John, Tony,
My copy of the manual has U128 and U230 as 74S122 (not LS122), my memory says that most LS devices are rated to around 40MHz which is why I suggested HCT as a possible alternative. Anyway, until there is some evidence against them there is no reason to replace. The 470R load resistor should be a good test, a quick look at 74S series specs says that you might have to sink 4 x 2mA while keeping the output voltage below 0.8V.
A quick look on ebay (search 50MHz crystal) comes up with a 50MHz crystal oscillator in a 14 pin DIL outline for under ?3, you might find a branded one if you look a little harder. I have a similar 50MHz oscillator in an 8 pin DIL package, I will check its performance into 470R tomorrow!
You can, as a temporary measure, add another large electrolytic capacitor across C820 using flying leads to place the (much smaller) new capacitor somewhere convenient. The old capacitors dry out, going low capacitance and high series resistance. They don't do any harm, they just don't do what they are supposed to do. I think the 100mV ripple John measured was with part of the 5V logic circuitry disconnected so it may be much worse with everything connected up.
Roger
