Adhesion loss on passives over wave

Hi folks,

I am looking for inputs on a long-standing problem encountered in our shop: Bottom-side parts (passives) falling off in the Wave...where the only remnants of the part is a flattened glue dot of proper diameter and volume, and sometimes flux residue (from the wave).

I suspect there is some incompatibility between one of our pastes and the glue. Our shop uses two types of paste and one type of glue. After some experimenting, I am seeing a STRONG correlation that suggests one paste type is not suited to the glue.

Short of listing the specific materials used and getting myself in trouble, is there anyone who has encountered the problem I am talking about? What was the cause and how did you resolve it?

If you want more specifics, please call me at 509-921-4311.

Thanks, Marc

reply »

| Hi folks, | | I am looking for inputs on a long-standing problem encountered in our shop: Bottom-side parts (passives) falling off in the Wave...where the only remnants of the part is a flattened glue dot of proper diameter and volume, and sometimes flux residue (from the wave). | | I suspect there is some incompatibility between one of our pastes and the glue. Our shop uses two types of paste and one type of glue. After some experimenting, I am seeing a STRONG correlation that suggests one paste type is not suited to the glue. | | Short of listing the specific materials used and getting myself in trouble, is there anyone who has encountered the problem I am talking about? What was the cause and how did you resolve it? | | If you want more specifics, please call me at 509-921-4311. | | Thanks, | Marc | Have you looked at the possibility that the parts are being knocked off in the handeling proccess prior to flow? If the glue is losing integrity from the use of an specific paste your torque readings should be considerably less with that paste type. This is just my opinion, I could be wrong??? |

reply »

| | Hi folks, | | | | I am looking for inputs on a long-standing problem encountered in our shop: Bottom-side parts (passives) falling off in the Wave...where the only remnants of the part is a flattened glue dot of proper diameter and volume, and sometimes flux residue (from the wave). | | | | I suspect there is some incompatibility between one of our pastes and the glue. Our shop uses two types of paste and one type of glue. After some experimenting, I am seeing a STRONG correlation that suggests one paste type is not suited to the glue. | | | | Short of listing the specific materials used and getting myself in trouble, is there anyone who has encountered the problem I am talking about? What was the cause and how did you resolve it? | | | | If you want more specifics, please call me at 509-921-4311. | | | | Thanks, | | Marc | | | Have you looked at the possibility that the parts are being knocked off in the handeling proccess prior to flow? If the glue is losing integrity from the use of an specific paste your torque readings should be considerably less with that paste type. | This is just my opinion, I could be wrong??? | | | | | Marc,

Are you suggesting there may be some chemical incompatibility between paste/flux chemistries and the glue being used - thus intefering with the glue's ability to bond properly? I've only heard of this possibility, but never experienced it. Justin, Steve, Dave, or Wayne must have the answers to this issue. Where are you guys and all you others - the man has a serious problem and I can't answer it (obviously) though sitting but one aisle and four cubes away.

Regards,

Earl Moon

reply »

| | | Hi folks, | | | | | | I am looking for inputs on a long-standing problem encountered in our shop: Bottom-side parts (passives) falling off in the Wave...where the only remnants of the part is a flattened glue dot of proper diameter and volume, and sometimes flux residue (from the wave). | | | | | | I suspect there is some incompatibility between one of our pastes and the glue. Our shop uses two types of paste and one type of glue. After some experimenting, I am seeing a STRONG correlation that suggests one paste type is not suited to the glue. | | | | | | Short of listing the specific materials used and getting myself in trouble, is there anyone who has encountered the problem I am talking about? What was the cause and how did you resolve it? | | | | | | If you want more specifics, please call me at 509-921-4311. | | | | | | Thanks, | | | Marc | | | | | Have you looked at the possibility that the parts are being knocked off in the handeling proccess prior to flow? If the glue is losing integrity from the use of an specific paste your torque readings should be considerably less with that paste type. | | This is just my opinion, I could be wrong??? | | | | | | | | | | Marc, | | Are you suggesting there may be some chemical incompatibility between paste/flux chemistries and the glue being used - thus intefering with the glue's ability to bond properly? I've only heard of this possibility, but never experienced it. Justin, Steve, Dave, or Wayne must have the answers to this issue. Where are you guys and all you others - the man has a serious problem and I can't answer it (obviously) though sitting but one aisle and four cubes away. | | Regards, | | Earl Moon | Earl, et al:

Indeed I am, and I currently have two different vendors (of the materials in question) evaluating the chemistries, and the possibility that adhesion degredation is due to the paste/flux solvents or other compounds. The solvent, btw, is a very uncommon material, according to the folks I'm working with (not the mfr. of the paste).

As for the question immediately preceding yours, I have another no-clean paste enroute that will allow me to conduct comparison experiments, and further validate (or not) the no-clean paste interaction theory. I doubt very much that handling is at the root here, although arguably all the torsion the boards go through prior to wave may have some cumulative effect that is not understood (at least by me). Getting the other paste will help bound this problem.

Meanwhile...I wait for some lab results!

Marc

reply »

Are you suggesting there may be some chemical incompatibility between paste/flux chemistries and the glue being used - thus intefering with the glue's ability to bond properly? I've only heard of this possibility, but never experienced it. Justin, Steve, Dave, or Wayne must have the answers to this issue. Where are you guys and all you others - the man has a serious problem and I can't answer it (obviously) though sitting but one aisle and four cubes away.

Regards, Earl Moon

Hi ya Earl!

I just answered ya' on the TechNet...missed ya' at the Expo Moonman! It was pretty fun actually! Rock Bottom Brewery was just across the street from the Convention Center...and they brew a heavenly amber! Gots cute waitresses too!

Anyways, I had this happen once...it was all the same part, a capacitor. Turns out the tape had a mold release compound all over it that contaminated the caps where the glue wouldn't bond good (some sorta silicon material, you can get that same kinda stuff from the bags they package boards in sometimes too). Changed the vendors of the caps and it was okie-dokie after that! If it ain't the same part, well....back to the drawing board...

-Steve Gregory-

reply »

Hey Marc,

I am not sure about the kind of process you are following. Is it Gluing+paste printing on the bottom side or just gluing on the bottom side and paste printing on the top side.

The relationship which I could think of between Bottom side paste printing and the glue could be that the paste is less viscous, therefore flows into the glue and affects the properties of the glue. Check the viscosity of the two pastes. Is the one with lower viscosity a problem.

Also. you can check on the dot size of the glue. If you need dot size details as per component type you can reply back to me.

Regards, Vinesh Gandhi

| Hi folks, | | I am looking for inputs on a long-standing problem encountered in our shop: Bottom-side parts (passives) falling off in the Wave...where the only remnants of the part is a flattened glue dot of proper diameter and volume, and sometimes flux residue (from the wave). | | I suspect there is some incompatibility between one of our pastes and the glue. Our shop uses two types of paste and one type of glue. After some experimenting, I am seeing a STRONG correlation that suggests one paste type is not suited to the glue. | | Short of listing the specific materials used and getting myself in trouble, is there anyone who has encountered the problem I am talking about? What was the cause and how did you resolve it? | | If you want more specifics, please call me at 509-921-4311. | | Thanks, | Marc | |

reply »

| | | | Hi folks, | | | | | | | | I am looking for inputs on a long-standing problem encountered in our shop: Bottom-side parts (passives) falling off in the Wave...where the only remnants of the part is a flattened glue dot of proper diameter and volume, and sometimes flux residue (from the wave). | | | | | | | | I suspect there is some incompatibility between one of our pastes and the glue. Our shop uses two types of paste and one type of glue. After some experimenting, I am seeing a STRONG correlation that suggests one paste type is not suited to the glue. | | | | | | | | Short of listing the specific materials used and getting myself in trouble, is there anyone who has encountered the problem I am talking about? What was the cause and how did you resolve it? | | | | | | | | If you want more specifics, please call me at 509-921-4311. | | | | | | | | Thanks, | | | | Marc | | | | | | | Have you looked at the possibility that the parts are being knocked off in the handeling proccess prior to flow? If the glue is losing integrity from the use of an specific paste your torque readings should be considerably less with that paste type. | | | This is just my opinion, I could be wrong??? | | | | | | | | | | | | | | | Marc, | | | | Are you suggesting there may be some chemical incompatibility between paste/flux chemistries and the glue being used - thus intefering with the glue's ability to bond properly? I've only heard of this possibility, but never experienced it. Justin, Steve, Dave, or Wayne must have the answers to this issue. Where are you guys and all you others - the man has a serious problem and I can't answer it (obviously) though sitting but one aisle and four cubes away. | | | | Regards, | | | | Earl Moon | | | Earl, et al: | | Indeed I am, and I currently have two different vendors (of the materials in question) evaluating the chemistries, and the possibility that adhesion degredation is due to the paste/flux solvents or other compounds. The solvent, btw, is a very uncommon material, according to the folks I'm working with (not the mfr. of the paste). | | As for the question immediately preceding yours, I have another no-clean paste enroute that will allow me to conduct comparison experiments, and further validate (or not) the no-clean paste interaction theory. I doubt very much that handling is at the root here, although arguably all the torsion the boards go through prior to wave may have some cumulative effect that is not understood (at least by me). Getting the other paste will help bound this problem. | | Meanwhile...I wait for some lab results! | | Marc | Marc,

couple of questions and 1 suggestion.

Q1, are you gluing & pasting the bottome side as a double sided reflow operation or is it going onto go thro the wave ? if it's going onto the wave ...why use paste as well..?

Q2. if it's only for a double sided relow and you're adding glue to the passive's, ..why ? there should be enough paste to hold the component in place if you've got your profile right.

One other thing to look at, what is the finish of the component like...glassy, light / dark etc, if the finish is too smooth maybe your glue aint good enough.

reply »

| Hi folks, | | I am looking for inputs on a long-standing problem encountered in our shop: Bottom-side parts (passives) falling off in the Wave...where the only remnants of the part is a flattened glue dot of proper diameter and volume, and sometimes flux residue (from the wave). | | I suspect there is some incompatibility between one of our pastes and the glue. Our shop uses two types of paste and one type of glue. After some experimenting, I am seeing a STRONG correlation that suggests one paste type is not suited to the glue. | | Short of listing the specific materials used and getting myself in trouble, is there anyone who has encountered the problem I am talking about? What was the cause and how did you resolve it? | | If you want more specifics, please call me at 509-921-4311. | | Thanks, | Marc | | Marc: We use yellow colored glue (Locitite 36?? something or other, I think) to provide contrast with the solder mask so that, if components are missing or slopped on pads we can see it fairly easily. Our causes of lost second side glued components are:

1 Bumping or knocking the components prior to wave soldering 2 Poor control of glue curing, a fairly wide scale problem 3 Materials on the component (as Steve mentioned) or board that prevent good adhesion on a fairly wide scale

Ya know, that being said, I'm with you. I'm sure our wave is rippin' those little suckers off also, but I just can't find any of them in the solder pot. They must dive to the bottom of the pot and stay down there (You know, normally chips float on solder, if you just throw a handful on the wave as a kinda sacrificial thing?) (We do that each morning to appease higher beings. Hey it's better than salt!!!) or get masticated in the pumps or dissolve, or something.

But ya wanna know the really weird thing about not being able to find them in the wave? I CAN find them on the floor. Weird that the wave is putting them on the floor, isn't it? I think it's done at night. I keep pulling suprise bed checks, but the components are always on the floor before I get there ... sneaky eh?. And that's so out of character for the wave to do that. Weird, I'm tellin' ya. Mulder, Scully ...

Good luck

Dave F

reply »

| Hi folks, | | I am looking for inputs on a long-standing problem encountered in our shop: Bottom-side parts (passives) falling off in the Wave...where the only remnants of the part is a flattened glue dot of proper diameter and volume, and sometimes flux residue (from the wave). |

hi Marc.

Ah.....an age old problem. Really tough to pin down. Have you definitely pinned the missing passives down to the wavesolder work center? or is that just where you find the problem?

We've all experienced missing bottomside parts over the years. How thin is your substrate, and what mat'l is it? .031 teflon is lots harder to handle properly than .062 FR4. Here's something to try......

mid day, go out and walk the line. Start at ICT, and work backwards--through post wave ass'y, touchup / rework, cleaning (if you have the process), wave, and all areas in between the placement of the passives and the end of the process.

My money says hands down that in addition to a possible chemistry problem (which I've never experienced), you have a handling issue. Manual or conveyor....it IS there. Look inside the entrance and exit end of your cure oven. Look under tech benches, metro carts, conveyors, etc. Check trash cans from where the operators have swept up and dumped.

I have even observed passives go flying at the entrance and exit end of an in-line cleaner--they get pinched between the idler and drive sprockets and the wire belt links of the wire belt conveyor. No kidding. Check me out on this.

I found that by enforcing good handling disciplines, ensuring positive displacement pumps in good order, no air in glue / lines, and good glue (and ensure it's within the mfg'er's shelf life)......combined with buying brand new placement gear......my #1 defect on the floor virtually disappeared. BTW, I used Loctite 3609 with Camalot dispensing gear. Tried the Amacon yellow stuff, as well. Worked OK.

Please let us know what you find.

Scott Cook

reply »

| | | | | Hi folks, | | | | | | | | | | I am looking for inputs on a long-standing problem encountered in our shop: Bottom-side parts (passives) falling off in the Wave...where the only remnants of the part is a flattened glue dot of proper diameter and volume, and sometimes flux residue (from the wave). | | | | | | | | | | I suspect there is some incompatibility between one of our pastes and the glue. Our shop uses two types of paste and one type of glue. After some experimenting, I am seeing a STRONG correlation that suggests one paste type is not suited to the glue. | | | | | | | | | | Short of listing the specific materials used and getting myself in trouble, is there anyone who has encountered the problem I am talking about? What was the cause and how did you resolve it? | | | | | | | | | | If you want more specifics, please call me at 509-921-4311. | | | | | | | | | | Thanks, | | | | | Marc | | | | | | | | | Have you looked at the possibility that the parts are being knocked off in the handeling proccess prior to flow? If the glue is losing integrity from the use of an specific paste your torque readings should be considerably less with that paste type. | | | | This is just my opinion, I could be wrong??? | | | | | | | | | | | | | | | | | | | | Marc, | | | | | | Are you suggesting there may be some chemical incompatibility between paste/flux chemistries and the glue being used - thus intefering with the glue's ability to bond properly? I've only heard of this possibility, but never experienced it. Justin, Steve, Dave, or Wayne must have the answers to this issue. Where are you guys and all you others - the man has a serious problem and I can't answer it (obviously) though sitting but one aisle and four cubes away. | | | | | | Regards, | | | | | | Earl Moon | | | | | Earl, et al: | | | | Indeed I am, and I currently have two different vendors (of the materials in question) evaluating the chemistries, and the possibility that adhesion degredation is due to the paste/flux solvents or other compounds. The solvent, btw, is a very uncommon material, according to the folks I'm working with (not the mfr. of the paste). | | | | As for the question immediately preceding yours, I have another no-clean paste enroute that will allow me to conduct comparison experiments, and further validate (or not) the no-clean paste interaction theory. I doubt very much that handling is at the root here, although arguably all the torsion the boards go through prior to wave may have some cumulative effect that is not understood (at least by me). Getting the other paste will help bound this problem. | | | | Meanwhile...I wait for some lab results! | | | | Marc | | | Marc, | | couple of questions and 1 suggestion. | | Q1, are you gluing & pasting the bottome side as a double sided reflow operation or is it going onto go thro the wave ? | if it's going onto the wave ...why use paste as well..? | | Q2. if it's only for a double sided relow and you're adding glue to the passive's, ..why ? there should be enough paste to hold the component in place if you've got your profile right. | | One other thing to look at, what is the finish of the component like...glassy, light / dark etc, if the finish is too smooth maybe your glue aint good enough. | | John,

Our boards require paste and glue, rather than just glue because our cusomters are saving space by not designing in the "wave pad" used in the glue only application. Almost all of our hundreds of different boards are double sided paste-glue-reflow bottom side process flow.

The parts most affected are the capacitors (ceramic body), namely the very high use caps. This may be a phenomenon where the highest use parts also get the highest "visibility", but I am in the process of having their surfaces analyzed for any contamination that could be left over from the plastic tape punching process...it's a stretch, but at this point, I'm trying everything. Also, I recently found out from our paste mfr that a HIGH residue paste is NOT compatible with a polymer based epoxy (our exact situation). So I have 2 other labs on the outside checking the chemistries of the products involved.

Marc

reply »