In a T-pad, the shunt arm
has to be more than 600 ohms at zero attenuation. Think about it. The two series arms are 0-600 ohms, though, so you're two-thirds correct.
The data is there to figure this all out with only a small investment of effort. I know I'm starting to sound like Grouchy Gus

razz

, but I've also noticed that many are not willing to take the initiative and are just looking for a paint-by-numbers answer. You don't learn much that way.
A resistive attenuator in a circuit with known impedances is a fairly straightforward thing and it's a great way to cut your teeth in applying basic arithmetic to circuits.
Some people just wanna build a clone and don't give a damn about learning anything about electronics. That's valid, too, but I choose not to humor those people since nobody's paying me a dime to post here, and I'd rather focus on helping someone who wants to learn something about the art--as opposed to a project studio owner who's just looking to save a few dollars by building an SSL clone.
One last piece of advice: if you're getting custom pots made, might as well go with a bridged-T since it only requires two sections (and two external resistors). If running calculations for several steps of attenuation makes you cross-eyed, here's a web-based calculator which I've found to give good results:
http://www.microwaves101.com/encyclopedia/calcattenuator.cfm
Enter "600" in "Zo" and just plug in the attenuation values; 2dB steps should give you enough resolution to plot the resistances on an Excel graph, which you can then give to your pot manufacturer when specifying the tapers you need. The calculator doesn't accept an input of zero for attenuation, so you can start at a fraction of a dB to give you an idea of the total resistance of the shunt arm. (For instance, 0.1dB gives a result of 52K for shunt and 7 ohms for series).
OK, that's all I'm gonna say about it... I mean it this time.