You have a really good grasp of what it takes to make Grade 91 work. You also have a really good grasp of what you need to do to improve properties. This, unfortunately is a rare thing.
SAW, there is no other way to put it, is a *****. Yes, it can vary a great deal depending upon welding parameters but is overall not going to be very good no matter what. Small wires, flat beads, rapid travel speeds, REHEATING OF PRIOR BEADS, are all critical for this stuff. I also stay away from agglomerated fluxes.
What temp did you test at?
Back in the olden days they didn't worry about CVN's much with this stuff since it is always used in a high temp applications. Who gives a rats ass about CVN's at 1000deg? The thinking has changed dramatically.
PWHT can be tricky with this stuff because while you are relieving stress you may also be getting precipitations that could cause it to go the other way. I have not confirmed this macroscopically but I have had long cooks that actually INCREASED strength over shorter ones with the same EXACT materials. I have no other explanation.
It is absolutely critical to get a 100% martensitic microstructure. You can do macros, but this is impractical in production for obvious reasons.
Thanks for the reply, I believe I have learned a little bit thanks to your previous nudge,
Test temperature was room temperature, about 23degC. Failed to mention that other mechanicals was great. I did notice one phenomena that also tickled me a bit. When performing only Base material heat treatment UTS and Hardness in increments of 3hours at 770Deg C up to 18 hours the was secondary hardening/strengthening but al values was >630MPa.
However the base material hardness 25mm away from the weld joint of the welded specimens exhibited a definite reduction in hardness for similar heat treatments.
This leads me to believe that the influence of the cumulative welding passes on base material properties, beyond the conventional HAZ region, was quite significant.
The aim is to do some TEM, SEM work on the mechanical specimens in due time to see if anything can be gained to enable reverse engineering to parameter setup. Once done I will send you th results, May hap try and publish a light hearted article...
Cheers