Can anyone optimise this SQL please

Hi,
I am a newbie to coding T-SQL for SQL Server 2005  and have written the following code. However although it runs <1 second on a 200 record set it takes >11minutes (and counting) to run on a 100,000 set. My final set will be >11million  so i really need to optimise it.
Can you help please?

/////////////
explanation of code
I have a number of incidents per device. I need to run a sliding window of 1 hr over all incdients per device. I want to get back a result set of the 4th FK_AddressID 's involved in a set of 4 incidents in any given hour. I thus use 4 columns to track the previous 4 values so can do a difference between the first and last per record per device to see if <= 1hr
///////////////

SELECT IDOfEAC2_Fault
      ,FK_AddressID
      ,StartTime
      ,ClosedTime
      ,EventATypeID2=2
FROM
(
SELECT DISTINCT FK_FaultHASCID AS IDOfEAC2_Fault

FROM
(
SELECT
      t4.StartTime AS StartTime4,
      t3.StartTime AS StartTime3,
      t2.StartTime AS StartTime2,
      t1.StartTime AS StartTime1,
      t1.FK_AddressID,
      t1.FK_FaultHASCID,
      DATEDIFF(ss,t4.StartTime, t1.StartTime) AS span,
      (CASE WHEN (DATEDIFF(ss,t4.StartTime, t1.StartTime) <=3600) THEN 1 ELSE 0 END) AS Type2_Outage
FROM Processing_FaultDurations t4
JOIN Processing_FaultDurations t3
on t4.StartTime<t3.StartTime
JOIN Processing_FaultDurations t2
on t3.StartTime< t2.StartTime
JOIN Processing_FaultDurations t1
on t2.StartTime < t1.StartTime
AND t1.Duration<300
AND t2.Duration<300
AND t3.Duration<300
AND t4.Duration<300
AND t1.FK_AddressID = t2.FK_AddressID
AND t2.FK_AddressID = t3.FK_AddressID
AND t3.FK_AddressID = t4.FK_AddressID
) t1
WHERE Type2_Outage =1
) retrieval
join Processing_FaultDurations
on IDOfEAC2_Fault = Processing_FaultDurations.FK_FaultHASCID