Interesting though I think I won't be as simple as it may appear...

Many CPU problems are down to bad filtering of the 5V supply.
Check (or replace) the capacitor on the right side cooling, where the voltage regulators are.

Martin

Thanks! Tried that, but no. I do notice the 5V is 4.82V in reality, so 3.6% less than 5V.

Perhaps I should replace the 5V regulator, it's just a 7805..

rick

Replaced the 7805, voltage now is 4.98V, a bit closer to 5V, but still: nope.

Really odd thing there is something with a memory in this process. If I unplug and replug in, say 5 seconds, the Beomaster computer still is unresponsive. If I retry a minute later the processor works fine, for about 30 seconds, before hanging again.

This is consistent behaviour. So there is something with a memory that gives up after ca 30 seconds. And it's still messed up if you replug and retry too quickly.

Replaced C85, the 22uF capacitor on the processor board, too, still no.

rick

Dillen wrote:

Interesting though I think I won't be as simple as it may appear...

Reading the Tech.Produktinfo now (German), and it does not seem that horrible. It'll be quite some work, but it'll be fun too!

- the volume control is simply a 60 steps (6-bit) pattern, stored in the computer, controlling a volume ic. That pattern is in/decremented with the rotary encoder volume.

- the tuning is not as "digital" as you think it is, it is basically an analog tuning voltage generated by the microcomputer, by integrating an PWM signal generated by the computer into a voltage, then reading back the actual tuning frequency by measuring that, and adjusting while no match. Either controlled with the tuning rotary encoder, or by the computer. When the computer sets the tuning frequency it also sets a "fast" signal on the integrator circuit, so the tuning is achieved much faster than by spinning the wheel. It's all analog though :-) Generating PWM signals is something Arduino is very very good at. That old 6500 in the Beomaster is getting warm, I think I now understand why.

- FM frequencies are stored as a (serial) bit pattern in a TMS3529 non-volatile ram ic, controlled by the computer. (It uses a 33V programming voltage, switched on by a transistor, controlled by the computer.) Odd, I haven't found a single datasheet yet, though it's made by Texas Instruments.

- the display has basically 3 multiplexed normal 7-segment decoders, and a few extras. A fun extra is that they needed to display "PH" for PHONO, but the 7-segment decoder cannot display a "H". So what they did was display a "4", and light one extra segment with a dedicated transistor.. haha, great stuff!

- a bit special is that the various devices (beomaster, beogram, beocord) are linked, and they can control each other by sending serial bit patterns.That is what the extra pins in the DIN connector are for!

- the keyboard and the switches on the other panel (monitor, filter, manual/auto tune etc) are read by scanning/multiplexing with the computer (3 columns, the same as the display 7-segment multiplex signal, 7 rows).

rick

Way back when I determined that my BM5000 had a dead uC chip, I almost started a similar project. I was going to use a Teensy board though. Then I found an original replacement processor board and used that - it would have been impossible to implement 2-way remote control without a working system as a reference anyway. At least you don't have that problem, Good luck - I'm following with interest!

- Replacing the crystals did not fix it. Well, €1,50 down the drain. At least the processors still work.

- made sure the remote control signals could not reach the processor, in case that circuit is messed up, but no.

- unsocketed the TMS3529, that also has a direct connection to the processor, but, again no.

- grounded the input of the RESET schmitt trigger, so RESET=1: nope.

What now? Replace is the HEF40106 6-schmitt trigger? That is about the only thing left supplying the processor with inputs... (The tuner is disconnected, the rest of the processor board controls the display.)

rick 

(eyeing the Arduino Mega on the shelf here)

Look at Tim's solution!

Wow, he has replaced the individual 6500 processors!

I intend to replace the entire processor board with an Arduino, unless I get it fixed after all. That board is horrible anyway.

Arduino Mega has 54 digital pins, or I can use a regular Arduino with an I/O expander, e.g. with an I2C bus. Or, I'll use two Arduino Mega's.. (Arduino boards are so cheap these days!)

rick

@Dillen: Saw a post regarding the prescaler in the archived forum: /forum/general-discussion-and-questions/beoplay-e8-earbuds asking wether the u264b and u664b prescalers are  equivalent, Well, my Beomaster 8000 has a u664b, so the answer is yes.

The prescaler is something I need to keep, or copy, to measure the tuning frequency. So perhaps I will leave the processor board in place when switching to Arduino, although stripped.

Um, my schematic has just one ic8 HEF4013 dual FF for further dividing by :4 the :64 FM frequency generated by the 664, but on the actual board there's also a floating board with another HEF4013.. Does that further divide by :4 ? It does seem so, as the output of the floating board goes to pin 21 of the slave processor, which is the frequency counter pin, with the trace going directly to ic8 severed. So, not divided by 256, but by 1024? It would make the work for the slave processor easier.

I'll try replacing the two HEF4013's and the HEF40160 next.

rick