So what you would have to do is to capacitively couple this hammond to the grids, drop two high value (100k?) resistors one from each grid then the center point going to the bias circuit. This would permit only AB1 operation and keep the transformer from ever having to deliver current... one would likely want to drop a resistor across the secondary to provide some sort of resistive load as well... CT to ground. drive it with whatever you like... you just need the requisite voltage swing.
_-_-bear
EDIT: another thought - the exact ratio of transformer for this application is almost irrelevant. What is needed is the ability to drive the grids into AB2 (current), which implies a suitable ratio (to get the required voltage) and enough current on the primary to back it up. Then the core size needs to be sufficient for the driving power - what did we say a watt or two? The enough current to back it up is really the key, which is why there was a suggestion to use a small power amp. There's no DC on either side of the the thing, so nothing exotic or fancy is needed in terms of a transformer, just the turns ratio that results in the source's swing providing the required secondary side swing. Of course you want halfway decent frequency response, but that ought to be reasonably non-critical.
My thought is a "line to 70v" transformer - the type used for distributed speaker installations. Some have a "140v" tap as well, I have a few of them... they're cheap and available, have good freq response and come in various "wattage" sizes. Drive the primary with an "8 ohm" little power amp, and ur good to go... I think I've found a CT in the winding scheme on the multi tapped ones I have... and IF the "CT" is off a little, you could set it up so that you automatically get a bit more postitive swing than negative!
_-_-bear