Grain will increase for one of two reasons.
-overdevelopment
-reticulation
If the film is overdeveloped, grain size increases but this also results in increased density on the neg. Overdevelopment can result from increasing the developer concentration, increasing the development time, increasing the agitation, increasing the temperature. Small variations in these are cumulative so a small increase in agitation and temp can result in far more overdevelopment than you would think. Usually, though, you have them working opposite ways and they cancel out.
Micro-reticulation (I have posted about this before) is caused by thermal shock. If you plunge a film that you have processed at 20C into iced or hot water the emulsion contracts and produces a result like baked earth in a drought.
At less extremes of temperature (sometimes only 1 or 2 degrees) the effect happens at a much reduced scale. The result is a microscopic crazing of the emulsion that looks, even under enlargement, like bad grain. There is no increase in neg density.
If your negs are of similar density to your earlier ones then I would suspect micro-reticulation. In 90% of 'grain' cases this is the culprit.
If you have changed tanks your new one may be more thermally efficient and so loose heat quicker. Your dev - and therefore film - may be several degrees lower when you put in stop/fix at 20C and this is enough to cause low-level micro-reticulation and look like increased grain. The same thing will happen if the ambient temperature in the room is higher than the developer. The temperature of the film will increase above that of the stop/fix and thermal shock will result.
The Cure: A band from the 80's that was part of the New Ro...
Sorry!
Process using a water jacket (stick the tank in a big bucket of water at 20C) and monitor with a mercury photo thermometer. Keep all temperatures within 1C of each other. This includes the wash water!
Timings should be precise (within 10 secs). Measurements for mixing should be as precise as you can manage.
Agitation should be precise and consistent.
If you control all aspects of processing precisely you can then start making small changes in one aspect at a time to get optimum processing.
This will deal with micro-reticulation and over/under development.
I would suspect, though, that a water jacket should solve your problem.