What looks like an absolute mayhem has a structure in it – working on revision B of the modified tape deck CPU. In the new PCB the BC858B do have now the correct orientation. I replaced the 1N4001 with faster BVY27/200 and removed the inverter which should have prepared the clock for the CD-CPU. Unfortunately this was not working. As a solution I was thinking of using the internal timer of the AT1284P as a supplier for the transistor circuit. I have got this finally working, the new panel is now plug & play and can easily replaced with the original CPU. There is only one cable that needs to plugged in.
Whilst I had to use a small piggyback PCB for the clock of the CD-CPU revision B also provides the clock. The only disadvantage is that the clock has now only 1MHz instead of 5.8MHz. To be honest, I have no idea why the clock is so high in the original one since the CPU idles most of the time. However, 1MHz is enough and so far I have not seen any drawbacks.
I still kept a led just for some testing purposes but removed unnecessary connectors. Also added the opportunity to use the I2C bus of the 1284P if necessary. This required some realignments of the pins in the source but finally everything is working smoothly.
If there would not have been the tape mechanism of unit 3 which has problems in fast forward/rewind since the motor is slow and takes too much currency which leads to the a reaction of the reel wheel sensor and the motor switches off. Will try some grease and hope the best…
