A small amount of ice was found in the bottom of the cold well, surrounding Thermocouple 0. It appears that water condensed on the walls, flowed to the bottom of the well, and froze.
The problem was found after reviewing the overnight data log, which shows that early this morning T1 (sensor on refrigeration tube) minimum values dipped lower and lower with each refrigeration cycle while T0 (sensor in bottom of cold well) minimum values struggled to descend below 0 C and remained there longer and longer with each cycle. This is consistent with ice surrounding the T0 sensor. The phase change from liquid to solid requires significant energy, explaining why the trajectory flattens near 0C, and lingers there after refrigeration is switched off.
The cart had been operating continuously for at least 7 days (168 hours), fully loaded with 12 RBC units and 12 FFP units. During normal operation, the cart will likely not store products longer than 48 hours. However, current normal operations are designed around wet ice coolers, which do not have capability to maintain temperature control for multi-day periods.
Neither water nor ice is desirable. They form conductive thermal pathways between the blood products and the cold well walls that move heat at a different (and faster) rate than air and therefore lead to freezing of the blood products.
Cold well dividers that provide a small air gap between the blood products and the bottom of the cold well. Any condensation will be kept away from the blood products.
Improved lid seal to reduce the temperature differential between the cold well walls and the units. This will allow the walls to be warmer, reducing the amount of condensation.