Texas Beomaster 8000: Sorting out problems with the microcomputer board

In the blog post where I checked out the Beomaster 8000 power supply I used a spare microcomputer board to run the Beomaster. The reason for that was because the original microcomputer board was not in good shape. During the recapping of that board I could see there were problems with it.

In this post I dig into the original microcomputer board with the goal of cleaning up bad solder joints, investigating some unusual board wiring and making sure the two processor devices are working.

The starting point of this task is with the recapped microcomputer board.

On the trace side the picture below shows this microcomputer board compared with a typical Beomaster 8000 board...in this case the microcomputer board from the Canada Beomaster 8000 unit.

The orange and green connection wires are what I expect to see on these boards. The other wires on the Texas Beomaster 8000 microcomputer board are unusual.

Checking them against a known good board I found that they do connect points that are supposed to be connected. The question is why and if they are necessary.  I started by removing one end of each jumper wire and testing if the connection was still good. In some cases it was so the wire really wasn't necessary. In several wires however the jumper wire was the only thing making the connection.

My suspicion is that board vias (connections between the top and bottom board layers) are not making contact.

Some of the jumper wires appear to have been made as a change to the normal wiring between vias on the component side.

One task on this board restoration is to replace the two crystal oscillators as I did on the Beomaster 8000 from Canada. Starting with that task will give me better access to board vias I want to resolder.

The following pictures show the removal of the two processor devices and the replacement of the crystal oscillator devices. The same care was taken regarding static discharge and the shorting of the oscillator leads while handling the processor devices.

When removing the processor chips I felt the sockets were showing their age so I removed the original sockets and plan to replaced them with modern, tulip style sockets. The pins are sturdier. This picture shows the old sockets (in the rear) and new sockets in the foreground.

While the processor devices and sockets were removed I installed the new crystal oscillators and added three jumper wires (blue, green, orange) between board vias as shown below. That is the normal way the Beomaster 8000 microcomputer board makes those connections. That eliminated three wires that were on the trace side of the board.

I also resoldered the vias involved in all of those connections.

After this first round of work I was left with the following...The green and orange jumpers that should be on the trace side plus three additional jumpers that I found I could not remove yet.
Note that the green and blue jumper wires were white wires earlier. I changed them when I started cleaning up bad solder connections. I did the same thing with the original yellow jumper wire.
I changed it to the white in the photo below.

Testing this state of the board resulted in a partially working Beomaster.

The standby LED dot illuminated and the keypad selection switch put the Beomaster in the power on mode. However the source selection did not appear to function and some displays were not working.
I swapped the two processor devices with some spares I had to make sure the processor chips were not the problem. It doesn't appear to be an issue with those two devices but I need to make some additional checks.

My prime suspect at this point is still connections between board layers. I want to see why those remaining three jumper wires have to be on the trace side.

For the green and blue wires (from two vias next to one of the oscillators) ... to IC7 I discovered that the board vias were the problem.

I removed IC7 to get to the component side of those vias and resoldered them there. I also resoldered the vias on the trace side. That resulted in those connections being good without the jumper wires.

Here is the original state of the microcomputer board trace side compared to the current state.
Note again that I changed the yellow jumper wire to a new, white wire when I cleaned up the ugly solder joints for that.

I am left with just the white wire jumper. That jumper is connecting the board ground between two devices (IC5 and IC6). With the wire removed those two points are not making a good connection.  The wire shouldn't be necessary so there is another via or two I need to check.

In this current state the microcomputer board still isn't functioning correctly. I can control quite a bit of the functions. The volume and tuning dials operate. The channel balance control works as do filter and tone control switches (for the three illuminated lamps below).

There are a couple of things I need to do next.

I want to put the original processor devices in one of my good spare boards to be 100% sure those devices are good.

I want to try and eliminate that grounding layer connection the white jumper wire is trying to do.

If finding and fixing the problem (that requires the white jumper wire) doesn't result in a fully functional microcomputer board then I will have to move forward on this project with one of my spare microcomputer boards and return to this problem board at a later date.

This microcomputer board is the only component in this Beomaster 8000 that is keeping it from moving to functional testing.

Beomaster 8000: Input Selector Trouble - Tuner Audio Mixed with Phono or TP1 Inputs (Pt. 2 - Conclusion)

I thought this was fixed! But it wasn't! Last year I reported about a strange intermittent issue with a Beomaster 8000 that I was restoring: When switching the input to Phono it sometimes mixed in the tuner signal that both turntable signal and tuner signal were audible at the same time...a pretty useless situation. My investigation back then determined that there was a discontinuity around P82 on the uProcessor board, which occasionally caused the input switches on the preamp board to malfunction. P82 carries the control signals for the two electronic multiplexers that control which input is routed into the preamp-circuit of the Beomaster. I implemented a bunch of jumpers, shown here:

and the problem appeared to have gone away.

As it turned out recently, this was an illusion as the Beomaster again showed this issue when I tried to play a lovely 1950s vinyl by Cal Tjader of which I bought a significant number recently. 

My initial thought was that maybe the ribbon cable that connects the input selector pins of the uProcessor to the preamp board had an issue, and so I ordered a custom manufactured 5-lead ribbon with appropriate 0.1" headers and put it in. Here is an impression of the end plugged into the uProcessor board:

And a photo of the preamp end:

While this new cable made a great connection and looked very pretty, it did not cure the problem. But at least another potential root cause was ruled out.

Back to the bench it went. I opened it up (again) and wiggled P82, and there it was (again)...depending on the pressure on the brand new connector the input behaved correctly or malfunctioned. I took the board out and had a closer look. All the jumpers were in good shape, so the problem had to lie somewhere else. However, all the traces checked out o.k. (again) and so I was a bit at a loss.

I put the board back in and played the Beomaster for a while scratching my head. Then it occurred to me that such problems are often related to ground connection issues. I prepared a jumper cable from the preamp board to the uProcessor board. Then I induced the issue and subsequently made a separate ground connection between the boards with the jumper. And, eureka, the input out of a sudden behaved normally. With the jumper in place I was not able to produce the issue again. I took the board out again and inspected it more. And I finally understood how the grounding plane is setup on the uProcessor board. I realized that this board is actually a double sided PCB, even though it has no etched traces on the back side. However, they used the continuous Cu layer as a ground plane from which several connections are made through the board to the traces. This is hard to see since the surface just looks a uniform green and there is no obvious mention in the manual where they show the PCB layout. After this it became clear to me how the ground connection to the preamp board was made. This shows the pertinent part of the board:

C67 connects through to the patterned side of the board and so makes the ground connection that connects to pin 2 of P82. The picture shown on top shows the backside of this area. I had soldered a jumper in from the C67 leg to Pin 2 of the header to bridge potential cracks in the trace connecting the two. What I did not consider was that the problem was rather caused by a 'cold' solder joint on the other side causing an intermittent contact. I resoldered the area and then the problem went away. Another demonstration that most control issues in the 8000 are caused by problematic solder joints on the uProcessor board.

Back to testing it for a bit longer, but I think it may finally be ready for its journey to its new owner in the UK.