Revival of an old thread - but at last I've got round to doing something that's been on my list for well over a year.
You may recall that the original demise of this machine was due to the lithium 'Tadiran' on PCB RAM back up battery having let out its juices and dissolving much of the Siemens CNC controller, so when I got it all working I mounted a remote battery in a leak proof case away from the electronics. This worked well but a few months (years maybe?) later we had an electrical storm IIRC that took out the little ram card that the battery had been mounted on.
This card backs up many setting and parameters of the machine and is vital to it's operation. However they are rare as hens teeth and come in several incompatible versions. I eventually found one that 'almost worked' - every thing was fine EXCEPT for a look up table that is stored on this card that translates error numbers to plain (ish!) English. Load sequence worked but errors still show as numbers not text. I spent weeks reverse engineering the RAM card - and manually proved that all the addressing and other peripheral logic on the card was fine - but what about the RAM chips themselves - was one faulty.
I then spent several fruitless weeks trying to make a RAM tester using an Arduino but never got a good setup, and the bits have been om my desk in the workshop for well over a year - glowering and nagging me.
Well I decided - hang it, source the RAM chips and change them ! They are four 28 pin surface mount NEC D43256GU-12L -I found a source of the chips - bought myself a cheap SMD rework station and . . . .bottled it ! My eyesight was never good and with anno dominae isn't getting better.
However today I braced myself and made a determined effort using a bench magnifying lens.
To remove the original chips I used loads of flux and Quick-Chip ultra low melting point solder. Strange stuff - extremely brittle and stays molten for a long time allowing the old bits to be removed. However due to this brittleness it is important to remove ALL of it before soldering in the replacements. I used normal Solderwick again with loads of flux, then cleaned up the mess with IPA.
Much watching of YouTube videos told me the virtues of 'drag soldering' - never heard of it before - but essentially you anchor the chip at opposite corners by conventional soldering with a very fine tip iron - then load the tip with excess of solder and drag it across the remaining pis, where the excess of flux and surface tension pulls the solder into the places that it's needed. Amazingly the process seems to work - I had a couple of solder bridges but they were cleared fairly easily.
So after a lot of flux clean up I re-assembled the RAM card with it's retaining handle and umbilical cord, and now I need to find the time / courage to try it in the machine. Lengthy re-load procedure so not today.
Still I overcame the SMD soldering issue and fears
