Issue with the expected tolerance in the Low Power Timer test

[jeromecoutant]

While LPT implementation, I sometimes get some issue with the expected precision for Low Power Timer new test... (#5074)
I noticed that you created a DELTA to take in to account this extra time. It's good, but maybe still too precise...
It's OK will small duration, but becomes difficult with higher values

Example with a STM32 at 80MHz => DELTA=500
[1510235869.37][CONN][RXD] :332::FAIL: Expected 1.000000 Was 0.999302

Note that in tests-mbed_drivers-lp_timeout, precision is 100000 for the same 1s duration...

Could we increase the TOLERANCE_FACTOR or add an additional tolerance time depending on the timer value ?

Thx

@mprse @bulislaw

[mprse]

Fix has been provided in PR #5403.

TOLERANCE_FACTOR has been increased (multiplied by 2) for Timer and Low Power timer tests.
Please let me know if this solution is sufficient.