MSP430F5438A: Double variable looses precision

Hi,

while using double variables I encountered an unexpected behaviour. But first some words what I'm doing:

I'm sampling data with an ADC and to reduce the sampling noise, I am implementing a low pass filter in code. So far no problem. The problem is, that I never reach the final value due to numeric problems. 

float LowPassResult=0;
LowPassResult = (*LowPass) + (ADC - (*LowPass))/K_LP;
*LowPass = LowPassResult;

Debugging this code, I can observe the following:

LowPassResult has a value of 1.481625E+3, but in *LowPass it is saved as 1.481E+3. 

I don't understand that, cause LowPassResult and *LowPass have the same data type... Can someone explain that behaviour and/or know how I get the desired behaviour??

Thanks in advance!