If you operate with a stop at Heathrow, Reading and Bristol Parkway Wast (I propose a station in the middle of that giant diamond junction thing with split immediately to the west with the Cornwall line curving away in a tunnel) only, you could easily manage 90 minutes, it is only 340km via Bristol if you go directly from Bristol to Plymouth in a straight line.
You could possibly even add a third stop at either Tiverton (or Taunton) Parkway/Exeter and stay under 90 minutes.
As to building a line in Cornwall itself, many people would balk at the cost, but without it journey times to Cornwall don't really drop that much.
Do we want the beaches of Western Cornwall to be day trip destinations for London (with all the attendant benefits, including being able to shed the expensive to operate Sleeper services) or don't we?
EDIT:
If we assume my Option 'W' capacity enhancement on HS2 as the base, we have 18tph at London Paddington of which 8 will support the primary north-south High speed network.
This leaves us ten paths to play with:
I propose deploying them as:
1tph Hereford/Great Malven (leaving the HS network at Oxford, possibly dividing if this is required by the timetable to allow a second 200m unit to proceed to a secondary destination on the north-south network).
2tph Bristol via Bath (I propose 400m Classic Compatibles on this route) leaving the network at a slip line somewhere north of Chippenham)
2tph South Wales via Bristol Pkwy with a new Severn rail directly east of Bristol Airport.
2tph Bristol via Bristol Pkwy (using the Pkwy-Temple Meads line which will be captive cleared).
2tph South West via Bristol Pkwy (using said tunnel I mentioned).
1tph Spare.
This requires captive clearing on relatively little line (Bristol Pkwy-Bristol Temple Meads) and would allow the remaining destinations currently served by Great Western (chiefly, stations to Cheltenham and Westbury) to be served over the now rather emptier fast lines out of Paddington, alongside some sort of expanded outer commuter service for the THames valley.
Platform extensions to 400m at Cheltenham are likely to be extremely trivial, while at Bath the platforms could be extended to 400m alon the viaduct/bridges without major property acquistions other than car parking space and already public land (appears to be a road central divider or somesuch).
Bristol could have its platforms extended towards the south if the southern road bridge is replaced.... which isn't going to break the bank.
The high speed line would have a triangle east of reading, have an Ashford style station at Reading and then proceed towards Bristol, with a Parkway possibly being added near Swindon before the divergence of the Bath Spa slip line (this would be served entirely by Cross Country services from Oxford though).
The paths to Oxford would be provided by local services as now or by the 8tph that will be sent through the station, as well as the 1tph Classic Compatible that will divide there.
2tph to the South West would give us a lot of opportunities (especially if we add in 2tph "XC" services, 1 from each leg of the main network, presumably formed of coupled CC sets from various places for transit along the crowded Leeds-Birmingham axis).
We could call at multiple parkway stations and then have 1tp2h terminating at multiple places in Cornwall. (I do note that the Atlantic Coast Line has no real tunnels and only a handful of height restricting overbridges, almost all of them in rural areas where cheap humpbacks are acceptable, and that if it was cleared we can have captive sets terminating in Newquay, and all the intermediate stations are also cheap to extend). Therefore I recommend constructing GC cleared structures all the way to Penzance by one means or another.
It would be expensive.... but I think it is worth it because it completely transforms an entire side of the country that seems rather transport poor and which has so much potential.
(unfortunatley only having 9 paths for non through-Oxford routes is annoying but that could be solved if capacity really warranted it by another line, my Option 'M' from the main thread, which would reduce the neccesary Oxford-route paths to however many trains Oxford needs).
