Beomaster 6000 Display Repair - Intensity Adjustment

After refurbing the TP8 display of my Beomaster 6000 I had another idea how to adjust the brightness of the new LEDs to the dim glow of the old ones. I did not like the 'halo' around the assembled displays coming from the small crack between the aperture plate and the PCB of the displays. Furthermore, the new (brighter) LEDs actually created quite a bit of light intensity behind the displays through the PCB. This caused a slight glow behind the displays even with the black cover 'bezel' mounted on top of the displays. I remembered that I had some "Safmat" laser printable foil. This foil is very thin and flexible, much more so than a standard transparency. It is also adhesive.

The idea was to put the grey filter right on top of the LEDs and not on top of the aperture plate. I printed one for the foils with the same "75% black" that I used for the transparency based filter (see last blog entry). Then I cut a strip a bit wider than the PCB, and of about twice the length and adhered it to the display:

Then I cut Safmat foil and wrapped it around the back:

Now I wrapped the remaining part also around the back:

Then I poked holes into the Safmat film where the mask and red filter positioning pins needed to be inserted into the PCB. These holes were small, with the idea being that the remaining material in the holes would help hold the mask and filter in place with a press fit. This worked fairly well, but I also secured the red filter with Scotch adhesive tape on three sides. This seems to do the job quite well. I did not want to use epoxy or similar since I wanted to be able to take things apart in the future if need be...here is the result:

Note that with the Safmat approach the segments are visible as they were before, even if the display is off...The transparency filter on top of the mask obscured the segments, i.e. the displays were not visible without illumination. Maybe that is more perfect, but certainly not original. Here is the display fired up:

Quite nice! the intensity is very close to the original LEDs. No crosstalk around the segments either-another advantage of the Safmat filter approach, since it covers the LEDs on all sides, i.e. they cannot shine light into adjacent segments.

Here is a pic of the installed TP8 display (sorry for the jitter-the camera was a bit slow due to the low light level). Note the halo around the frequency display-it still has the "transparency based filter" on top of the mask instead of the Safmat filter. The TP8 display in contrast is completely dark around except for the segments.

A note with regard to installing the cover bezel: The displays need to fit perfectly into the black clamps of the bezel. It took me a few approaches to get the angle and orientation of the displays right. After this experience it dawned on me that the better way would have been to just tack the displays on on the left and right most pins. Then adjust the angle and position as good as possible according to the marks that were hopefully made before taking them out. Now put in the bezel, and see if the displays fit in there. If not, adjust them by heating the two solder joints and moving the display into a better position while the solder is soft. Once the bezel fits smoothly, solder the remaining pins into place.

Beomaster 6000 Display Repair - Practice Does It!

I decided that I also needed to put new LEDs into the input/program selector display of my Beomaster 6000 to make both displays appear exactly the same. There is a very slight color difference between the old and the new SMD LEDs...the new ones are very slightly more orange...I am thinking one could probably fix this by using a bluish filter instead of the gray one I put into the frequency display...well, something for the next days. Now on to replacing the LEDs:

After removing the display and opening it up, I scraped the old LEDs out with a scalpel. This works very well. If it is moved parallel to the pads, it is virtually impossible to damage them. Then I studied a few more SMD soldering videos by Norman Mier on You Tube. I used his "dry tack" method to first fix the LEDs in place. To do this one first places a generous amount of rosin on the pads. Then grab the LED with precision tweezers and put it in an aligned location between the pads. Then take a very small amount of solder with the iron tip and briefly touch one side of the LED. After doing this with all of the LEDs, I then did Norman's "put solder wire halfway across the SMD tab to be soldered, and move the tip quickly in from the outside towards the pad". I held the tip for about 0.5 sec in place touching the LED to ensure good solder flow and bonding. The temperature of the iron was set to 340C. Here are a couple of pics of the result-much nicer than the mess I generated with the frequency display:

Even if some of the LED pads appear to be not coated by solder on the pic, they are. It seems that the pads usually only get a fairly thin coating during this soldering process. I tested the LEDs for mechanical stability with a tooth pick, and I was not able to damage the bonds even with quite some force.

Here is a picture of the display running. I plugged the pins into the female plug of an old parallel computer cable. The female plug at the other end of the cable was then connected to a breadboard's pos and neg rails with individual breadboard jumpers.

Beomaster 6000 Display Repair - Pain and Suffering!

****note: This post is 'historic'...Please, visit this page for my latest display repair procedure: http://beolover.blogspot.com/2012/09/beomaster-8000-display-repair-hopefully.html****



I guess I was too happy too early with my Beomaster 6000 display repair...While the LEDs survived my 12 hrs stress test under full load, they did not survive putting the apertures and lenses back on...two solder joints failed. I realized that the low temperature soldering paste did not work very well...after watching a few more SMD soldering tutorials by Norman Meir (yes, I am a fan!!) on youtube it became clear to me that I had to re-solder all of them. I doused everything in rosin flux and re-soldered with a fine tip at 340C using standard 60/40 solder (sparingly). It turns out that one needs to put the tip on the pad for about 5 sec to heat the part high enough that the solder actually flows on the contact pads of the LED. I tried not to touch the LEDs directly to avoid damage. This seemed to work fairly nicely. After wiping the board with alcohol it looked like that:

I tested each LED with a toothpick for mechanical stability. They seem to hold quite a bit of force without damage now...I hooked them up again to my connection manifold and turned them on:

We will see tomorrow, if they survived this torture. I must have heated each of them up about 10 times by now...

Beomaster 6000 Display Repair

****note: This post is 'historic'...Please, visit this page for my latest display repair procedure: http://beolover.blogspot.com/2012/09/beomaster-8000-display-repair-hopefully.html****


I am glad to report that I was able to replace all LEDs in the frequency display of my Beomaster 6000, which also had a dead segment. I decided to give my Beomaster 8000 some rest, and try my skills first at the BM 6000, which is, I have to say, not as dear to my heart than the 8000. This pic shows the display with missing digit 2e:

Here is the display in disassembled state (just cut the red and white circles on the backside of the display with a razor blade. Be careful to not destroy the traces on the PCB when you do this!

This pic shows the red diffusor removed. The next picture shows the board without the aperture:

This pic shows the old LEDs as close as I was able to go with my camera:


I ordered some 14 mcd SMD LEDs from Newark.com (Lumex SML-LX0603IW-TR). These LEDs have a turn on voltage of 2V and run 20mA. With a razor blade I scraped the dead LED from its pad on the 2e location. Then I soldered one of the new LEDs in that location. This was surprisingly easy: I put some solder paste on one of the pads with a tooth pick. Then I held the LED with precision tweezers in the right location and touched the pad briefly with the soldering iron tip (which was at only 200C - I used the low temperature paste). Then I put some paste on the other side and touched the pad again. This was it. This picture shows the display with this LED inserted under standard operation condition inside the Beomaster 6000:

Works, but is way too bright...not satisfactory! I ordered some more LEDs with 9 mcd and 2.8 mcd output. I replaced some of the other LEDs with one of each of them - unfortunately pretty much the same result. It turns out that the 2.8 mcd LED was the brightest due to its lower turn on voltage of 1.8 V...It seems the old LEDs have a higher turn on voltage than all of the new ones. The next step was to try to change the resistors. Indeed the 14 mcd LED had a similar brightness after adding a 220R resistor in series with the already present 68R resistor for the e digits. I chose to not remove the old resistor since its second lead is soldered right underneath the volume indicator plastic foil, which is quite flimsy. I did not want another disaster to happen. It is probably difficult to replace the volume indicator...Here is the pic with the 220R resistor in place:

In case you wondered: The paper clip at R24 is in there to play with the other segments in digit 3 and 4 with the lesser mcd LEDs inserted at the c locations...these are off in this picture since no resistor is connected.

I decided to go ahead and replace all of the LEDs with 14 mcd ones. Here is the display board with the new LEDs:

I made a custom connector for the pins of the display board with some female header pins to run all the LEDs for 12 hrs at full output to test if they made it o.k. through the soldering process (three died right when I soldered them, but at that point I still used 280C-after lowering to 200C none died anymore). Here is a pic of the display under full output at 2V...A beautiful sight!:

Let's see if they will still be running tomorrow...

Beomaster 8000 Speaker Switch Repair

I guess I was happy too early. Two days after putting the beomaster 8000 back together, out of a sudden the left channel cut out...Headphones still worked properly, so I immediately suspected the speaker switch. I plugged the speakers into the #2 outlets...still nothing on the left channel. Strange! Both switches broken? I never tried the secondary outlets, though...maybe the #2 switch was already broken...

I opened the Beomaster up again and checked...indeed both speaker switches dead on the left channel. I tried to find exact replacements, but of course drew blanks wherever I looked. I finally settled on model S201031MS02Q made by C&K. I ordered them from Mouser.com (Mouser#: 611-S201031MS02Q). These switches have a very similar sliding knob like the knob attachments of the old switches. They also travel 4 mm, identical to the old ones. This is important that they fit into the slots of the heat sink cover. But of course these new switches are lacking the PCB base of the original switches, i.e. do not fit into the slots in the Beomaster 8000 chassis. The pictures show one of the original switches (with the knob attachment removed)

This picture shows the PCB base plate removed from one of the old switches:

I modified the base plates with a Dremel tool until the new switches would fit on there exactly centered

The next step was to glue the plates onto the new switches. I used superglue gel and secured the plates with two grip pliers:

The modified switches were exactly the same height as the old ones with their knob attachments in place, i.e. the heat sink cover fit perfectly on top, and held them snugly in place like the old ones. I put everything back together only to realize that another display segment died...time to get back to SMD LED soldering practice...;-)

SMD Soldering Practice

Today I hot air soldered my first SMD components. I used an Aoyue 2702A+ rework station. This is a cheap unit that "does it all". So far I am pretty happy with it, except that it had a very strong chemical odor when I unpacked...I kept it in the garage for a few days for outgassing. Now it got better, but it is still noticeable. Anyway, on to the SMDs!

I used 14 mcd LEDs from Lumex (SML-LX 0603IW-TR) for this test. They seem to be of similar brightness as the 1982 LEDs that are in my Beomaster 7-segment displays, and they have a very small package, i.e. should hopefully fit into the diffusers of the displays. For practicing and testing I used a "Surfboards" SMD breadboard for using SMDs on a standard wire component breadboard. On the picture the final result is shown. Even though this does not look perfect, they all worked. I used a low temperature solder paste, the lowest airflow setting (10) and a restricting 3 mm diameter nozzle on the blower tool. Temperature was set to 178C (the solder paste apparently already melts at 138C....). The entire process was surprisingly easy. The only real difficulty being not to blow the LEDs from the board while heating the paste up. Surface tension seems to put everything in place 'automatically' when the paste is hot enough.

Beomaster 8000 Back Together!

After an uneventful recap of the tuner section (boards 1&2) I put everything back together. Here is a pic of the finished boards:

Testing revealed no further issues except the missing display segment...I will take care of that soon. But first I need to practice how to solder SMD LEDs....

Beomaster 8000 Pre-Amp Re-Capping

Next, I recapped the pre-amp/input selector board (#3). This board can be completely taken out, which is what I did. I exchanged all electrolyte capacitors. One noteworthy detail on that board is that the printed capacitor symbols on the backside of the board for C120/220 (two tantal 10uF caps) are apparently showing the reverse polarity (this was the same in my first BM8000, i.e. I think this might be the case in all BM8000s...). I replaced the caps in the same polarity like the tantals, and everything seems to be fine. Here is a pic of the finished board:

Beomaster 8000 No-Load Current vs. Time II

After I installed the reservoir caps, I adjusted the no load trimmers one more time. Since the previous voltage vs. time measurements yielded a different behavior in each channel (left changed a few more mV than the right channel), I came to the conclusion, that adjustment in cold condition does not make much sense. I repeated this experiment a second time. This time, I waited the prescribed 30 sec after cold start and set the trimmers to yield 18 mV in both channels. Then I let the BM run for about 1 hr at zero volume setting. After the hour I measured a constant 25.3 mV in the left channel, and 22.4 mV in the right channel. The heat sink on the left channel was very slightly hotter reflecting the higher power consumption.
I thought, both channels should have the same condition under steady state operation conditions. Therefore, I decided to set the left channel trimmer that the no-load voltage would also be about 22.4 mV. This means of course that right after turn on this channel will run at a too low voltage (maybe 16 mV). However, since it took only about 3 min for the voltage to rise 3 mV in each channel after turn on, I do not think this poses any problem.

Beomaster 8000 Main Reservoir Capacitor Replacement

I finally received the 105C ECES1JA103EZ Panasonic replacements for the 10000uF main reservoir capacitors of the beomaster. They are almost the same size as the original ones: 80 mm long, but only 35 mm diameter (vs. 40 mm original). I wrapped them with a layer of corrugated cardboard to prevent bouncing around in the BM8000. The exhange required taking off the heat sinks and the aluminum cover of the control panel. The pictures show the steps for the left channel:

New and old capacitors:

Caps before exchange and after removal of the leads:

And after exchange:

An inspection of the old caps yielded capacitances within tolerance and ESR values of ~0.11-0.13 Ohm...not too bad. The new caps had 0.08 Ohm. Not sure if that means anything.

Beomaster 8000 No-Load Current vs. Time

I guess I was wrong about the voltage drift I observed during the left channel no-load current adjustment. I connected the right channel again to the power supplies and the voltmeter to R236/7 and turned the output stage on. I adjusted the voltage to 14 mV after about 30 sec. Then I observed the voltmeter over time. After about 40 min the voltage reached a constant 18.2 mV. It appears this voltage strongly depends on the temperature of TR211 in the constant current source that feeds current into the trimmer.
I blew some air on TR211 with a straw, and immediately the voltage decreased by several mV. Even air current like from an AC can have significant influence.

I repeated this experiment with the suspect left channel. A similar result. After about 20 min the voltage was at 19 mV and after 1:30 hours at 19.6 mV remaining constant there after. Same sensitivity to air flow across TR211.

I think this experiment shows that both output stages may be running correctly. I guess the rule to adjust the trimmers when the Beomaster is still cold assumes that the final no-load voltage spec is higher than 18 mV after the Beomaster has warmed up.

Output Stages Test/Recap

The next step was to check the main output transistors to see if there was an 'event' in the BM8000's previous life. I measured the resistance between E/C on all TIP141 and 146 on the heat sinks. It turned out that both channels are o.k....even though the left channel has blackened R236/7's, which is a sign for high current in these transistors....
I set out to exchange the electrolyte caps as well as the two trimmers in each of the boards (#5). The pictures show before/after:

Once the components were replaced, I hooked the board up to three external power supplies with +54V (to 'red' lead connector on the board), -54V (to 'black), GND (to 'blue'), and +15V to pin 2 of connector 38 (I really connected to R205 for easier access). The current limiters were set to 200mA on each of the three supplies. Before turn-on I set R226 (the trimmer that controls the voltage of the bases of the main TIPs) to zero Ohms. This turns the output completely off, i.e. the current into the board is lowest/has the least potential to create mayhem. Then I connected a voltmeter to R236/7. After turning on the three power supplies simultaneously, the currents in the three supplies were near zero. So far so good!!

Then I started increasing the resistance in R226 until the prescribed 18 mV appeared on the voltmeter. The output offset trimmer will be adjusted when everything is plugged back together and the amp runs on the BM power supply.

For better control I replaced the original corroded single-turn R226 trimmer with a modern 11 turn encapsulated cermet 100 Ohm trimmer. Since modern footprints are smaller, I had to extend the slider pin a bit. This is shown here:

Now I did the same thing to the left channel. Board up, capacitors and trimmers removed and replaced. Start-up with the external power supplies etc...I attached the voltmeter and adjusted the current. Like the other board the external power supplies showed the following currents after adjusting to 18 mV:

+54V--> 0.1A

-54V-->0.11A

+15V-->0.01A

Unfortunately, soon after turning things on, the voltage increased to about ~21 mV, indicating that something may not be stable...

I reduced the voltage to 15 mV and began to monitor. The voltage again crept up over ~20 min to about 17 mV...I turned off, and let things cool down. Again on: Initial voltage at about 14 mV, then over ~20 min again up to ~17 mV. The right channel did not show this behavior. I guess this warrants further investigation.

Recapping of Board #6 (Power Supply)

I finally received the axial capacitors for board #6. I replaced all electrolytic and Ta capacitors on board #6. The pic shows the board after reinstallation.

Replacing C6 was a bit tricky since it sits right next to the remote receiver, and one of its leads terminates underneath the shield below the board. This makes it difficult to unsolder it. However, using a thin solder tip, it was possible to reach the joint w/o needing to remove the shield. Other than that the work proceeded uneventful. After reinstalling I connected the transformer and plugged the BM in. All voltages were present. Encouraged I reconnected everything, and the BM still worked!

Next step: Evaluation of the output stages and replacement of the four main reservoir capacitors, which look a bit corroded.

Circuit Diagrams/Info for Vintage Bang & Olufsen...

In case you would like to better understand some of my descriptions referring to connectors etc..., I recommend to download copies of the original service manual/circuit diagrams.

There are a few nice sites across the internet aiming to provide technical support. Possibly the best and most in-depth is http://www.beoworld.org/. Unfortunately, one needs to become a paying member to download manuals...but it is great for discussing technical things in their fora.

For free manuals I recommend http://www.audiocircuit.com/Utilities/Manual/Bang-and-Olufsen. They allow a maximum of three downloads per day, and one needs to create an account.

First Light!

Encouraged by the low currents, I removed the external power supplies, and reconnected the internal transformer (P52), plugged the BM into the mains outlet and measured the voltages at the various connectors. Everything seemed to be as it should!

Encouraged by this test, I connected the microcontroller PCB to the power supply board (P48). and turned the BM on by pressing the "0" bar on the key pad. Life signs! Input selector and volume wheel resulted in the proper display reaction!

This looked pretty good, so I reconnected all plugs to board 6. Then I removed all power connections from the two output amplifiers (boards 5), i.e. I disconnected the red and black single leads (±55V) and the control voltages (P38/39). The purpose here is to prevent damage to the power transistors, which can be fried easily if trimmer R226 is corroded.  

I plugged the BM back into mains (I always unplug before changing connectors inside etc...), took a deep breath, and pressed the "0" bar. First light!! It seemed everything was working. Tuner responded, lights came on when filter, mono etc...were pressed on the control board. A healthy double click was audible a few hundred ms after turning the BM on indicating that the two relays in the power supply turned the big transformer on to supply power to the output stages. The picture shows a snapshot of the display. Everything looks good, but the right digit in the frequency display seems to have a dead element (it should read 102.1).

After this life sign, I connected my voltmeter to the output stage supply leads to check the voltages. I measured ±55V as prescribed...a good chance that this BM might be more or less working soon!

First Fixes

I opened the BM and I hooked up board 6 (5/10/±15 power supply) to external regulated DC supplies. I also disconnected all cables from the board to isolate it from the rest of the BM8000. The pic shows the connections directly to the main reservoir caps (C31/33/34).


I limited the current in each of the supplies to 250 mA. Then I turned them on simultaneously via a switchable power strip. Immediately the -24V arm went to zero volts (and 250 mA), indicating a short circuit. Inspection of the reservoir caps immediately showed that C34 was short circuited...nothing unexpected in these BMs!

I replaced all three reservoir caps, and C21/39/40 for good measure (they were those red cans that also like to malfunction after 30 years). There was a little snag when I tried to replace the big ones: They were glued to the circuit board with silicone. Heating up everything with my wife's blow drier softened the silicone, enabling removal w/o jeopardizing the cheapo easily crackable circuit board (one really wonders why they did not use fiberglass boards in these quite expensive receivers...). The pictures show the silicone blobs and the board after replacing the caps with new ones.


I ran the board again with all other connections not connected, and the currents now are ~30-40 mA in the ±23V arms, while the 10.4V supply showed ~10 mA (note that the grounds of the 5V and the ±23V systems are not connected when all plugs are disconnected from the board, i.e. a connection must be made between both grounds to run both systems with the external supplies at the same time-the picture only shows a ground connection to the ±23V system).

First Inspection of New Beomaster 8000

Today, I opened the fuse panel on the back of the cooling tower and it became immediately clear that there is a problem: Fuse F2 (500 mA, for ±15V and 10/5V systems) was blown. This means the Beomaster has to be opened up for an inspection!

My Second Beomaster 8000 Just Arrived!

Addiction in full swing! Yesterday, I received my second Beomaster 8000 in original packaging! Exciting! It appears to be in great cosmetic shape. But I also know it is not working...I cannot wait to take it apart and inspect its innards. Here is a pic as it sits in the box.