Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

Baking time for PCBA rework

Dason C


Baking time for PCBA rework | 31 January, 2002

What is the baking time/temperature standard for the PCBA before BGA rework?

reply »


Baking time for PCBA rework | 3 February, 2002

There is no standard for baking an assembled board prior to BGA rework, nor should there be. The amount of time required to bake a board varies with: * Board material. * Board construction and layers. * Board size. * Components on the board.

If someone added a PCA rework standard for BGA reowk to J-STD-001, it would probably go like: Baking of the assembly prior to BGA rework and repair shall be documented and not affect the performance specification.

reply »


Baking time for PCBA rework | 4 February, 2002

Dave highlighted all the elements that must be considered to determine the optimal temperature and duration for a pre-rework PCBA bake process.

As far as the component is concerned, it really depends if you care about not damaging the package for re-balling or failure analysis reasons. Since PBGAs are moisture-sensitive you must make sure that you remove enough moisture before you perform a localized reflow. Otherwise this will induce internal cracks and delaminations.

You will find the minimal requirements in the upcoming revision to J-STD-033. It includes a bake table at 90C for populated boards that cannot withstand the 125C bake temperature normally used for individual components.

By the way the bake duration at 90C for PBGAs will range from 11 hours to 10 days, based on package thickness, MS level and prior exposure (i.e. "saturated" or "at limit of floor life").

I can e-mail you a copy of the proposed bake table if you are interested.

Francois Monette Cogiscan Inc.

reply »


Baking time for PCBA rework | 6 February, 2002

I agree with Francois. We have just announced a 55 deg. C LTVP drying process that can greatly speed moisture removal. It is also effective for drying components that remain in tape and reel, conductive plastic shipping tubes or trays. Process times from 16 to 18 hours are common.

The paper that was presented to the JC-14 Moisture Sensitivity JEDEC subcommittee is available on request. It gives you all of the details on this subject

reply »

Dason C


Baking time for PCBA rework | 6 February, 2002

Francois/Sleech, thank you very much for your information, by the way where I can get the copy of the paper, what is LTVP stand for and what is the drying requirement expected ie less than 5%?

I am also trying to run my evaluation to dry the parts at 55C/<5%RH and looking for the drying time.

Can you also advise what is the allowable moisture remain on the assembly and I assume that we can leave 25% of the allowable moisture gain from the soaking parts(I soaked the parts at 30C/60%RH for 18 hours and check the weight gain form the dry parts)?

Thank & My Best Rgds. Dason

reply »


Baking time for PCBA rework | 7 February, 2002


Please be aware that weight gain is only a very crude estimate of the component moisture content. This is only accurate when the moisture is uniformly distributed inside the component (which is actually never). This measurement is usually not conservative enough and this method was abandonned a few years ago by the IPC/JEDEC committee.

Since the diffusion process inside a package is very slow, there is always a moisture gradient. Moisture is either moving in or out of the package. When you bake a component the last moisture concentration that will be removed is the one in the middle of the package near the die (critical interface), which is the exact location where cracks and delaminations will initiate. A package could have a high moisture content by weight that will be harmless if it is not too far inside the part. Conversely, even a very small moisture content by weight can create defects when it is near the center of the package.

The most accurate method to account for drying conditions is to actually calculate the time required to remove the minimum amount of moisture from the critical interface, using known diffusion models (an excellent paper on this subject was written by Rick Shook from Lucent. I can send you a copy if you want).

This method was used to compute the current bake tables in J-STD-033. This also explains why the bake cycle was increased significantly, from 24 to 48 hours at 125C and from 8 to 68 days at 40C. This is the difference between the simplistic weight gain analysis and the more accurate moisture diffusion model. For a comprehensive list of technical papers on this subject I invite you to check the MSD knowldedge base at, click on Moisture Sensitive Devices.

Francois Monette Cogiscan Inc.

reply »


Baking time for PCBA rework | 23 April, 2002

Here's a related question: How hot can a PBGA get before popcorn delamination becomes a concern? 100 C? 150C?

I realize that the answer will depend upon the conditioning of the part, so let's assume that the part is saturated with moisture.

A practical answer to this question might allow us to eliminate the bake times prior to reworking boards with PBGAs. Currently, we bake them for 24 hours at 125 C in order to protect the parts (PBGAs and PQFPs) adjacent to the rework site(s). But if we find that the adjacent parts are only reaching, say, 130 C on a specific board, then can we skip the baking for that board?

reply »

Matt N-Stott


Baking time for PCBA rework | 23 April, 2002

I would sugest the lowere the baking temperature the better, the main cause of popcorning is the water expansion and this is greatest as the water boils, 100 degrees, better to elongate the time.


reply »


Baking time for PCBA rework | 23 April, 2002

My first posting was not very clear -- sorry. Let me try again.

What I want to know is: can I eliminate the drying (baking) process completely IF I know that the good parts (i.e., those not being reworked) are going to stay under some threshold temperature during rework?

If I'm removing PBGA "A", and PBGA "B" is 2 inches from PBGA "A", and PBGA "B" is saturated with moisture, then will PBGA "B" crack/delaminate during rework if it reaches 190C? What if it reaches only 170C? What if it reaches only 150C? What is the threshold?

Note: I realize that PBGA "A" will probably crack/ delaminate in this example.

reply »


Baking time for PCBA rework | 24 April, 2002

Ooooo, much better.

Tough to say, because of all the variables [ie, different materials, moisture level, temperature, etc].

Plastic encapsulated devices, especially IC, absorb water from the air, which is violently released during soldering. Typically, 1000 ppm of absorbed moisture is considered a maximum content beyond which device failure due to body cracking may result. [from IPC-HDBK-001]

Creating a possible analog �

According to IPC-HDBK-001, for bare boards, the bake-out removes water accumulated during the fabrication process and absorbed during storage. Recommended baking times and temperatures are: Baking Temperature [�C]||Baking Time [Hrs] 120||3.5 to 7 100||8 to 16 80||18 to 48 Longer bake-out times and higher temperatures are not recommend, as this can degrade PWB and component solderability.

So if people bake to remove moisture from a board at 80�C, then it might ne possible to damage a component at that temperature.

Have you considered: * Shielding moisture sensitive parts from the heat? * Impact of the temperature change on the bare board?

reply »


Baking time for PCBA rework | 25 April, 2002


This concern is well covered in J-STD-033, section 8.6 Board Rework. "...If the component temperature exceeds 200C, the board may require a bakeout prior to rework..."

In other words, you only need to worry about moisture related defects when the maximum component body temperature exceeds 200C, as measured on top of the package. This is the practical threshold that is established by the industry standard.

Francois Monette Cogiscan Inc.

reply »


Baking time for PCBA rework | 25 April, 2002


By the way, the default bake duration should be 48 hours at 125C instead of 24 hours. The shorter duration was based on a standard that is now obsolete. The current revision of J-STD-033 includes a variable duration bake table, based on component MS level and body thickness. (Re : table 2). Keep in mind that a new revision is just about to be released which includes significant changes to that table. You can download a copy of the standard at or I can e-mail you a copy if you want.

Francois Monette Cogiscan Inc. Tel : 450-534-2644 Fax : 450-534-0092 E-mail :

reply »