[ARM] Environmental testing of Unit inside Chamber

Environment Test
Part Numbers Used:
RAM : MT48LC16M16A2TG-6A:GTR
MCU: LPC1788FBD208

Day 1: Cold Cycle Test
Entire unit( PCB Board fitted inside the casing ) is kept in a chamber and following tests are performed.
step 1: Chamber Temperature is Brought down From Ambient temperature to -5°C(Takes about 30 Min)
step 2: It is maintained at -5°C for 2 Hours
Step 3 : Now the unit is turned ON for 2 Hours.
Step 4: Now the temperature is brought down from -5°C to 0°C and unit is kept in OFF condition for 2 Hours
Step 5 : Now the unit is turned ON for 2 Hours.

Day 2: Dry heat Test
Step 1: Chamber Temperature is Brought up From Ambient temperature to +55°C(Takes about 35 Min).
Step 2: At +55°C , unit is turned ON for 2 Hours and kept OFF for 2 Hours
Step 3 : At +50°C, Unit is turned ON for 2 Hours and kept OFF for 2 Hours
Step 4 :Chamber Temperature is bought down to ambient Temperature (Unit is in OFF condition)

Day 3 : Damp Heat Steady State
Step 1 : At +40°C, 95%Rh , Unit is Kept in OFF Condition for 24 hours.
Step 2 : Just 5 Minutes Before completing 24 hours, unit is Turned ON, check for its proper functioning and then turned OFF. While performing Step 2, Unit Failed to turn ON.

NOTE : In The Program Initialization, Step for turning ON the unit comes after initializing MCU , RAM.

While Debugging, MCU was initialized, RAM was not functioning.
Hand solder touch-up was done to RAM on PCB, Unit was switched ON and was functioning accordingly.
Unit was put in chamber again and Environment Tests were followed up.
Again While performing Step 2 on Day 3, Unit Failed to turn ON.
While Debugging, MCU was initialized, RAM was not functioning.
Hand solder touch-up was done to RAM on PCB, Unit was switched ON and was not functioning.
For Checking RAM, on all Data lines, Read and Write was performed (data bits consisted of alternate 1's and 0's).
Using CRO, On all data pins of both RAM IC's , wave-forms were Appearing as in Fig 1, Except on one of the data pin of 2nd RAM IC, wave-forms appeared as in Fig 2(Controller side waveform was proper, RAM pin waveform was different).
I replaced the 2nd RAM IC, on which wave-forms were appearing as in Fig 2. Unit Started functioning properly.
What would be the cause for such issues ? Should i Consider a different part number for RAM ?
Will there be any issue with PCB design ?

It could a faulty components but the environmental conditions are not particularly harsh.
Can you show a schematic, obviously something is between the MCU and RAM if the signals are different between them.
Why was the solder touch up needed?

My first suspicion in a case like this is the handing during touch-up has caused the damage.

Brian.

Schematic is split into pages,but Signals are coming directly from MCU to the RAM. There is nothing in between.

are the ICs all speced to work properly at at least the conditions of your test?

was there any condensation at 40 deg C and 95% RH?
does the humidity come up first, or does the temperature come up first?

if the humidity comes up first, there may have been hidden condensation that did not
bake out in the high humidity

if there was any condensation, all kinds of things could happen
is your board conformal coated ("painted" with a clear ?? to keep moisture and dirt off the circuits) or not?

Since its inside the chamber, we will know the condition of unit only by the end of the day whether the unit has passed or failed the test, and PCB is inside the unit, and we cannot come to know whether there was any condensation at 40 deg C and 95% RH. After the test if the unit fails to switch ON and perform the required operation, we declare the unit as failed . There was No conformal coating on both the IC's(MCU and RAM).

Are the two chips connected to the same power? Did you measure Vdd on both chips when doing the measurement on Fig2?

From your curve in Fig2 it looks like the RAM has a low Vdd voltage or a too high load on the signal.
If the signal measured is two ends of the 'same' line, you could have problems with the PCB. Like iffy connection in a Via plating, bad soldering or workmanship.
Are the board cleaned after soldering, so there is no remaining flux? Flux remains may be hygroscopic, and causing a temporary decreased isolation between signals.

You could try to heat the board with a blower, to see if the temperature is the problem. If this is ok, then you need to see what the high humidity affects on the board.

Unit failed to turn on after high heat and humidity test

there's more to turning on than the MCU
can you attach wires to the supplies and ground and bring them outside the chamber?
when you turn the unit on after test, check the supply voltages.

is the RAM IC case sealed against humidity at high temperature?

try a chip with a different date code or from a different manufacturer
its possible you got a batch that's "bad" in your particular case