The issue is the amount of protrusion of the heater matrix tubes above their support plate within the FOHE. These are notoriosly difficult to manufacture and the protrusion height is largely driven by their method of attachment, i.e. brazed joints.
If they protrude above a certain dimension then it is likely that any upstream release of a "snowball" will not pass cleanly through the FOHE and will "stick" or "cling" at the FOHE heater matrix inlet plate and hence restrict fuel flow.
The solution is to provide a bypass within the FOHE, similar to that provided for Fuel Filters, which is where, I believe, RR are heading.
The problem is that the current EASA fuel icing cert requirements do not include testing with fuel at a temperature conducive to "sticky ice", ("Snownballs" or "Sticky Ice" occur in fuel at or around -8°C). However sticky ice does not seem to be an issue if, as above, heater matrix tube protrusion is below a certain value.
ps, according to Today's Telegraph (online) the FOHE is there to cool the fuel................
