Re: Waypoint memory after power cycle
Posted: Tue Sep 15, 2020 7:25 am
If you look on the processor board - you will find a chip labeled CY62... from Cypress. That is a 2MByte static RAM. Next to it you will find a 16374 chip.
The latter is used as address/data demultiplexer. It goes straight to the CPU (BGA, nothing much you can do there).
Using a very good magnifier or microscope carefully check each pin on the above two chips for tin whiskers or solder breaks - if you have access to a hot air reflow station (pencil type) and a drop of liquid flux reflow the solder joints.
Tin whiskers have been a problem since lead free solders have been forced on the industry - tiny (really tiny) tin whiskers grow for no apparent reason between two pins on a chip and eventually join creating a short - if there is little power on that line it will stay that way (otherwise the connection will just be destroyed as initially it is just a few atoms wide.
This has been becoming less of an issue in the past years though as the industry has learned to formulate solders and fluxes that help prevent it.
If they occur you cannot really see them with the naked eye - you need a good magnification.
It was a problem still with the boards we did for the old Enigma (This was very annoying) but has been almost not an issue anymore - but please check anyway - if say two address lines are shorted - you could very well have issues like you are reporting.
The RAM chip is permanently powered. The multiplexer not.
Rainier
The latter is used as address/data demultiplexer. It goes straight to the CPU (BGA, nothing much you can do there).
Using a very good magnifier or microscope carefully check each pin on the above two chips for tin whiskers or solder breaks - if you have access to a hot air reflow station (pencil type) and a drop of liquid flux reflow the solder joints.
Tin whiskers have been a problem since lead free solders have been forced on the industry - tiny (really tiny) tin whiskers grow for no apparent reason between two pins on a chip and eventually join creating a short - if there is little power on that line it will stay that way (otherwise the connection will just be destroyed as initially it is just a few atoms wide.
This has been becoming less of an issue in the past years though as the industry has learned to formulate solders and fluxes that help prevent it.
If they occur you cannot really see them with the naked eye - you need a good magnification.
It was a problem still with the boards we did for the old Enigma (This was very annoying) but has been almost not an issue anymore - but please check anyway - if say two address lines are shorted - you could very well have issues like you are reporting.
The RAM chip is permanently powered. The multiplexer not.
Rainier