This is a follow up to my recent post about the redesigned Beogram Commander remote control board, which now works in both (DC-motor) Beogr...
Saturday, May 30, 2015
Like most Beocord 9000 and Beomaster 8000 units the Beocord 9000 that I am rebuilding right now has a broken 7-segment display. In this case, only one segment was missing, but of course this needed to be fixed, these big early LED displays are one of the stunning aspects of these units. I can understand that Jacob Jensen was fond of using them:
This display is very similar to the frequency display of a Beomaster 8000. However the design is somewhat different, so they are not interchangeable. The restoration process however is the same. I made many posts in the past about it, see here for example.
Taking the display out is slightly different due to the bezel that is clamped to PCB #1 that holds them in place and also prevents any light from the back and sides of the display to escape. This is superior to the earlier design of the Beomaster 8000. The first step is to remove PCB#11 and then the bezel. Then the display can be unsoldered. A desolder gun is essential for this task since all pins need to be able to move freely when it is taken out:
After opening the display unit up, the first task was to determine the correct current limiting resistors for the SMD LEDs (Newark 75K1441). The Beocord 9000 circuit has no easy way like the Beomaster 8000 for tuning the display intensity, so the original resistors need to be exchanged with values appropriate for the much more efficient modern LEDs. I went at it by replacing the dead segment LED with a SMD LED and then I connected the old LED with a 100 Ohm resistor to 6.2V (the spec value from the diagram). Then I experimented a bit with various resistor values on the new LED until I achieved the same brightness:
It turned out that a 470 Ohm resistor did the job very well to achieve the exact same appearance.
However, I always felt that the original intensity of these displays is a bit too low for well-lit rooms, and I prefer to give them a bit more intensity. A 220 Ohm resistor, it turns out yields exactly the same intensity as my typical Beomaster 8000 intensity adjustment:
The difference seen on the photo is actually stronger than it appears to the human eye...a digital camera artifact. The higher intensity allows a very nice appearance if the room is bright, but still does not look overly strong in a dark room. A matter of taste at the end, since it is easy to do one or the other. After this experiment, I went ahead and replaced all the original LEDs with new ones and performed my usual 24hrs test to make sure that things are stable:
This shows the current being drawn by the display during the test at 2.2V:
My work passed the test, and I went ahead and closed the display unit up. I had to deviate a bit from my latest black-gluegun-glue solution in that the display sits very tight in the bezel, which makes it impossible to put the glue on both sides. So I put glue in those locations that were accessible and did not hinder the bezel fit:
To add a bit of strength to the seal, I used the soldering iron at a 225C setting to widen the remnants of the original plastic posts that they would also help to keep the display together:
I think this resulted in a pretty strong overall bond between cover and PCB. One advantage of this design is also that the bezel already holds things together due to its snug fit.
Before I put the display back in, I replaced the original current limiting resistors. Here is a picture of the original setup:
My client decided to go with the brighter setup, so the 100 Ohm resistors of the seven segments were replaced with 220 values, while I used 750 Ohm for the 390 Ohm original values on the three dots of the display. Mathematically this is a bit too low (390 x 2.2 = 858), compared to the 100/220 pair, but it turned out that the dots look perfectly adjusted with it. They have a different situation in the display due to their much smaller aperture in the diffusor of the display. This is the new setup:
After this it was time to give the display a spin. I reinstalled it. It is important to put it at the exact right angle that the bezel can fit properly. This is easily achieved since the display aligns perfectly with the peak program meter unit. So I put it in, tacked it on with one pin, then adjusted it to align and then soldered the remaining pins. Here you see it with the bezel installed again:
Looks very nice and balanced! However, all this tugging and pulling on the PCB apparently killed the scale illumination bulb for the peak program meter. When I started out, it still worked:
Oh well...this would have happened soon anyway if a bit of wiggling and messing with the PCB broke the bulb. Another interesting repair task coming up! This is Beolove!
Friday, May 29, 2015
When I opened up the Beocord 9000 that I am currently restoring I found a loose part in the enclosure. The part is shown in this picture on the left:
It turned out that it is the left leaf spring (#12035 in the exploded drawing in the service manual on p. 4-6) that holds down the head carrier. The part on the right is the right spring that I extracted to take a measurement of the ball that takes care of the friction between the carrier and the spring when the heads are pushed into the cassette. The ball of the right side was of course lost...there are many openings in the bottom of the Beocord...
The ball turned out to be a 2.5 mm diameter ball bearing ball. These are widely available. I ordered the smallest batch I could find, which was a lifetime supply of 100 on Amazon.
The black block that is attached to the spring that I found is a broken out piece of the plastic assembly that holds the spring. This assembly is pretty weak, and so it is no surprise that the spring finally cracked the plastic. It is under constant tension.
Luckily it is a straight forward repair to reattach the plastic block and the spring. All that is needed is a 2-56 3/8" screw and a nut and a washer. This picture shows the screw already inserted into the plastic part. The 2-56 screw is tight enough that it threads itself into the plastic part, which is convenient since it allows the fixing of the leaf spring in place before bolting everything together:
Before putting in the screw it was necessary to sand the bottom part of the plastic block down a bit so it would not stick out beneath the hole in the tape mechanism carrier plate into which it fits to hold it precisely in place:
After this the spring assembly can be bolted through the carrier plate. The ball is held in place during the procedure by the grease in the groove (I used a dab of silicone grease). This picture shows the underside of the carrier plate where the sanded round flange fits through the alignment hole:
The next step is to put the washer and the nut on:
Tightening the bolt while aligning the spring completes the repair. Here are two views of the spring in its place underneath the erase head:
I tested the mechanism after I put it back into the enclosure. Seems this repair did the job.
On to the 7-segment display, which has a missing segment.
Thursday, May 28, 2015
It seems the Beomaster 4400 (2419) that I recently restored runs stable. I have it connected to my Beogram 4002 and I also tested it with the Beocord 9000 that I am working on right now, and everything functions very nicely. The final task was to replace the missing preset cover. There must be a big heap of these lost covers somewhere on this planet, but as long as we do not know where this Preset Cover Shangri-La is, getting authentic reproductions (injection moulded as the originals) from 'Dillen' (beoworld.org) is a great option. He recently sent me a few, and I installed one. Fits very well, maybe a bit tough to remove, but that will keep it (hopefully) away from their hiding place...;-). Here is a pretty shot of the cover:
Time to double box this Beomaster 4400 and send it back to its owner. That is the hard part of Beolove! Constant separation trauma...love can be pain...;-)
Wednesday, May 27, 2015
It turned out that fixing the broken out DIN 7 jack of the Beocord 9000 that I just started to work on was easy. The part is exactly the same as they used for the Phono input of the Beomaster 8000. Last year I developed a 3D printed part that can resurrect these sockets when their flimsy tabs break off that normally prevent them from falling into their panel cutout. When the tabs break off due to the force applied while inserting a cable plug the jack gets pushed into the enclosure. Here is a picture of the broken jack:
These are the 3D printed tabs that replace the broken off parts:
I made a short video that shows how to install them and gives a demonstration of the repaired plug:
This shows the end result:
After the DIN jack was fixed and the tape drive was installed back into its spot, I finally gave this Beocord its first spin. I made a recording using the Beomaster 4400 that I restored recently, and I played back a nice mix tape of Donald Byrd tunes from the early 70s that I had recorded a while ago on my own 9000. Lovely!
However: One segment of the display does not work,
and I also found a 'mystery piece' in the Beocord enclosure when I initially took out the drive:
The left part is the one that I found lying at the bottom of the enclosure. The right part is its counter part from the right side of the head carrier assembly. The broken part holds down the left side of the carrier, i.e. it is mandatory that it gets fixed. Otherwise, the heads might be in a non-spec position. I found out about this part in an excellent thread on Beoworld that was started by 'Sonavor' when he rebuilt one of his eight Beocords (kudos to owning that many! It appears, and this should please my wife, I still have a long way to go until my addiction reaches this level!!...;-).
Unfortunately, I was not able to find the 2.5mm bearing ball that goes along with the spring. The ball is needed to let the head carrier slide beneath the spring without generating too much friction. Luckily, I had the second ball for taking measurements and I was able to find small batches of these balls on Amazon Prime, so this may get fixed in a couple of days.
Tuesday, May 26, 2015
I was recently sent a Beocord 9000 for some TLC. It came in its original B&O box and even double boxed. Unfortunately, whoever packed it put the B&O box upside down into the external box, so it probably did the trip upside down. Here are a couple pictures of the package. The original box saw some water damage at some point:
But the Beocord 9000 sat safely inside:
I never plug units in directly after unboxing, I rather like to open them up first to do a visual inspection. So I took it out of the box and opened it up:
I really like how accessible these units are. No comparison to earlier model series. Two bolts at the back, and then it can be opened up like a clamshell. All that is missing are hydraulic shocks to open it up automatically..;-).
These decks usually have degraded belts at this age, so I unbolted the drive mechanism and flipped it around. This is a detail shot of the drive section:
Three bolts hold the drive to the chassis. After taking them out it can be carefully turned around without having to unplug anything. It is a good idea to mark the position of the drive before loosening the bolts. This helps to get it into the right position when putting it back in, which is important for making the proper connection to the eject button:
It is a good idea to cushion it with some cloth to avoid damage to the circuit boards and the delicate drive mechanism itself.
And as expected the belts were gone:
The curious thing about these belts is that they literally melt. This does not happen to the belts they used in their 1970s designs. Those belts just crack into pieces. But these here melt:
And what a mess this is. Since the belts seem break up into pieces and then go slimy, one can expect that the entire mechanism and the sections beneath the drive are covered with gooey nasty stuff that really sticks well to your hands and the wiring around the drive etc...The only good news here is that the stuff can be removed with ethanol. It is a good idea to wear gloves and have an ample supply of paper towel handy. This stuff is definitely messy! In this particular unit, someone already had a look before me, and that explains why the power cord and the DIN-to-RCA cable it came with were full of that black stuff. It just gets everywhere.
I removed the shield and the retaining plate for the flywheel and cleaned everything:
A lot of the black mess was on the drive pulley. I scraped it off as good as I could:
After that I put the pulley into ethanol over night
This removed most of the black goo. The last remnants came off when wiping it with some cloth. The it was time to install the belts. It is remarkable how easy it is to get the new belts installed in comparison to the Beocord 5000, where this process turns into a Saturday afternoon adventure. Here is almost as easy as popping in a tape into he deck: They just go on. The only thing to remember is that the flat belt goes first. Here is a picture with the belts installed:
After I was done and had put the metal plates back on the drive mechanism, I realized that the DIN 7 jack of the deck had the usual broken tabs and was hanging loosely in its panel cutout.
The jack is located right below the drive mechanism, so I decided to fix it before putting the drive mechanism back in. Luckily, this jack seems to be the exact same design as the Beomaster 8000 phono plug, one of which I fixed last year with 3D printed inserts. All I need to do is get some more of these inserts printed, and this should be fixed nicely. This will enable me to plug in a DIN cord without fuss for testing the functions of the deck. It is pretty difficult to get a plug in, if this jack is loose without opening up the deck to push it out, while plugging the cable in.
Saturday, May 23, 2015
It seems that the Beomaster 4400 that I am currently restoring is now mostly working. After I bridged all the cracked traces on PCB#6 I finally had a stable Phono input and a working AFC. But then I realized that the right channel Phono input did not yield much in terms of low frequencies and was lower in overall volume compared to the left channel. The tape inputs were working, so this indicated that the RIAA equalization on that channel did not work properly. Here is the relevant part of the circuit diagram:
I added some benchmarks measured on the left channel in the diagram. Input was 10mVpp. This yielded ~3.9V at 100Hz, 900mV at 1kHz, and ~190mV at 10kHz, corresponding to an about 20x drop between 100Hz and 10kHz (=26dB). This conforms with the RIAA curve as shown here (from this very interesting article about different ways to achieve RIAA equalization - another proof that anything can become a complicated topic if one only digs deep enough...;-):
The red curve is the one that the Phono input needs to achieve. What is coming from the record follows the green curve, i.e. low frequencies are recorded at a much lower volume than high frequencies.
The main reasons for doing this are (1) to increase the playback time and (2) to reduce high frequency noise. Lower frequencies need more space in the groove for the same amplitude output than higher frequencies. This is a consequence of the fact that the current induced in the moving coil of the cartridge is based on the speed at which the iron moves relative to the coil. If one would want to achieve low freq reproduction at the same level as high freqs, the grooves would need to be much wider, which would significantly reduce the play length. By accepting a lower amplitude for low freqs the grooves can be kept narrow and so the record plays longer.
Higher frequencies benefit from being cut at higher amplitudes since that gives a larger signal-to-noise ratio at the upper end of the spectrum. The 'natural' noise of records comes from roughness of the vinyl material, which generates a noise floor. Since during playback the RIAA curve reduces the high frequencies relative to the lower frequencies, the vinyl noise is reduced. This is a similar concept to Dolby NR where low volume sections of the recorded signal are amplified on the tape, and then reduced during playback, which also reduces the tape noise.
Anyway, anything that changes the signal depending on frequency immediately 'smells' like a capacitor issue. The two capacitors in the Phono preamp circuit that are responsible for the RIAA equalization are C105 and C104. And sure enough one of the corresponding units on the right channel circuit, C205, gave me a strange reading. My capacitance meter yielded a fluctuating signal, while C105 yielded a solid 10nF reading.
So I put in two replacements (it is best to do such things in pairs to keep the channels even). Since I did not have 2% tolerance caps available, I went through he batch I had and measured them until I found two that had the exact same value of 10.2nF, i.e were in spec. I soldered them in:
And tested the unit with a 10mV signal on both inputs. The signals were very even on both channels through the entire frequency range. Very good! I put things back together and hooked up the Beogram 4000 that I just restored, and pure bliss! I listened to Autobahn by Kraftwerk on the original German first pressing double album, and all was good!
While I listened, I decided to explore the topic a bit more and I made a model of the Phono pre-amp in iCircuit (an absolutely great app for spontaneous circuit explorations - not as powerful as a Spice simulator, but so much more pleasant to use - give it a try: iCircuit costs less than a couple high-cal drinks at Starbucks..;-). Here is a screenshot of the model:
The simulation pretty much confirmed the measurements that I made on the 4400 unit. I rearranged the circuit components a bit to make more clear how this RIAA equalization works.
The voltage divider at the output formed by the three 36.5, 220k, and 470 Ohm resistors and the emitter resistor R103 (330 Ohm) determine the gain of the amplifier via feedback. Without C104 and C105 the gain of the amplifier is ~460 across the entire frequency spectrum, i.e. 10mV at the input generate 4.6V at the output. The RIAA equalization is produced by adding the two capacitors that reduce the 36.5k and 220k resistor values depending on frequency. This reduces the upper end of the voltage divider more and more with increasing frequencies, hence, increasing the feedback and in consequence reducing the gain. This has the effect that the output of the amp is reduced at higher frequencies, and we get the RIAA equalization.
In this context I was wondering about the failure mode of C205. I think it must have partially short circuited the 220k resistor causing a reduction of the base gain of the amplifier. Since this affects the low end more than the high in this circuit, it appeared to the listener (me) as if the low end was missing.
Anyway, this seems to be fixed now! Kraftwerk! Autobahn!
Wir fahr'n fahr'n fahr'n auf der Autobahn
Vor uns liegt ein weites Tal
Wir fahr'n fahr'n fahr'n auf der Autobahn
Vor uns liegt ein weites Tal
Die Sonne scheint mit Glitzerstrahl
Die Fahrbahn ist ein graues Band ...
This is Beolove!
Friday, May 22, 2015
After performing all necessary restoration tasks on the Beogram 4000 that I restored during the last few weeks, I put the aluminum panels back on. I quickly realized that the platter was not flush nor centered relative to the surrounding aluminum panel. This meant I needed to adjust the chassis and platter positions. The service manual prescribes to adjust the platter first relative to the chassis, and then adjust the chassis poison relative to the enclosure, i.e. the aluminum platter surrounding the platter to create a situation where the platter is flush with the panel and also centered in the opening. In other words the task was to achieve a Beoloving situation where the Beogram looked just perfect!
I made two short videos describing the the process. Here is the first one, that describes how to adjust the platter relative to the chassis:
After the platter is adjusted the chassis can be aligned. This procedure is shown in the second video:
Well, this pretty much concludes the restoration process. I am almost sad, since I really got used to this most beautiful piece of engineering, and I will have a hard time to let this Beogram 4000 go back to its owner! I am grateful to Jacob Jensen for coming up with such wonderful designs. They enrich my life and they also got me turned on to electronics. It has been a wonderful experience the last few years! Of course I never met him, but it feels like I just lost a friend! Rest In Peace! It is Beolove!
Thursday, May 21, 2015
After I replaced all electrolytic capacitors a preliminary test of the Beomaster 4400 (2419) that I am currently restoring yielded that the phono input would only work intermittently. Further testing revealed that also the automatic frequency control (AFC) did not work. Furthermore, in my first happy excitement I also did not understand that the tuning lights are supposed to be off during non-FM listening, while the signal meter is supposed to be at zero. Both conditions were not met when I switched the unit to Phono.
After some measurements I figured out that the muting function, which should only operate when FM is selected and mute the tuner output in-between stations, came on intermittently, which muted the phono signal resulting in a fluctuating phono experience.
In combination with the lit tuning lights and the non-zeroed signal strength meter, this suggested that the tuner still had an issue. When I started working on this unit, the tuner did not work at all. It turned out that PCB 6 has a crack which prevented the tuning voltages to be applied to the front end. I jumpered the tuning voltage traces on this PCB and thought I was done after the tuner started working. But not so! After consulting the relevant section of the circuit diagram,
I realized that all these problems would occur if the ground connection of the preset and FM switches were disconnected (left center of diagram). This ground connection becomes active when all FM switches are in their off position. In this case, the signal strength meter ("radicator") and the power supply of the tuning lamps (blue dashed box) are turned off via D25 and D18. A look at the PCB schematics
immediately suggested that the crack in the PCB was a bit longer than I had assumed. Since this part of the board is not easily accessible I was not able to assess the damage initially. The black bar in the above picture indicates the approximate location of the crack. It disconnects the GND/black lead from the FM buttons. It is obvious that this also disconnects the AFC switch. This explained why AFC had no impact on the tuning experience (it helps to find the center of the frequency band for a station).
After I realized the extent of the damage and the number of jumpers that would be necessary to fix these issues, it was obvious that the front panel of the Beomaster needed to be removed a bit for better access to the solder points of this PCB.
So I started to disassemble. It is not too difficult an operation to pull the front plate one inch away from the main PCB. This gives enough access to allow some careful soldering activities on this board. The first step is to remove the 'slide rule'. Once it is removed the screws that hold the front panel to the enclosure can be accessed:
Here a detail shot of the small adjustment wheels that are in the volume and FM dial sliders. They come out easily, i.e. one should be careful when removing the slide rule to not lose them:
After that was done and the four screws were removed, the two PCB clamps (pic below shows one of them below the resistor) that hold #6 to the main PCB #5 needed to be pushed through their holes:
After that was done I carefully pried the front panel away from the main board by an inch or so to gain some access to the top of PCB #6:
The picture shows the jumpers installed (red, blue and grey). After the jumpers were in, I gave it a quick test to make sure everything was fixed. It was. So I put everything back together.
Then I hooked the Beogram 4000 up again and put one of my favorite records on: Miles Davis - In a Silent Way. I have the MoFi reissue, which sounds delicious. Unfortunately, while the muting problem was gone, I realized that the two channels were not balanced. The left channel sounded great, but the right channel had very weak bass. Only higher frequencies made it through. This indicated to me that there may be an issue in the RIAA equalization in the phono pre-amp. Oh well, back to the bench for some more testing!