Criterion of PCBA bending degree

Hello all! Recently we got a problem about the Funtional test. Some boards passed tests by using the auto-fixture(vacuum), but failed by using manual test fixture(need to plug CPU, DIMM etc.). The error symptom is no picture. We doubt it's the problem of board bending. Anybody can tell me the criterion of board bending? Or tell me how to find out in some articles or journals. Thanks in advance.

reply »

| Hello all! | Recently we got a problem about the Funtional test. Some boards passed tests by using the auto-fixture(vacuum), but failed by using manual test fixture(need to plug CPU, DIMM etc.). The error symptom is no picture. We doubt it's the problem of board bending. Anybody can tell me the criterion of board bending? Or tell me how to find out in some articles or journals. Thanks in advance. | |

reply »

| | Hello all! | | Recently we got a problem about the Funtional test. Some boards passed tests by using the auto-fixture(vacuum), but failed by using manual test fixture(need to plug CPU, DIMM etc.). The error symptom is no picture. We doubt it's the problem of board bending. Anybody can tell me the criterion of board bending? Or tell me how to find out in some articles or journals. Thanks in advance. | |

Deon Response:

Hello Gyver,

Try IPC-TM-650. I beleive bow and twist criteria is not to exceed 1.5% per linear inch on all thru-hole assemblies and .75% per linear inch on SMT assemblies. Exceding this criteria will jeopordize reliability due to stress at time of mounting final assembly. I would also recommend checking your bare boards at receiving inspection. I have seen and rejected bare boards coming in from the fab house exceeding this criteria. Bow and twist can also affect stretch which can lead to paste print registration and other process problems. Point being, it is an important part of incoming inspection. Hope this helps.

Regards,

Deon Nungaray SMT Mfg. Engineer GMI USA CA

reply »

| Hello all! | Recently we got a problem about the Funtional test. Some boards passed tests by using the auto-fixture(vacuum), but failed by using manual test fixture(need to plug CPU, DIMM etc.). The error symptom is no picture. We doubt it's the problem of board bending. Anybody can tell me the criterion of board bending? Or tell me how to find out in some articles or journals. Thanks in advance. | | Deon has quoted the requirements correctly. However I am concerned about flexural considerations. Bow and twist requirements as received are one thing important. Flexural capabilities are another.

Rigid printed circuitry is just that. There is no forgiving "bending" rigid boards. If they are out of spec, they most often "break" when attempting straightening. Bow and twist is laminated into the board as wrong materials or press cycles.

Also, If you are not getting continuity, consider hole wall failure of several different types. Upon test pin/fixture compression, often opens "heal" themselves as failed hole walls (usually cracked) are reunited and show temporary continuity.

Earl Moon

reply »

| | | Hello all! | | | Recently we got a problem about the Funtional test. Some boards passed tests by using the auto-fixture(vacuum), but failed by using manual test fixture(need to plug CPU, DIMM etc.). The error symptom is no picture. We doubt it's the problem of board bending. Anybody can tell me the criterion of board bending? Or tell me how to find out in some articles or journals. Thanks in advance. | | | | | Deon Response: | | Hello Gyver, | | Try IPC-TM-650. I beleive bow and twist criteria is not to exceed 1.5% per linear inch on all thru-hole assemblies and .75% per linear inch on SMT assemblies. Exceding this criteria will jeopordize reliability due to stress at time of mounting final assembly. I would also recommend checking your bare boards at receiving inspection. I have seen and rejected bare boards coming in from the fab house exceeding this criteria. | Bow and twist can also affect stretch which can lead to paste print registration and other process problems. Point being, it is an important part of incoming inspection. Hope this helps. | | Regards, | | Deon Nungaray | SMT Mfg. Engineer | GMI USA CA | Thanks Earl and Deon! Your messages are very useful for me. Well, I am sorry to ask another question. For BGA components, is IPC-TM-650 also a standard? And, could you tell me how to measure the bow and twist degrees on thru-hole assemblies? Thanks in advance!

Gyver

reply »

| | | | Hello all! | | | | Recently we got a problem about the Funtional test. Some boards passed tests by using the auto-fixture(vacuum), but failed by using manual test fixture(need to plug CPU, DIMM etc.). The error symptom is no picture. We doubt it's the problem of board bending. Anybody can tell me the criterion of board bending? Or tell me how to find out in some articles or journals. Thanks in advance. | | | | | | | | Deon Response: | | | | Hello Gyver, | | | | Try IPC-TM-650. I beleive bow and twist criteria is not to exceed 1.5% per linear inch on all thru-hole assemblies and .75% per linear inch on SMT assemblies. Exceding this criteria will jeopordize reliability due to stress at time of mounting final assembly. I would also recommend checking your bare boards at receiving inspection. I have seen and rejected bare boards coming in from the fab house exceeding this criteria. | | Bow and twist can also affect stretch which can lead to paste print registration and other process problems. Point being, it is an important part of incoming inspection. Hope this helps. | | | | Regards, | | | | Deon Nungaray | | SMT Mfg. Engineer | | GMI USA CA | | | Thanks Earl and Deon! Your messages are very useful for me. Well, I am sorry to ask another question. For BGA components, is IPC-TM-650 also a standard? And, could you tell me how to measure the bow and twist degrees on thru-hole assemblies? | Thanks in advance! | | Gyver | | Dear Deon, Could you tell me where to find out the tolerance mentioned in your response? I've checked the MIL-P-55110D and it says"The maximum allowable bow and twist shall be 1.5%". Thanks in advance!

Gyver

reply »

| | | | | Hello all! | | | | | Recently we got a problem about the Funtional test. Some boards passed tests by using the auto-fixture(vacuum), but failed by using manual test fixture(need to plug CPU, DIMM etc.). The error symptom is no picture. We doubt it's the problem of board bending. Anybody can tell me the criterion of board bending? Or tell me how to find out in some articles or journals. Thanks in advance. | | | | | | | | | | | Deon Response: | | | | | | Hello Gyver, | | | | | | Try IPC-TM-650. I beleive bow and twist criteria is not to exceed 1.5% per linear inch on all thru-hole assemblies and .75% per linear inch on SMT assemblies. Exceding this criteria will jeopordize reliability due to stress at time of mounting final assembly. I would also recommend checking your bare boards at receiving inspection. I have seen and rejected bare boards coming in from the fab house exceeding this criteria. | | | Bow and twist can also affect stretch which can lead to paste print registration and other process problems. Point being, it is an important part of incoming inspection. Hope this helps. | | | | | | Regards, | | | | | | Deon Nungaray | | | SMT Mfg. Engineer | | | GMI USA CA | | | | | Thanks Earl and Deon! Your messages are very useful for me. Well, I am sorry to ask another question. For BGA components, is IPC-TM-650 also a standard? And, could you tell me how to measure the bow and twist degrees on thru-hole assemblies? | | Thanks in advance! | | | | Gyver | | | | | Dear Deon, | Could you tell me where to find out the tolerance mentioned in your response? I've checked the MIL-P-55110D and it says"The maximum allowable bow and twist shall be 1.5%". Thanks in advance! | | Gyver | Dear Earl, Sorry to ask you one question. You mentioned "Hole wall failure" in your response. Well, I don't understand this term. Could you decribe in detail or tell me where to get related description? Thanks in advance!

Gyver

reply »

| | | | | Hello all! | | | | | Recently we got a problem about the Funtional test. Some boards passed tests by using the auto-fixture(vacuum), but failed by using manual test fixture(need to plug CPU, DIMM etc.). The error symptom is no picture. We doubt it's the problem of board bending. Anybody can tell me the criterion of board bending? Or tell me how to find out in some articles or journals. Thanks in advance. | | | | | | | | IPC-A-610-B lists 0.75% for SMT and 1.5% for PTH in accordance w/ IPC-TM-650. Fomula= D x .075 for SMT and D x .15 for PTH. (D= greatest diagonal measure). E.G. if a board measures 13.5 inches (diagonally) the max allowable warp would be 1.013in. (13.5 x .075) for SMT, and 2.025in. 913.5 x .15) for PTH. To verify you could use standard pin guages and a flat surface. The height of the bottom of the pcb (anywhere on the board) from the flat surface should not exceed the max allowable warp. | | | Deon Response: | | | | | | Hello Gyver, | | | | | | Try IPC-TM-650. I beleive bow and twist criteria is not to exceed 1.5% per linear inch on all thru-hole assemblies and .75% per linear inch on SMT assemblies. Exceding this criteria will jeopordize reliability due to stress at time of mounting final assembly. I would also recommend checking your bare boards at receiving inspection. I have seen and rejected bare boards coming in from the fab house exceeding this criteria. | | | Bow and twist can also affect stretch which can lead to paste print registration and other process problems. Point being, it is an important part of incoming inspection. Hope this helps. | | | | | | Regards, | | | | | | Deon Nungaray | | | SMT Mfg. Engineer | | | GMI USA CA | | | | | Thanks Earl and Deon! Your messages are very useful for me. Well, I am sorry to ask another question. For BGA components, is IPC-TM-650 also a standard? And, could you tell me how to measure the bow and twist degrees on thru-hole assemblies? | | Thanks in advance! | | | | Gyver | | | | | Dear Deon, | Could you tell me where to find out the tolerance mentioned in your response? I've checked the MIL-P-55110D and it says"The maximum allowable bow and twist shall be 1.5%". Thanks in advance! | | Gyver |

reply »

| | | | | Hello all! | | | | | Recently we got a problem about the Funtional test. Some boards passed tests by using the auto-fixture(vacuum), but failed by using manual test fixture(need to plug CPU, DIMM etc.). The error symptom is no picture. We doubt it's the problem of board bending. Anybody can tell me the criterion of board bending? Or tell me how to find out in some articles or journals. Thanks in advance. | | | | | | | |

| | | Deon Response: | | | | | | Hello Gyver, | | | | | | Try IPC-TM-650. I beleive bow and twist criteria is not to exceed 1.5% per linear inch on all thru-hole assemblies and .75% per linear inch on SMT assemblies. Exceding this criteria will jeopordize reliability due to stress at time of mounting final assembly. I would also recommend checking your bare boards at receiving inspection. I have seen and rejected bare boards coming in from the fab house exceeding this criteria. | | | Bow and twist can also affect stretch which can lead to paste print registration and other process problems. Point being, it is an important part of incoming inspection. Hope this helps. | | | | | | Regards, | | | | | | Deon Nungaray | | | SMT Mfg. Engineer | | | GMI USA CA | | | | | Thanks Earl and Deon! Your messages are very useful for me. Well, I am sorry to ask another question. For BGA components, is IPC-TM-650 also a standard? And, could you tell me how to measure the bow and twist degrees on thru-hole assemblies? | | Thanks in advance! | | | | Gyver | | | | | Dear Deon, | Could you tell me where to find out the tolerance mentioned in your response? I've checked the MIL-P-55110D and it says"The maximum allowable bow and twist shall be 1.5%". Thanks in advance! | | Gyver | Deon Response :

Hi Gyver,

Sorry it took me a while to repsond, we don't work on Fridays here. The way I have measured the bow and twist of any assembly is by placing it on a granite table and making sure board is flat or perfectly horizontal on at least two sides. I then take a measurement on what the difference is in height if any from the opposite side of the PCB. If this exceeds the mentioned criteria then you must rejct the assmebly for excess bow and twist. Hope this helps.

Regards,

Deon Nungaray SMT Mfg. Engineer GMI USA CA

reply »