QA401 should see (across the dummy load resistor assembly) is 22 dBV.  So I can place my QA401 differential measurement probes across the full 8.1 ohm dummy load safely.

If I really want to be at a safe measurement difference I could put my QA401 probes across one of the other sections of the dummy load (i.e. across the 4 ohm resistor, the 2 ohm resistor or the 0.08 ohm resistor).  However, the best results should come from using the highest output level I can safely use.

Here are my distortion measurements at maximum rated output power (or as close as I could get) across the 8.1 ohm dummy load. (The Tape input source was a 1KHz sine wave at -12dBV).  

Left THD: 0.038%

Right THD: 0.062%

The volume control for this Beomaster 1900 version is only adjustable in sixteen steps.  Later serial number units (serial number 1826011 and on) increased the volume selection steps to 128.  The output power is the same power but the volume adjustment is much finer.

Because of the coarser, sixteen step volume control, I played around with the volume control value and my source sine wave generator amplitude to achieve an output that was close to the maximum rated power for this amplifier. 

I am pleased with the harmonic distortion measurement results.  Although, as you can see, the Left channel is producing 20% more power than the Right channel.

I measured the audio test signal through the Beomaster 1900 pre-amplifier to its voltage amplifier stage in the output amplifier.

At the input to the output amplifier (Test Points TP201 and TP302 for the Left and Right channels respectively), the Right channel is about 10% lower than the Left channel.
When the signal is at the collector of transistors TR212 (Left channel) and TR312 (Right channel), the difference in the channels is 12%.

I will do some more investigating on the difference in output level between the Left and Right channels.
It appears that most of the difference occurs in the output amplifier.

Meanwhile, I continued on to a frequency response measurement at maximum power. 
I ended up with a little more volume for the measurement and the Left channel output was 5 Watts over its rated value.
The measurements are within the expected value for this amplifier but again, the Left channel output is higher than the Right channel (by 23%).

A Pair of Beomaster 1900 Receivers for Restoration: Finishing up Unit #1

The first of the two Beomaster 1900 receivers is now complete.  

I wrapped up the testing tonight by checking the Total Harmonic Distortion (THD) at 1KHz with 20 Watts of power out of the Beomaster 1900 into my dummy 8Ω speaker load as well as the Frequency Response from 20Hz to 20KHz at an output level of 20 Watts.

Here is a photo again of the dummy speaker load test setup.

The Total Harmonic Distortion (THD) with a 1Khz sine wave into the Beomaster 1900 Tape source input and maximum rated power (20 Watts) across the dummy speaker load measured 0.068% which is well below the service manual specified maximum THD level of 0.2%.

Next I performed a frequency response measurement from 20Hz to 20KHz frequencies with and input level of 260mV into the Tape input and the power level at 20 Watts again.

The result was a ±1 dB swing from the 20Hz to 20KHz frequency range. That is pretty much the same as I measured for the Beomaster 2400 receiver I restored earlier.

 

Those are very satisfactory measurements for this Bang & Olufsen series amplifier.

This Beomaster will now go into a listening room while I get the second Beomaster 1900 unit restored.

Here are a couple of final photos from the workbench.
This Beomaster 1900 receiver shouldn't be seeing the workbench again for fifty years.