DateDiff in Measures

I have the following 3 measures in the Application User Table. This is connected to the AccessLog table via user id.

1stToLastLogin = CALCULATE(

MAXX ( TOPN ( 1, 'AccessLog', 'AccessLog'[AccessDate] ), 'AccessLog'[AccessDate]),

FILTER('Date','Date'[DateId]=SELECTEDVALUE('DateSelector'[DateId],MAX('DateSelector'[DateId]))))+0

2ndToLastLogin = CALCULATE(

MINX ( TOPN ( 1, 'AccessLog', 'AccessLog'[AccessDate] ), 'AccessLog'[AccessDate]),

FILTER('Date','Date'[DateId]<=SELECTEDVALUE('DateSelector'[DateId],MAX('DateSelector'[DateId]))))+0

RepeatTimeToAccess = DATEDIFF([2ndToLastLogin],[1stToLastLogin],SECOND)

On a power bi page I have a selected single date like the 25/02/2020 say.

So I get values similar to below : ( No real data is being shared here just sample data)

Appuser 1stToLastLogin 2ndToLastLogin RepeatTimeToAccess

2 25/02/2020 05:09:42 25/02/2020 05:09:32 10
2 25/02/2020 05:09:40 25/02/2020 05:09:38 2
TOTAL 25/02/2020 23:59:48 25/02/2020 23:59:47 1 --- ?

However the total for RepeatTimetoAccess is wrong . It does not add each individual row to give a grand total . It recalcualtes the total based on a max of 1stToLastLogin - a max of 2ndToLastLogin to give a completed nonsence high level figure .

I cannot use calculated columns as I am using direct query and the restrictions around max and min , and I understand datediff does not really have the sum x and other x aggregration functionality. What is the simplest way round this to give me a proper total RepeatTimeToAccess figure ?