Switch failure due to touch panel de-lamination seems to be "the" fatal issue for Newall systems. Objectively almost certainly not that common, there would be a fix or spares around if so, but a right pain when its your turn to be unlucky.
Hafta wonder how difficult it would be to strip a dead touch panel down and sort out a replacement using ordinary momentary contact switches. The laminated system is intrinsically sealed which is good in workshop environment but ultimately they tend to be less reliable than conventional switches as being unable to handle as many switching cycles. But thats the difference between a very large number and very, very large number something like 1 million as against 10 million.
Agree with John about the inherent robustness of spherosyn scales. In these Arduinio / Rasperry Pi et al and days I'm a little surprised that no-one has reverse engineered the output side of the Newall system. All pretty straightforward so long as you have a decent, stable, drive for the induction side. Think I'm right in saying that all the intrinsically really hard bits are in the scale read head. Differential transformers of any kind are not DIY beasts!
Clive.