Post tds needs to be zero. The resins when exhuasted release ammonia, PO4 and other troublesome solids first as more highly charged ions bump them out. So even .01 tds from di could be a problem when it wouldn't be singnificant for just ro.
Test the membrane output with a tds meter. Typically, membranes have a rejction rate of 95% to 98%,so depending on your membrane's rating .I'd expect tds from the membrane to be 5 to 13 tds on an input of 250tds. If it's higher the resin will exhaust quickly.
Personlly, I use two di canisters with a tds meter in between them . When the output of the first one in line shows above 0, I change the resin in it and rotate it to second position. This insures 0 tds and eliminates the guesswork as to whether resin is exhausted; color changing is not especially helpful in gauging exhaustion ,ime. With two canisters and the tds meter in between I not only get 0 tds all the time but use teh resin more efficiently because I don't change it prematurely.
Also test after water is flowing for 30 seconds or so to get a fair reading to avoid reading the higher tds( tds creep) that can occur on start up. The in line meters are most effective when water is flowing.