Surely, knowing that your water has negligible alkalinity is the one bit of information you can trust.
Perhaps a Chalk addition to raise the alkalinity and calcium content is all you need initially.
Another thought is malt, did this problem coincide with crushing your own. Long shot, but could a change in malt colour between pre-crushed and uncrushed be an issue with virtually no alkalinity?Similar to how they use very pale malt with Pilsen water.
I'm happy assuming negligible alkalinity, but I appear to be faced with negative alkalinity which is silly.
Chalk is one of those things I'm happy to be binning. It's unpredictable manner of dissolving just heightens the "where the hell am I" concerns. I've heard of people overcoming its insolubility by using CO2, but I think I give that a miss. Apart from sodium bicarbonate I will use Slaked Lime which adds Calcium too, but I am aware that adding Lime directly to the water can precipitate out CaCO3 (chalk!) and actually have the opposite effect to the intended one (
decrease alkalinity, not increase it, but I believe the water must be pretty alkaline for it to happen).
(EDIT: I should be more careful what I say - Slaked Lime can be dissolved in water more easily than chalk (it produces "lime water" for detecting CO2 in school lab experiments) but it isn't what could be described as "soluble").
Further checking back (notes can be handy can't they) tells me I was self-crushing my grain before the problem occurred, and since the problem I've used Warminster Maris Otter Malt, Crisp Maris Otter Malt and Crisp Chevallier Malt. Using the whole Warminster Malt pre-dates the problem (by just one brew) yet returned 73% brewhouse efficiency, which rather blows-out-of-the-water my idea that my current very low mash pHs and low mash efficiencies are related.
I've another brew on later this week, but as it's a low-alcohol offering with 1kg grain in 5L water (not physically possible to get a stiffer mash) so not very representative.