First of all, go and find the piece of paper that came in the Rodinal box. Now find ‘Diagram 2’. There are two graphs, one for 1+25, and one for 1+50. Being a little on the ケチ side, we’ll use 1+50.
Two things to initially note, i) the vertical scale (temperature) starts at 18ºC, ii) there are two roughly parallel lines, for 100 and 400 speed film.
Focusing on the APX 400 line, we see that it indicates that an increase in temperature of 3ºC reduces the development time by 4 minutes. If we assume that Tri-X will have similar behaviour to APX 400, it will have a (roughly) similar, and therefore parallel line on this graph. If you now draw a line, parallel to the APX 400 line, that passes through your known time and temperature (13min @ 20ºC, in my case) you can get a starting point time for higher temperatures.
Unfortunately the Agfa graph is skewed towards high values by the long development time for APX 400, but simply judging by eye, it’s easy to see that 13min @ 20ºC corresponds to 9/10min @ 23ºC. I’ve tried this with Tri-X and got results that i can’t distinguish from my usual efforts… as i’ve read that higher temperatures can yield more grain, i limited agitation to an initial 30s of gentle inversions, and two gentle inversions every minute.
Personally i’m not going to push this any further than 24ºC, as that’s the limit of the Agfa graphs. This doesn’t seem like it will be an issue… unless it stays really warm for several weeks.
Ice cubes? You’re all nuts – grow a pair and read a graph!