I have personally experienced the insulation failure mode regarding floating the center tap of a high voltage plate transformer. Back in the '60s, I was using a UTC S-46 plate transformer in a bridge configuration, with two oil caps in series, floating the center tap between the two capacitors. I actually trashed two of the new S-46 transformers.
I later used a UTC S-48 transformer, in the same circuit configuration, and never had a problem. (Even though it was not recommended, the S-48 probably survived due to its higher voltage rating.)
I had the opportunity to speak with a support rep at UTC in NYC, and, as I recall, he explained the problem regarding the insulation between the winding layers. The primary winding is applied to the core first, followed by some insulation. Then the start of the secondary is applied, two windings that each begin at the center tap. Subsequent layers of windings are added, with interleaving insulation, to arrive at the desired end-to-end voltage at the outside of the windings. He explained that the voltage between the individual layers close to the primary was not as high as the voltage at the outer end, therefore the insulation between the beginning of the secondary (actually the center) and the primary did not need to withstand the full voltage of the secondary. His advice: If it is desired to use a circuit where the center tap is elevated from ground, then select from the CG (Commercial Grade) devices that are specified for the additional insulation voltage tolerance.
There was actually a requirement that for the S series transformers, the CT should not be disconnected when in standby mode. I copy an excerpt of the catalog to this post, from page 49 in this link:
http://www.junkbox.com/electronics/utc_transformer_catalog_1963.pdfYou may be able to determine whether the overall insulation rating of the Heathkit DX-100 plate transformer is sufficient for your planned use by performing a hi-pot test. This test will show leakage at a voltage where the insulation breaks down, but since it is a DC test, it will not be possible to determine where in the transformer the breakdown occurs. This should not be an issue, so long as when using the transformer, the overall combined AC and DC potentials to not come dangerously close to the hi-pot test voltage. If you find this test voltage to be significantly in excess of the voltage you expect at the CT, then that specific transformer may be safe to standby the CT. If the hi-pot test indicates a lower safe voltage than the working voltage of the transformer, I would suspect an insulation deficiency between center tap and core or primary windings.
It is also important to note that floating a plate transformer is not analogous to floating a filter or modulation reactor. Even if the core of a plate transformer is floated, remember that there is a low impedance path from the primary winding to ground, so floating the core does not avoid the potential for insulation breakdown.
Personally, I believe it would be simpler, and more reliable, to bias the final and modulator to cutoff if it is desired to leave the plate voltage on continuously. Rather than decrease bleeder resistance to eliminate soaring, or add series resistance to the primary with a relay, you might consider using filter capacitors that can withstand the unloaded output of the power supply. With choke input, you still should not need any soft-start circuitry.
By the way, I have been told by several folks that the BC-610 plate iron has the same restriction, no bridge rectifiers, and don't float the CT. This apparently is a cost-saving design consideration on some transformers....
Good luck on the modifications.