Featured Post

Beolover SyncDrive: DC Platter Motor Replacement for Beogram 4002 and 4004 (Type 551x and 552x)

Late Beogram 4002 and the 4004 (Types 551x and 552x), which have DC platter motors instead of the earlier synchronous AC motors usually suff...

Friday, July 31, 2015

MMC Cartridge Box: A First Impression

I accumulated a few MMC20 and MMC4000/6000 cartridges over the years, and I only have a couple protective covers. I always wanted a better solution for keeping them safe. Protecting them definitely makes sense considering the prices they fetch these days in good condition. So I set out to design a 3D printable small box. First I took exact measurements of a MMC20 cartridge and built a model in Inventor. Then I created the lower part of a box moulded to the metal housing of the cartridge and a cover.

Yesterday I received the third version of the design from Shapeways, which finally seems to fit the cartridge perfectly, and also has enough friction between the cover to give a nice open-close feel. The issue with 3D printing really is that every technology and material behaves differently, while having specific tolerances and 'quirks'. So it takes a bit of trial and error to get a working product from a CAD design. Well, here are a few impressions:

This shows the empty container. The bottom part shows the characteristic slant of the bottom part of the cartridge. The front part is cut out deeply to make sure that the cantilever and tip never touch anything:
Pretty amazing the resolution that Shapeways can achieve. The slightly grainy surface is a result of their Selective Laser Sintering (SLS) process that blasts the plastic granules in the printing bed to fuse them into a solid body. 

And here with inserted MMC20CL cartridge. The cartridge body makes a nice smooth press-fit with the part to ensure protection even if turned upside down:
And closed. The cover shows the cartridge type on top:
I think I will redesign that a bit with bigger lettering. I will make a specific cover for each MMC20 and the MMC4000 and MMC6000 models that one can easily find the desired one. Shapeways also offers a bunch of nice colors, i.e. color coding will be possible, too.
Stay tuned for the final version...They will be offered at the Beolover Shapeway store.

Wednesday, July 29, 2015

Beocord 9000: Total Recap - Replacement of all Electrolytic Capacitors

I also did a full recap for my earlier Beocord 9000 restoration. While I think it is less urgent for Beocords and Beograms due to the generally lower operational temperature (than amplifiers/receivers), a replacement of all electrolytic capacitors is always a good idea if long term stability is desired. While the Beocord 9000 contains quite a few electrolytic capacitors, the good news is that its design is pretty service friendly. They definitely learned something between the 70s and the 80s. Even in comparison to the Beomaster 8000, the Beocord offers much easier access. In my opinion it is best to simply leave the main PCBs in the unit and instead remove the front and back panels. Here is a picture with the front panels removed. Their removal is straight forward and described in the service manual. Here is an impression of the top after removing the control panels and the glass cover:

The bottom PCBs can be accessed in a similar way by removing the bottom panel:

A good 'service position' for exchanging the capacitors is to put the unit upright opened like a book:

I put it on a towel, which allowed me to turn and move the unit while in this position. Once in this position it is time for some patience and concentration. One after another the caps are removed and new ones soldered in (in the correct polarity - as I said: concentration!! listen to some calm music and you will find some BeoZen in this..;-). Here are a few pictures:

Board #11 (DAC) is one of the two boards that need to be removed for the recap since they are sandwiched on top of another board:


The other one is #13 (Oscillator Detector)


The following boards can be done while installed in the frame:

#8 Microphone Amplifier:

Before (the shield needs to be removed; it simply comes off when unsoldering the two spots that hold it in place):
After (shield already installed again):

#3 Rec/Play back circuit:


#6 Rectifier

It is worthwhile noting that the big 5000 uF cap was very close to be out of spec. It only yielded 4000 uF. This is still good enough for stabilizing the supply, but it shows that doing the recap was probably a good idea.

#4 Power Supply:


#1 Microcomputer/display circuit:

There is one cap on the upper part of this board. This shows it after replacement:

#2 Dolby circuits:

O.k. done!

Tuesday, July 28, 2015

Beocord 9000: Restoration of the Peak Program Meter Scale Illumination

While the scale illumination light bulb was still working in the Beocord 9000 display that I am rebuilding right now, my customer decided to replace the bulb anyway to make sure that there are no unwanted surprises during return shipping. I can only agree since these light bulbs have a tendency to break their leads of during high vibration-situations (think speeding FedEx truck and a bunch potholes...;-). Once one is in there, I think it is best to simply replace everything and be done with it.
I did this procedure already once in the last Beocord 9000 that I rebuilt, which had a broken bulb.

Here are a few pictures to document the process for this particular unit:
This photo shows the original light bulb held in place by the 'cork' that sticks through the PCB

and from the back:
This is the PCB that I implanted. It is the same that I use in the Beogram 4000 to replace the position-scale incandescent bulbs.
It has almost the same width as the 5 mm bulb that they used in the Beocord 9000. I only use one red-green LED on the board for this application, i.e. I bridged the second LED position with SMD jumpers ("000"). It is important to replace incandescent bulbs with red-green LEDs if the light is being filtered or reflected by red or green features (like in this case where the scale is both green and red).

This picture shows the LED board inserted into the bulb space.

And from the back. It can directly be soldered to the solder points of the bulb, the polarity matches:

Here is a picture of the test drive:

On to the 7-segment display.

Monday, July 27, 2015

Beocord 9000: Replacing the RAM Battery

The Beocord 9000 has a 3V battery that supports the 16 word memory chip (an interesting MCM144102 chip that is mentioned as a peripheral memory unit in the Motorola 6804 manual, but whose detailed data sheet appears to be unavailable on the web). This chip stores the calibration settings (and may also back up the system time...not sure at this point - need to play a bit more with it). Here is a snapshot of the circuit:

The good news is that the battery is secured against connecting it with reverse polarity via D27. This suggests that replacing it with a battery holder instead of a soldered in tabbed version is a good way to proceed. 

The battery is located right underneath the TIME CAL indicator cabinet. The original battery started leaking the Li compound from the inside (that is the green-grey stuff around it). When I unsoldered the battery I suffered a small explosion due to this leaked compound. One should keep in mind that Li  is a highly reactive compound where a bit of heat immediately causes an exothermal reaction with the oxygen in the air. Luckily, I wore a magnifying visor and glasses, so nothing bad happened. But I would definitely recommend to cut the battery off with a wire cutter before starting to remove the tabs. Also wiping the area with ethanol would probably be a good idea to remove as much of the stuff as possible. Anyway, here is a picture of the vacated battery location after cleaning it as good as possible:

I used a 1/3N battery holder since tabbed 1/3N batteries are rare and pretty expensive. Also, a battery holder will enable to replace the battery in a few years without much fuss (the glass panel is removed very quickly without the need of a screwdriver, and then the battery can be accessed). Here is a picture of the new battery and the holder:

And the holder (Newark 25T0455) magnified:
Here is a picture after soldering it in. It is important to note the correct orientation that the + and - symbols match up with this on the PCB to avoid confusion down the road:

And with battery inserted:

Looks nice! Like it was designed that way (sort of strange to find a soldered battery in such a rarefied device, but then engineering of consumer devices usually assumes a 10 year use horizon...).

Beocord 9000: Restoration of the Peak Program Meter with SMD LEDs (III)

This is the third installment of my report on how to rebuild the peak program meters (PPM) of a Beocord 9000. The first two posts report on the installation of SMD LEDs to replace the original faulty LEDs on the PPM PCB, and the re-attachment of the plastic covers. This post demonstrates the installation and brightness adjustment of the rebuilt PPM.
The first step was to replace the original current limiting resistors with the correct values for the modern SMD LEDs. Modern LEDs typically need much less current and achieve higher brightness than the original LEDs. This is reflected in the much larger resistor values that my initial experiments yielded to achieve similar or slightly increased brightness with the new LEDs. The original values were 180 Ohm for the red and 120 Ohm for the green ones. The SMD types required 470 Ohm for red and 330 Ohm for green for slightly enhanced output (see first post). The resultant currents were in the 10 mA range, while the original LEDs easily draw 40 mA, while having a lower intensity.
Here is a picture of the original resistor bank (the 8 leftmost resistors)
This shows the new resistors installed. Despite the PPM having 16 LEDs there are only 8 resistors. This is due to the fact that the two left and right PPMs are multiplexed at 60 Hz, i.e. only 8 are on at a time. This is seen in the signals of the two common anodes (pin 1 and 2 on the display board) which each drive one of the PPMs:

This shows the new resistors installed:

And here after installation of the rebuilt PPM:
And in action:

It turned out to be a good idea to fix the two display boards in sequence, i.e. leave one in while taking out the other. this allows to realign the one that was out with the one that is still in when putting it back. That way it is easy to get it back into the right position that it fits perfectly into the plastic bezel. Another helpful thing to do is to mark on the headers of the displays how far they were inserted into the PCB. This is important to get the right slant of the displays that matches the cover.
On to replacing the scale illumination bulb with a red/green SMD LED.

Saturday, July 25, 2015

Beogram 4004: START Key Not Working

A customer from New Zealand recently ordered a set of Beogram 400x transport bushings from the Beolover Shapeways Store for his 4004 (5525) model. He sent me a picture after he installed them:

Pretty! I like them in white!

A while later he contacted the Beolover Customer Service Department ;-) about an issue with the START button on his deck. When pressed nothing would happen, while he was able to start the deck by first pressing the 33 button and then using the "<" button to drive the tone arm towards the start groove of the vinyl. Furthermore, the START button worked once the arm was on the vinyl. This indicated that the switch itself was o.k.

The fact that the 33 button worked told me that the circuit mostly worked, and that the fault must have its root cause in the SO-switch override circuit that is activated by the START button. Here is the relevant section of my annotated version of the circuit diagram:

This is how it works: When the carriage is in its 'home' position the SO (switch off) switch pulls the base of IC1 down which disables the 21V power supply, turning the deck off. When start is pressed, 30V are applied to the base of IC1 via R5/D9/R3/D6 (follow the red dashed line from the << ON (=START) switch). The 22V Zener diode D7 makes sure that the voltage at the base is 22V. This turns the deck on. D5 prevents a short circuit to GND via the SO switch, while this process is unfolding. The <<ON switch also turns on the carriage motor (via the red dashed line to he left at the branch point below the <<ON switch) and then the carriage starts moving. After it clears the SO switch the <<ON override is no longer needed. C2 makes sure that the override is in place long enough until the SO switch is open.

The 33 button provides another way to override the SO switch, but it does not turn on the carriage motor. Only the platter starts turning. That is why my customer had to press the < button to move the carriage while pressing the 33 button to drive it far enough until the SO switch was open. After that the deck would work normally.

The above considerations immediately suggested to look for a problem in the R5/D9/R3/D6 path. And indeed, there was a dead solder joint on D6, which prevented the override voltage from making it to the base of IC1. A bit of soldering, and the START button started working properly.
Wonderful! Another Beogram back in business! This is Beolove...;-)!

Friday, July 24, 2015

Beocord 9000: Restoration of the Peak Program Meter with SMD LEDs (II)

The 24 hrs test of the rebuilt Peak Program Meters (PPM) of the Beocord 9000 display that I am rebuilding was successful:

So I set out to reattach the red-green covers. Encouraged by my recent rebuild of a Beocord 9000 counter display, I decided to rely on melting the plastic posts of the red-green covers with my soldering iron set to 200C to recreate the original attachment method. First I needed to clamp the display together, however, to ensure that the covers would make a tight bond with the PCB to avoid light escaping through cracks at the bottom:

I applied the clamps in a way that the cover posts were still accessible and then I melted them by rolling the solder iron tip cone over them. This worked quite nicely:

When I cleaned the PCB with ethanol after soldering the SMD LEDs, the glue that held the light guide down dissolved. This told me that this clear glue was most likely some kind of polyvinyl ester based carpenter glue. So I used 'Aleene's Original Tacky Glue', which is my mainstay for bonding speaker foams to the cones. This glue dries clear, recreating the original look of the bonds:

After the glue was hardened, I put the display again to a 24hrs test to ensure that nothing bad happened during the reassembly:

What a beautiful design! I can't wait until I can make a test-recording with this deck using these lovely big rebuilt PPMs.

Beocord 9000: Repairing a Cracked Clutch Arm

This is a follow up to my initial entry about replacing the idler wheel rubber 'tires' in a Beocord 9000 tape deck. One of the issues I had was that the clutch arm showed a crack where the clutch shaft bearing is press-fitted 

It seems that many of these decks have this issue. A design weakness. While I thought that this is maybe not too dramatic since the strain was released via the crack and the bearing seemed stably in place, I felt that a proper restoration should not leave such a crack unattended. There is an interesting thread on Beoworld.org that was started by Sonavor, who is an expert on these decks. His approach was to buy some fitting copper pipe and cut off small rings that he was able to force over the cracked plastic part. While this is a really pretty looking and solid fix, the disadvantage is that one has to take the clutch apart. I did not want to risk to break the part, so I came up with a different approach:

I designed a 3D printed two-part ABS sleeve that I glued together under pressure using acetone-ABS slurry. This makes a virtually bulk material-strength bond since it dissolves the material at the interface and after evaporation of the solvent an almost uniform piece emerges.

Here is a picture of the parts together with the clutch:
While it was hardening I forced the parts together with a pair of grip pliers

And here is the end result:

I installed this in the Beocord and it works very nicely.