Beolover provides professional maintenance and restoration services for vintage Bang&Olufsen stereos. We give one year warranty on parts and labor. All parts featured on the blog are available at our Store for DIY. Please, send an email for inquiries. Enjoy the blog!
Interesting! When I opened the shield around the processors on PCB #9 of the Beomaster 8000 I am currently restoring, I found this:
Note the green, blue and orange jumpers. They bridge traces that had vias in them on older versions of this board. These vias were not through plated but soldered in instead on those boards and prone to generate open circuits as the solder joints aged. A well-working remedy is to simply re-soldering them. However, I looked into my Bang and Olufsen Service Bulletins folder for the respective bulletin, which actually recommends to bridge them directly between the solder points at the ends of the respective traces. Here it is:
Anyway, this allowed me to directly move on to replacing the two decoupling capacitors on this board. Here is a shot of the board after replacing them:
Next steps are the main reservoir caps and the speaker switches.
Today I worked on the display PCB (#8) of the Beomaster 8000 that I am currently restoring. While I worked on the power supply board I noticed that the display segments flickered occasionally when touching the display board. The contacts between the processor board (#9) and the display board is really one of the weaker points of the 8000 internal design (no weak points on the outside, in my opinion!) and prone to issues. So I took the display board off and cleaned the contacts with a fiberglass pen and then coated them with DeoxiT D100. I also bent every second pin of the headers slightly to increase the pressure between the male and female contacts. This usually does the trick. Here is a pic of the contacts - it really does not take much bending:
I also replaced the two electrolytic capacitors on the board (there are only two - I'll let you find them...;-).
Before:
And after:
Detail (C7):
After I put the board back in, wiggling it would not produce missing segments anymore...hopefully this is now stable!
It is noteworthy that this display board features fiberglass PCBs for the volume, input and frequency displays - note the green color on the display in the upper right corner of the photo - previously, I only saw these boards used for the balance display, and these usually then had no missing segments. Maybe this explains that this Beomaster has not had display issues so far.
On to the inner sanctum, the processor board (#9)...problematic vias may need some attention there!
Another Beomaster 8000 just arrived. From ebay on route to Australia (yes, beolover went 'global'!...;-) it made a stop at my bench for some TLC. This was a better ebay experience. While the seller did not communicate in response to my requests for double boxing, he/she at least did a pretty solid job with a single box (and used FedEx - I am done with the post office since a recent disturbing experience where they dented the heat sink of a just repaired 8000...):
And it came with the original warranty document and an original owner's manual. Did I mention the remote? and speaker cables. Even an antenna is in the box. Very nice.
And check out the almost pristine veneer corners!:
Seems this baby comes from a home where people care about their equipment! I should keep it for myself! (just kidding!..My wife says three 8000s is enough!...;-)
Let's see what awaits us inside!
Today, I rebuilt the power supply of the latest Beomaster 8000. Nothing very surprising. Two of the reservoir caps were out of spec. C31 was at about 2000uF instead of 4700uF and it also measured a slight current when hooked up to my bench supply (this was the one that felt slightly warm when I tested the outputs the other day)...C35 was disconnected due to a broken lead, but it also tested 20nF (instead of 10 uF), i.e. it did not really matter whether it was connected or not. Since the Beomaster worked fine even with these caps in place, it appears that they are not essential...but the 5V supply would definitely have shown an out-of-spec ripple. Her are a couple of pics before and after rebuild:
And after:
I also cleaned all the pin headers on the board with a fiberglass pen and then slightly bent the pins alternatingly to increase the contact pressure.
Another Beomaster 8000 arrived for restoration. As usual, I opened it up for rebuilding the output stages before turning it on. This is a preventative measure against burning out the output transistors due to potentially corroded quiescent (quiet) current trimmers. And I like my 8000s with their original Texan transistors...;-).
I put in closed 12-turn cermet quiescent current trimmers, new differential input offset trimmers and new 105 C electrolytic capacitors. It definitely needed new caps. All of the 'red caps' except one of the 220 uF and the two 4.7 uF signal path caps were dead. Here is a capacitance and Equivalent Series Resistance (ESR) measurement on one of the 220 uF cans:
413 nF and 149 Ohm ESR is definitely missing the mark! This suggests that this Beomaster definitely needs a full recap. It is great that these red caps at least seem to die open-circuit instead via short circuits...Interesting to note that the axial leaded 100uF caps were all o.k....it seems it is always the red/orange types that go.
Here are the pics of the right and left channels before and after:
Right channel before:
and after restoration:
Left channel before:
and after restoration:
After putting in the new parts I measured the quiescent current trimmers, and indeed one of them had a too high resistance between swiper and track (~200 Ohm). This would have caused at least a pretty hot heat sink if not fried the transistors. But first things first. I hooked the output stages up to external power supplies and fired them up with the quiescent current trimmers turned to 0 Ohm (all the way CCW), thereby initially turning off the current through the outputs. See my post at http://beolover.blogspot.com/2011/09/output-stages-testrecap.html for details. Then I carefully adjusted the trimmers to yield the prescribed 18 mV across the emitter resistors (the big green wire wound resistors in the center of the photos, R236/7 in the diagram)
The currents into the the power rails were:
+55V --> 0.10A
-55V --> 0.11A
The +15V control voltage for the constant current source (TR208/11) yielded 0.01A.
After the above soft-ramp-up, I turned everything off, and reconnected the power rails, and the signal and control connectors to the output PCBs, and turned the Beomaster on. It came on without fuss and the radio immediately tuned in a local station. Both channels seem to work, and all other functions seem to be working, as far as I can tell at this early stage. Rather remarkable in light of the dead caps in the outputs. It will be interesting to see if many of the caps on the other boards are also dead, or if the ones in the outputs lived a more stressful life...they probably did experience higher temperatures than those on the other boards. At any rate the next step is to rebuild the power supply board...it seemed that one of the lower voltage reservoir caps got a bit warm during the above test, indicating a potential near-short in it.
Oh well, another setback!!... While messing with the output amplifier and doing the tuner dial string etc...I had to witness the degradation of the Buna N O-rings that I installed a month back into the Beomaster 6000's motor unit. I could literally see how cracks in the Buna N rings developed that increased in depth every day in a noticeable way. I think in another month the rings would have broken and fallen off by themselves. Here is a picture of the damage:
Pretty uncool! Anyway, this meant I needed to redo everything, and I already had the indicator bands installed.
I studied up on O-rings and the materials they are made from. It turns out that the best material for 'drive belt' uses under tension is EPDM rubber (ethylene propylene diene monomer rubber). I found the "Parker O-ring Handbook", and guess what, it has a chapter on "Drive Belt Applications" (section 3.17). There it suggests material "E0751-65" as ideal for drive belt applications. A bit of internet sleuthing then yielded that this is EPDM...unfortunately it is very difficult to buy Parker products as a private person...it seems they have a distribution system that is loth of modern purchasing methods, i.e they want you to go to a local handyman location to pick up the O-ring...crazy! So I went to the O-ring store, and ordered six of their E70032 model, which is a 1-7/8 inner diameter 1/8" thick ring made from EPDM. I also went through a learning curve with regard to the inner diameter - the original rings measured 2" diameter, and I installed the E70033 model first, which has 2" diameter, only to learn that two of the potentiometer drives slipped with these due to a bit higher resistance in these pots...so I had to do it a 3rd time with the 32 size...live and learn!
Anyway, I was able to finally complete the repair, and I made another video about the process. This video shows the entire process: Removing the indicator bands, putting in the new rings, and putting the bands back in. There is also a demo of the rebuilt unit in action at the end. You can find it on YouTube:
********Please, note that the video does not make clear how to properly position the indicator bands relative to the potentiometers. See this post for details.**************************************
Another interesting issue about this repair procedure was the brittleness of the V-shaped plastic clips that hold the indicator bands to the purple drive pulleys on the potentiometers. The original clips immediately broke when I tried to squeeze them a bit when reinstalling the bands:
So I took measurements of one of the still intact ones, and made a model in Autodesk Inventor. Then I exported the finished model as an STL file and printed it on a Makerbot II in yellow PLA (I thought the yellow would work well with the purple drive wheels, adding to the colorful interior of the Beomaster 6000....;-).
Here are the new clips:
Pretty cool! But I had to manually Dremel the grates off at the side that connected to the plate of the printer to prevent damage to the indicator bands when sliding the clips in.
This worked very well. They can be installed by squeezing them together with needle nose pliers and pushing them in from the top while bunching the band ends together into the 'crack' (sounds more complicated than it is - watch the video posted above - it shows this process in detail).
In a second step, the clips are pushed into the wheels until they seat against the wall:
And that is pretty much it. I really like the yellow-on-purple theme...;-). Watch the video for a demonstration of the rebuilt unit and the clutch mechanism.
Following excellent advice from Jacques ('charz') at Beoworld.com, I decided to also replace the quiescent current trimmers with modern 25-turn types. The challenge is to mount them in a way to achieve 'trimmability' with the output board installed in its operational position. Luckily the space between the two big 100n capacitors (C1/2 in the LF channel)is perfectly dimensioned to take standard encapsulated trimmer. The only thing left was to extend the leads to be able to bend them in the right orientation to be able to replace the old single turn trimmer. Here is a pic of a new trimmer in comparison with the old after this procedure:
It is important to orient the trimmer in a way that the quiescent current adjustment is in the same orientation as specified in the manual, i.e. current increases in clockwise direction. Otherwise, there might be confusion and hot heatsinks the next time this is done (in another 40 years??...;-). So I made sure that a complete counter clockwise adjustment means zero resistance between the left and center leads (in the orientation shown in the photo).
By the way: I used 500 Ohm trimmers since I was not able to get multi-turn precision 250 Ohm types through my usual channels in the US...I could only find single turn types...so I decided to use 500 Ohm. A look at the circuit diagram shows that there is nothing to worry about this. The only impact is that the base of 11IC3 (LF channel) can be pulled up a bit harder relative to the emitter if one would turn the pot all the way to the high current end due to the larger drop along the trimmer...but this is nothing one would do when following the quiescent current adjustment instructions in the manual (turn all the way CCW then turn on and slowly go CW until desired current is obtained). The 500 Ohm value does not affect the gain of 11TR2 either, since the collector resistor of 11TR2 is provided via 11R10 and 11R16, and the gain is theoretically close to infinite anyway due to the lack of an emitter resistor...Furthermore, the DC network for the base of 11IC3 is 'locked in place' (voltage wise) by the drop across 11IC3, which is defined by the quiescent current adjustment to the specified value, i.e. should be the same no matter if there is a 250 or 500 Ohm trimmer...
Next step was pulling the circuit board out again...certainly not a procedure I like to do. I hope this was the last time!
Here is a pic of the RR channel with new trimmer:
Pretty, how it fits! As if Jacob and his friends knew that I would want to do this in 2014!..;-)
And here a shot of the other three outputs:
And finally: The old 250 Ohm trimmers (which are all in pretty good condition..):
A test of the amplifier with the tuner revealed happy operation. This hopefully concludes the output amplifier restoration!
