Finally managed to get 100 watts output using a test tone. Found a problem with the SDR-1000 hardware code where the PA_BIAS control bit was inverted in the Linux code. This should now be corrected in the svn repository for DttSP.
Hopefully, I can now make a working transmit version available soon for others to test. I expect the CW transmit code will require a lot of work to get running well.