OK, tried to get some numbers together. The graph is a touch "off" on the Y-axis due to mine own difficulties with Excel.
The original tables I've copied from are all available on Mike Williams' wonderful site.
First is the Boscombe Down test of HX908. They tested it to compare the performance of 150 octane fuel at +25 lbs boost to regular fuel at +18 lbs boost. The curve for the regular fuel speeds is at 95% of takeoff weight, with external tanks on and using saxophone exhausts.
HX809 is the starting point as the other FB.VI tested by Boscombe Down, HJ679, was described as being not representative in its tests there. I've used de Havilland's (not Boscombe Down's) tests of HJ679 as a reality check, though the Boscombe report for HJ679 does have useful information re: relative speeds with tanks on and off.
I'm trying to get from the speed curve from the HX809 tests (carried out with drop tanks attached and saxophone exhausts) to a speed with no tanks and with the ejector stub exhausts, then check it against another actual test.
So here's the HX809 test curve (note it's te left-hand one which is relevant; +18 lbs boost. The other, faster one is for +25 lbs boost, which requires 150-octane fuel, which is "another fine mess."
The next step is to "add back" the speed loss from having the external drop tanks attached. Here's the relevant test numbers from HJ679 (see above). The average speed loss is 5 mph (5.1 mph if you want to pick nits), growing much larger at higher altitudes. I believe these were 50-gal tanks.
The full test with the full range of speed diffrences at altitude is on Mike Williams' site - report for HJ679.
The next step is to find a value for the speed loss associated with sexophone vs. ejector stubs. I've posted the resulting graph on here before. The exact numbers for the speed loss for the range of altitudes tested is here, again from Mike Williams' site. Test was done on DK290, a B.IV, with max +9 lbs boost. The Merlin 25s on the FB.VI can use +18 lbs boost - I'll leave it to greater minds to speculate on whether there'd be any real difference in speed gain as a result.
Average speed gain for changing the exhausts across the range tested is 15 mph, 15.4 mph for nit-pickers.
So, then I put together a graph containing the original HX809 test, then a curve for the gain from dropping tanks, then a curve from the gain from using the stub exhausts, then a curve for an aircraft with no tanks and stubs, using the raw numbers from the tests. It's as near as dammit to simply adding 20 mph to the HX809 speeds.
To check for reality, I then put on a curve for another FB.VI, HJ679, as tested by de Havillands with no drop tanks and with stub exhausts. (Boscombe Down, as noted above, had complained that HJ679 was not performing as expected, so dH took it back and ran some more tests, which confirmed it had been about 15 mph too slow at Boscombe. The test data from dH for HJ679 with no tanks and stubs is from April '43, before the aircraft went to Boscombe).
As you can see from the orange curve, my calculations give a result which is very close to dH's, in fact mine are a few mph on the conservative side. There's only one data point with any real difference, and mine is lower than dH's). As dH tested HJ679 at 19,000-odd pounds, instead of 21,000-odd, this may account for the difference.
So, I get a deck speed of 352 mph, overall best TAS of 383 mph.
