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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Solder surface tension

Steve Restinetti (aka Bix)

#9068

Solder surface tension | 8 October, 1999

Can anyone tell me the formula/method to determine the amount of weight that solder's surface tension can support. (ie: Maximum weight of components soldered onto a bottom side of a board during a top side reflow on a Paste/Paste process).

Thanks, Steve stever@uttc.ca bix@netcom.ca bix@smtnet.com

reply »

Brian

#9069

Re: Solder surface tension | 9 October, 1999

| Can anyone tell me the formula/method to determine | the amount of weight that solder's surface tension can support. | (ie: Maximum weight of components soldered onto a bottom side | of a board during a top side reflow on a Paste/Paste process). | | Thanks, | Steve | stever@uttc.ca | bix@netcom.ca | bix@smtnet.com | Steve

Now, there's a good question, but I don't think it is really possible to give a cut-and-dried answer. Certainly, it inspired me to look through my library and I found nothing. My feeling is that it would depend greatly on a) the geometry of the joint b) the wettable areas and numbers of pad and lead (metallisation) c) the wettability of the pad and lead (metallisation) d) the peak temperature e) the smoothness of the conveyor motion f) the solder alloy, including intermetallics g) probably etc., etc., etc.

The one thing I did find was that a Mr. A. Latin published the surface tension of solder at 280�C in J. Inst. Metals in 1946. Eutectic 63/37 is 490 dyne/cm (goes up slightly at higher tin content, down slightly at higher lead content). This is a helluva figure, nearly 7 times higher than pure water at 20�C. Any good degree-level physics textbook will give you the formulae regarding meniscus formation, but I'm not sure that it is relevant here, as the intermetallic formation will probably upset the calculations.

Sorry can't help with a real answer

Brian

reply »

#9070

Re: Solder surface tension | 9 October, 1999

| | Can anyone tell me the formula/method to determine | | the amount of weight that solder's surface tension can support. | | (ie: Maximum weight of components soldered onto a bottom side | | of a board during a top side reflow on a Paste/Paste process). | | | | Thanks, | | Steve | | stever@uttc.ca | | bix@netcom.ca | | bix@smtnet.com | | | Steve | | Now, there's a good question, but I don't think it is really possible to give a cut-and-dried answer. Certainly, it inspired me to look through my library and I found nothing. My feeling is that it would depend greatly on | a) the geometry of the joint | b) the wettable areas and numbers of pad and lead (metallisation) | c) the wettability of the pad and lead (metallisation) | d) the peak temperature | e) the smoothness of the conveyor motion | f) the solder alloy, including intermetallics | g) probably etc., etc., etc. | | The one thing I did find was that a Mr. A. Latin published the surface tension of solder at 280�C in J. Inst. Metals in 1946. Eutectic 63/37 is 490 dyne/cm (goes up slightly at higher tin content, down slightly at higher lead content). This is a helluva figure, nearly 7 times higher than pure water at 20�C. Any good degree-level physics textbook will give you the formulae regarding meniscus formation, but I'm not sure that it is relevant here, as the intermetallic formation will probably upset the calculations. | | Sorry can't help with a real answer | | Brian | Both of you guy should look in the SMTnet archives.

reply »

Brian

#9071

Re: Solder surface tension | 9 October, 1999

| | | Can anyone tell me the formula/method to determine | | | the amount of weight that solder's surface tension can support. | | | (ie: Maximum weight of components soldered onto a bottom side | | | of a board during a top side reflow on a Paste/Paste process). | | | | | | Thanks, | | | Steve | | | stever@uttc.ca | | | bix@netcom.ca | | | bix@smtnet.com | | | | | Steve | | | | Now, there's a good question, but I don't think it is really possible to give a cut-and-dried answer. Certainly, it inspired me to look through my library and I found nothing. My feeling is that it would depend greatly on | | a) the geometry of the joint | | b) the wettable areas and numbers of pad and lead (metallisation) | | c) the wettability of the pad and lead (metallisation) | | d) the peak temperature | | e) the smoothness of the conveyor motion | | f) the solder alloy, including intermetallics | | g) probably etc., etc., etc. | | | | The one thing I did find was that a Mr. A. Latin published the surface tension of solder at 280�C in J. Inst. Metals in 1946. Eutectic 63/37 is 490 dyne/cm (goes up slightly at higher tin content, down slightly at higher lead content). This is a helluva figure, nearly 7 times higher than pure water at 20�C. Any good degree-level physics textbook will give you the formulae regarding meniscus formation, but I'm not sure that it is relevant here, as the intermetallic formation will probably upset the calculations. | | | | Sorry can't help with a real answer | | | | Brian | | | Both of you guy should look in the SMTnet archives. | Dave

Sorry!, can't find anything in the archives that really answers Steve's question. Most of what has been said in several threads that I found is summarised by words to the effect of the surface tension is enough, but there is nothing I was able to find that showed how to calculate the forces.

There is another point inherent in Steve's question that I don't believe has ever been addressed (subject to proof to the contrary). Imagine we have two physically identical QFPs, one of which weighs 3 grams and the other 15 grams because it has a built-in cooling finning, soldered on side 1. During side 2 soldering, we assume the solder on side 1 exceeds liquidus. The light QFP will probably continue to sit pretty tight to the board but the other one will be pulled further down by gravity (assuming it stays attached), so that each lead will sit with much more solder between it and the PCB pad. What will the effect of this be on the reliability? I would venture to suggest that there would be more cracking after thermal cycling. It would therefore be useful if some physico-metallurgico-mathematical genius could help us calculate what is likely to happen when we reflow PCBs with heavy components on side 1.

Perhaps I'm reading more into Steve's question than is really there ???????

Brian

reply »

#9072

Re: Solder surface tension | 9 October, 1999

| | | | Can anyone tell me the formula/method to determine | | | | the amount of weight that solder's surface tension can support. | | | | (ie: Maximum weight of components soldered onto a bottom side | | | | of a board during a top side reflow on a Paste/Paste process). | | | | | | | | Thanks, | | | | Steve | | | | stever@uttc.ca | | | | bix@netcom.ca | | | | bix@smtnet.com | | | | | | | Steve | | | | | | Now, there's a good question, but I don't think it is really possible to give a cut-and-dried answer. Certainly, it inspired me to look through my library and I found nothing. My feeling is that it would depend greatly on | | | a) the geometry of the joint | | | b) the wettable areas and numbers of pad and lead (metallisation) | | | c) the wettability of the pad and lead (metallisation) | | | d) the peak temperature | | | e) the smoothness of the conveyor motion | | | f) the solder alloy, including intermetallics | | | g) probably etc., etc., etc. | | | | | | The one thing I did find was that a Mr. A. Latin published the surface tension of solder at 280�C in J. Inst. Metals in 1946. Eutectic 63/37 is 490 dyne/cm (goes up slightly at higher tin content, down slightly at higher lead content). This is a helluva figure, nearly 7 times higher than pure water at 20�C. Any good degree-level physics textbook will give you the formulae regarding meniscus formation, but I'm not sure that it is relevant here, as the intermetallic formation will probably upset the calculations. | | | | | | Sorry can't help with a real answer | | | | | | Brian | | | | | Both of you guy should look in the SMTnet archives. | | | Dave | | Sorry!, can't find anything in the archives that really answers Steve's question. Most of what has been said in several threads that I found is summarised by words to the effect of the surface tension is enough, but there is nothing I was able to find that showed how to calculate the forces. | | There is another point inherent in Steve's question that I don't believe has ever been addressed (subject to proof to the contrary). Imagine we have two physically identical QFPs, one of which weighs 3 grams and the other 15 grams because it has a built-in cooling finning, soldered on side 1. During side 2 soldering, we assume the solder on side 1 exceeds liquidus. The light QFP will probably continue to sit pretty tight to the board but the other one will be pulled further down by gravity (assuming it stays attached), so that each lead will sit with much more solder between it and the PCB pad. What will the effect of this be on the reliability? I would venture to suggest that there would be more cracking after thermal cycling. It would therefore be useful if some physico-metallurgico-mathematical genius could help us calculate what is likely to happen when we reflow PCBs with heavy components on side 1. | | Perhaps I'm reading more into Steve's question than is really there ??????? | | Brian | May be I'm reading too little into Steve's question. I think he wants to figure-out which components will fall from the board during second sided reflow. Try:

DOUBLE REFLOW WITH HEAVY WEIGHT PLCC PACKAGE - Stoney Tsai 21:27:35 07/23/1998

Ta

Dave F

reply »

bix

#9073

Re: Solder surface tension | 9 October, 1999

| | | | | Can anyone tell me the formula/method to determine | | | | | the amount of weight that solder's surface tension can support. | | | | | (ie: Maximum weight of components soldered onto a bottom side | | | | | of a board during a top side reflow on a Paste/Paste process). | | | | | | | | | | Thanks, | | | | | Steve | | | | | stever@uttc.ca | | | | | bix@netcom.ca | | | | | bix@smtnet.com | | | | | | | | | Steve | | | | | | | | Now, there's a good question, but I don't think it is really possible to give a cut-and-dried answer. Certainly, it inspired me to look through my library and I found nothing. My feeling is that it would depend greatly on | | | | a) the geometry of the joint | | | | b) the wettable areas and numbers of pad and lead (metallisation) | | | | c) the wettability of the pad and lead (metallisation) | | | | d) the peak temperature | | | | e) the smoothness of the conveyor motion | | | | f) the solder alloy, including intermetallics | | | | g) probably etc., etc., etc. | | | | | | | | The one thing I did find was that a Mr. A. Latin published the surface tension of solder at 280�C in J. Inst. Metals in 1946. Eutectic 63/37 is 490 dyne/cm (goes up slightly at higher tin content, down slightly at higher lead content). This is a helluva figure, nearly 7 times higher than pure water at 20�C. Any good degree-level physics textbook will give you the formulae regarding meniscus formation, but I'm not sure that it is relevant here, as the intermetallic formation will probably upset the calculations. | | | | | | | | Sorry can't help with a real answer | | | | | | | | Brian | | | | | | | Both of you guy should look in the SMTnet archives. | | | | | Dave | | | | Sorry!, can't find anything in the archives that really answers Steve's question. Most of what has been said in several threads that I found is summarised by words to the effect of the surface tension is enough, but there is nothing I was able to find that showed how to calculate the forces. | | | | There is another point inherent in Steve's question that I don't believe has ever been addressed (subject to proof to the contrary). Imagine we have two physically identical QFPs, one of which weighs 3 grams and the other 15 grams because it has a built-in cooling finning, soldered on side 1. During side 2 soldering, we assume the solder on side 1 exceeds liquidus. The light QFP will probably continue to sit pretty tight to the board but the other one will be pulled further down by gravity (assuming it stays attached), so that each lead will sit with much more solder between it and the PCB pad. What will the effect of this be on the reliability? I would venture to suggest that there would be more cracking after thermal cycling. It would therefore be useful if some physico-metallurgico-mathematical genius could help us calculate what is likely to happen when we reflow PCBs with heavy components on side 1. | | | | Perhaps I'm reading more into Steve's question than is really there ??????? | | | | Brian | | | May be I'm reading too little into Steve's question. I think he wants to figure-out which components will fall from the board during second sided reflow. Try: | | DOUBLE REFLOW WITH HEAVY WEIGHT PLCC PACKAGE - Stoney Tsai 21:27:35 07/23/1998 | | Ta | | Dave F | Well, I really didn't want too technical or too simple of an answer, basically a kinda simplistic way to figure out which components will most likely fall off and which ones may not. But Dave, I will read the THREAD you last mentioned. Thanks for all of your help including Brian.

Steve...

reply »

#9074

Re: Solder surface tension | 11 October, 1999

| | | | | | Can anyone tell me the formula/method to determine | | | | | | the amount of weight that solder's surface tension can support. | | | | | | (ie: Maximum weight of components soldered onto a bottom side | | | | | | of a board during a top side reflow on a Paste/Paste process). | | | | | | | | | | | | Thanks, | | | | | | Steve | | | | | | stever@uttc.ca | | | | | | bix@netcom.ca | | | | | | bix@smtnet.com | | | | | | | | | | | Steve | | | | | | | | | | Now, there's a good question, but I don't think it is really possible to give a cut-and-dried answer. Certainly, it inspired me to look through my library and I found nothing. My feeling is that it would depend greatly on | | | | | a) the geometry of the joint | | | | | b) the wettable areas and numbers of pad and lead (metallisation) | | | | | c) the wettability of the pad and lead (metallisation) | | | | | d) the peak temperature | | | | | e) the smoothness of the conveyor motion | | | | | f) the solder alloy, including intermetallics | | | | | g) probably etc., etc., etc. | | | | | | | | | | The one thing I did find was that a Mr. A. Latin published the surface tension of solder at 280�C in J. Inst. Metals in 1946. Eutectic 63/37 is 490 dyne/cm (goes up slightly at higher tin content, down slightly at higher lead content). This is a helluva figure, nearly 7 times higher than pure water at 20�C. Any good degree-level physics textbook will give you the formulae regarding meniscus formation, but I'm not sure that it is relevant here, as the intermetallic formation will probably upset the calculations. | | | | | | | | | | Sorry can't help with a real answer | | | | | | | | | | Brian | | | | | | | | | Both of you guy should look in the SMTnet archives. | | | | | | | Dave | | | | | | Sorry!, can't find anything in the archives that really answers Steve's question. Most of what has been said in several threads that I found is summarised by words to the effect of the surface tension is enough, but there is nothing I was able to find that showed how to calculate the forces. | | | | | | There is another point inherent in Steve's question that I don't believe has ever been addressed (subject to proof to the contrary). Imagine we have two physically identical QFPs, one of which weighs 3 grams and the other 15 grams because it has a built-in cooling finning, soldered on side 1. During side 2 soldering, we assume the solder on side 1 exceeds liquidus. The light QFP will probably continue to sit pretty tight to the board but the other one will be pulled further down by gravity (assuming it stays attached), so that each lead will sit with much more solder between it and the PCB pad. What will the effect of this be on the reliability? I would venture to suggest that there would be more cracking after thermal cycling. It would therefore be useful if some physico-metallurgico-mathematical genius could help us calculate what is likely to happen when we reflow PCBs with heavy components on side 1. | | | | | | Perhaps I'm reading more into Steve's question than is really there ??????? | | | | | | Brian | | | | | May be I'm reading too little into Steve's question. I think he wants to figure-out which components will fall from the board during second sided reflow. Try: | | | | DOUBLE REFLOW WITH HEAVY WEIGHT PLCC PACKAGE - Stoney Tsai 21:27:35 07/23/1998 | | | | Ta | | | | Dave F | | | Well, I really didn't want too technical or too simple of an answer, basically a kinda simplistic way to figure out which | components will most likely fall off and which ones may not. | But Dave, I will read the THREAD you last mentioned. Thanks | for all of your help including Brian. | | Steve... | | With due credit to Phil Zarrow...

Weight of component in grams ________________________

Total pad mating area in square inches

Grams per square inch must be equal to or less than 30 for successful secondary side mounting John Thorup

|

reply »

Brian

#9075

Re: Solder surface tension | 12 October, 1999

| | | | | | | Can anyone tell me the formula/method to determine | | | | | | | the amount of weight that solder's surface tension can support. | | | | | | | (ie: Maximum weight of components soldered onto a bottom side | | | | | | | of a board during a top side reflow on a Paste/Paste process). | | | | | | | | | | | | | | Thanks, | | | | | | | Steve | | | | | | | stever@uttc.ca | | | | | | | bix@netcom.ca | | | | | | | bix@smtnet.com | | | | | | | | | | | | | Steve | | | | | | | | | | | | Now, there's a good question, but I don't think it is really possible to give a cut-and-dried answer. Certainly, it inspired me to look through my library and I found nothing. My feeling is that it would depend greatly on | | | | | | a) the geometry of the joint | | | | | | b) the wettable areas and numbers of pad and lead (metallisation) | | | | | | c) the wettability of the pad and lead (metallisation) | | | | | | d) the peak temperature | | | | | | e) the smoothness of the conveyor motion | | | | | | f) the solder alloy, including intermetallics | | | | | | g) probably etc., etc., etc. | | | | | | | | | | | | The one thing I did find was that a Mr. A. Latin published the surface tension of solder at 280�C in J. Inst. Metals in 1946. Eutectic 63/37 is 490 dyne/cm (goes up slightly at higher tin content, down slightly at higher lead content). This is a helluva figure, nearly 7 times higher than pure water at 20�C. Any good degree-level physics textbook will give you the formulae regarding meniscus formation, but I'm not sure that it is relevant here, as the intermetallic formation will probably upset the calculations. | | | | | | | | | | | | Sorry can't help with a real answer | | | | | | | | | | | | Brian | | | | | | | | | | | Both of you guy should look in the SMTnet archives. | | | | | | | | | Dave | | | | | | | | Sorry!, can't find anything in the archives that really answers Steve's question. Most of what has been said in several threads that I found is summarised by words to the effect of the surface tension is enough, but there is nothing I was able to find that showed how to calculate the forces. | | | | | | | | There is another point inherent in Steve's question that I don't believe has ever been addressed (subject to proof to the contrary). Imagine we have two physically identical QFPs, one of which weighs 3 grams and the other 15 grams because it has a built-in cooling finning, soldered on side 1. During side 2 soldering, we assume the solder on side 1 exceeds liquidus. The light QFP will probably continue to sit pretty tight to the board but the other one will be pulled further down by gravity (assuming it stays attached), so that each lead will sit with much more solder between it and the PCB pad. What will the effect of this be on the reliability? I would venture to suggest that there would be more cracking after thermal cycling. It would therefore be useful if some physico-metallurgico-mathematical genius could help us calculate what is likely to happen when we reflow PCBs with heavy components on side 1. | | | | | | | | Perhaps I'm reading more into Steve's question than is really there ??????? | | | | | | | | Brian | | | | | | | May be I'm reading too little into Steve's question. I think he wants to figure-out which components will fall from the board during second sided reflow. Try: | | | | | | DOUBLE REFLOW WITH HEAVY WEIGHT PLCC PACKAGE - Stoney Tsai 21:27:35 07/23/1998 | | | | | | Ta | | | | | | Dave F | | | | | Well, I really didn't want too technical or too simple of an answer, basically a kinda simplistic way to figure out which | | components will most likely fall off and which ones may not. | | But Dave, I will read the THREAD you last mentioned. Thanks | | for all of your help including Brian. | | | | Steve... | | | | With due credit to Phil Zarrow... | | Weight of component in grams | ________________________ | | Total pad mating area in square inches | | Grams per square inch must be equal to or less than 30 for successful secondary side mounting | John Thorup | | | | | John

This rule of thumb could rule out many ICs. For example, A QFP64 with an SOT 314-1 case has a lead of 0,20 mm wide, with a typical flattened part of 1,20 mm long (I take the lead, rather than the pad, as being slightly smaller). This gives a total surface area of 15,36 mm2 or 0,0238 in2, giving a weight limit of 0,714 g, for a typical component weight of 1,3 g. A similar calculation would give a SOIC 16 (SOT109-1/2) a max weight of 0,13 g, where its real weight is typically about 0,3 g upwards. Yet I know of companies side 1-soldering such components. I think your figure may be a tad conservative?

Brian

reply »

#9076

Re: Solder surface tension | 13 October, 1999

Brian You're quite right. I frequently break this formula. The source recommends it as an almost assured success rather than an absolute. Given all the variables involved Steve should use it as a reality check and then just see what works. John Thorup

reply »

Reflow Ovens thermal process improvement