Perti-Gyle batches

I'm just curious about how to figure out the gravity on a parti-gyle batch, namely a high gravity 3 gallon batch and a small beer 5 gallon batch. My thinking is that the 3 gallon batch will basically be the first running of the mash, and, using 2 quarts/pound of grain sparge calculation, get two beers from one mash and one sparge session.
I am deciding between a 3 gallon wheat triplebock/ 5 gallon dunkelweizen and a 3 gallon barleywine/ 5 gallon pale ale. Of course, I might do both.
Would a 13-15 pound grain bill do it?
I'm assuming OG around 100 for the big beer, and high 40s/low 50s for the 5 gallon batch (to be adjusted with DME if necessary).
Any reason this is out of the realm or plausibility?

here's a calculator that you might find helpful http://astrocaver.com/java/Parti-Gyle.html

I like to use Braukaiser's batch sparge/partigyle calculator. I've been doing the same thing you mention (3gal/5gal) for a couple of batches and its worked out great. I've already done an Imp Stout/Oatmeal Stout and an English Barleywine/ESB, and have the plans put together for a Weisenbock/Dunkelweiss.

To give you an idea, my English Barleywine/ESB was 19 lbs of grain and gave me a 3 gal batch at 1.090 and a 5gal batch at 1.050, while the Stout batches were 3gal of 1.095 and 5gal of 1.057 from 19.5lb of grain. I tend to save some of the grain back for the second batch (2lb or so) so I can up that gravity of the second batch a little and continue mashing it while I boil the first batch. That means I wind up mashing for an additional hour on the second batch, then sparging afterwards to get my full second batch.

Awesome. Thanks for the responses. I have to admit I'm still a bit fuzzy on the math,at least as far as figuring out what I'll get with my mash.
15# is about the max I can do on my equipment, so I will have to figure out if it will be enough; not that a brew in the 80s is bad, or that I can't add DME to get my gravity up. But I should probably keep the grist as simple as I can until I determine conversion.