vanekp wrote:
What I find curious why does the LxCoPro work for most others and not for me? maybe I have something borderline on my BBC but what then?
As ICs age, their timing characteristics and logic thresholds will change. So no two Beebs will be identical. Also, as ICs fail, they sometimes do that in very subtle ways. You only have to read some of the Beeb trouble-shooting threads to observe that in practice.
The Model B is a particularly harsh electrical environment, because the data bus is unbuffered and there are (I think) 18 loads. The tube is simply an extension of the data bus, without any additional buffering.
Correctly designed hardware systems should be resilient to this normal variation (i.e. there should be a certain amount of margin for error), but there will always be limits.
In your case, something is pushing the system as a whole over that limit, resulting in marginal operation.
It could be any number of things:
- power supply noise (due to degraded electrolytic capacitors)
- the SY6502A being out of spec timing wise
- some other component starting to fail in a way that places additional load on the data bus
- the particular Lx9CoPro being slightly more sensitive to noise
Recently when working with Dave Hitchins on the Electron AP5 which includes a Tube interface we saw some examples where the Matchbox would fail, yet PiTubeDirect worked perfectly. Changing almost anything in the system fixed the problem. e.g. using a shorter cable, removing an unrelated piece of hardware, leaving the system on for a while, etc. After a lot of debugging, this turned out to be crosstalk from the data bus switching inducing noise onto the reset line into the Matchbox, causing the Matchbox to very occasionally be randomly reset. I'll try to dig out a photo of this happening.
This is just an example, it's probably not what happening on your system.