Well, that's interesting. With the pump set to I (lowest) the water
slowly circulates around, but seems to heat up all the rads barring the
furthest (downstairs in hall, but fed from upstairs circuit) reasonably
quickly (burnyhot within 5 minutes). Boiler temp on the flow reaches
80C and sits there quite happily while it's burning, and with all the
rads wide, it happily sits there for a good long while burning at this
temp, with the pump set to I.

This includes the two rads on the ground floor circuit, so I don't
think it's just convection doing the work - I'd expect both downstairs
rads to stay cold if that was the case due to the way the plumbing is
at the point where upstairs and downstairs diverges (sideways T with
22mm continuing up, 15mm continuing down, 22mm sideways horizontally to
boiler).

However, with all the rads shut except the bathroom (and this wide
open), the boiler kettles. Same if I have the kitchen rad wide
(closest downstairs one to the boiler, and the only other candidate for
having the TRV disabled on it). So two wide, pump on min, kettles and
short cycles. Set pump to middle and it stops the kettling, though a
little short cycling going on, with periods where it's doing it right
(and modulating up and down in the longer burn cycle).

I'm not sure where this has got me. If there were no TRV's I'd set the
pump on minimum, balance the entire lot on the basis there would
probably be enough flow with that pump speed, and if not, knock the
pump speed up to half way, and balance for that.

But the TRVs mean that if I balance the rads, when they shut down, it
throws the *entire system* out of whack, not enough flow rate through
the remaining radiator, boiler kettles, and I'm back to having the pump
on full tilt (which leads to very noisy valves, and all the other
things too).

If I leave the kitchen open wide enough to stop the kettling (in
addition to the bathroom rad being wide open), then I can see in order
for it not to turn off on the TRV it's going to be open that wide
constantly, the roomstat will have to be set to turn off before the TRV
starts to shut, which then leads to the roomstat turning off
prematurely if a) the radiator's kicking out heat at a constant rate,
b) I'm cooking, c) for some reason it's warmer in there than another
part of the house.

Fair enough that another part of the house may have TRV's 'open', but
if the flow rate through those other 'on' radiators isn't at least
equal to the flow rate lost from the kitchen one having turned itself
off (or even down), the boiler's going to start kettling again...

I know the theory behind them, but... are TRVs really such a good idea
in practice? I'm *really* beginning to wonder if they're suitable for
this boiler and this house!