Vapour phase Soldering

[SeanB]

You are only using the vapour, so it is not worried much by flux and other residues. Eventually you just run it with a drip tray collecting the condensed liquid till it is dry below and with the now distilled liquid up in the top. Then clean the boiler area and reuse the fluid.

[mrpackethead]

(1) do you need to do anything in terms of pre-heating your board before it goes into the vapour?

(2) How "thick" is the vapour layer above the liquid.

(3) Im thinking that a 'lid' that has a stack of aluminium heat sinks that sits over the tank might be quite good a well.

(4) If the board is sitting flat, after it is raised out of the vapour,  the vapour could potentially condense and sit on the board, ( once its cold ).   Maybe when you pull it up, it needs to be angled once the solder is solid?

(5) Double sided..  Just rely on surface tension?

[coppice]

(1) do you need to do anything in terms of pre-heating your board before it goes into the vapour?

Not really. Just don't heat too quickly, so the vapour has time to warm the components through.  While the core of the components is still drawing heat from the surface you won't get a good flow of solder. That might not be a big issue with small components, but it can be with big chunky power ones.

(2) How "thick" is the vapour layer above the liquid.

(3) Im thinking that a 'lid' that has a stack of aluminium heat sinks that sits over the tank might be quite good a well.

I'm not clear what you are trying to achieve there? It seems like that just going to make the fluid condense on the lid instead of the board. The whole things works by the latent heat of condensation ending up in the components.

(4) If the board is sitting flat, after it is raised out of the vapour,  the vapour could potentially condense and sit on the board, ( once its cold ).   Maybe when you pull it up, it needs to be angled once the solder is solid?

That isn't a problem when things are done right. If the vapour pressure is greatly reduced while the board is still warm, the fluid evaporates. The board should come out dry.

(5) Double sided..  Just rely on surface tension?

Yes. Make sure nothing jolts that board before soldering.  Alternatively glue spot.

[jeremy]

(1) do you need to do anything in terms of pre-heating your board before it goes into the vapour?

(2) How "thick" is the vapour layer above the liquid.

(3) Im thinking that a 'lid' that has a stack of aluminium heat sinks that sits over the tank might be quite good a well.

(4) If the board is sitting flat, after it is raised out of the vapour,  the vapour could potentially condense and sit on the board, ( once its cold ).   Maybe when you pull it up, it needs to be angled once the solder is solid?

(5) Double sided..  Just rely on surface tension?

1) yes, otherwise you will risk tombstoning or wicking. It's actually more of a problem with smaller components apparently because they will heat much faster than the pads. 150C preheat makes this problem basically go away. Source is in one of my previous posts in this thread. Usually an different batch oven is used, but that's only because in production everything needs to go super fast.

2) depends on the amount of heating. Seems like as long as the board is in the vapour and not the fluid it will be ok. If you're talking about density, it is heavier than air.

3) yes, or maybe some of those watercooling blocks from overclocking kits too? coppice: you want the fluid to condense on the lid so that it doesn't escape.

4) I don't see why you couldn't just wash the board with distilled water in place after everything had cooled down, then just run the thing on 120C for a bit to evaporate the water. When I finally get hold of some galden or fluorinert (I'm trying to get 500mL, but the cheapest I've got so far is about AU$1200 for 3.8L), I think it would be good to do some experiments to see just how much loss you get when you do something like this.

5) thermosetting epoxy designed for smd is not that expensive, and I've got it from RS in the past in plastic syringe form. I'd post the link, but I'm getting "RS Online is temporarily unavailable."

Cleaning the fluid should be straightforward. You can get a liebig condenser on ebay from china for like $30.

free_electron: nice pic. what did you draw it in?

[coppice]

3) yes, or maybe some of those watercooling blocks from overclocking kits too? coppice: you want the fluid to condense on the lid so that it doesn't escape.

As the vapour is heavier than air, why would it escape? You aren't going to lift the lid until the vapour pressure has dropped.

5) thermosetting epoxy designed for smd is not that expensive, and I've got it from RS in the past in plastic syringe form. I'd post the link, but I'm getting "RS Online is temporarily unavailable."

The material is cheap, but applying it can be a pain. People don't really like doing this, unless its being done by a fully automated machine that tightly controls quantities. If you dispense just a little too much, or don't place it accurately,  it can spread onto the pads, especially if the parts are very small.

Take what I say with a pinch of salt. I'm basing what I say on results in the 1980s with lower temperature fluid and leaded solder.

[jeremy]

3) yes, or maybe some of those watercooling blocks from overclocking kits too? coppice: you want the fluid to condense on the lid so that it doesn't escape.

As the vapour is heavier than air, why would it escape? You aren't going to lift the lid until the vapour pressure has dropped.

Actually, this is not true in a batch-like system. See the CPU soldering video above; there is no lid on the vessel, just a cooling jacket. You will also still get some loss due to natural diffusion even though it is denser. I'm not sure if this is worth worrying about, but the stuff is expensive so someone needs to test that.

Also, I'm not sure I would feel comfortable having a 100% sealed container being vigorously heated. It might also increase the BP of the fluid. Again, I think someone needs to actually measure the pressure and/or BP changes to see if these are negligible or not.

Does anyone have any contacts at 3M/dupont/solvay that they can put me in touch with? I've already tried contacting 3M AU but they (understandably) haven't got back to me yet.

Take what I say with a pinch of salt. I'm basing what I say on results in the 1980s with lower temperature fluid and leaded solder.

Thanks for your opinion anyway. It's useful to get many points of view and many people's experiences.

[helius]

There are some interesting videos from the manufacturers of these machines.
This company has several videos: This one demonstrates a jig that can desolder a BGA using the same machine and process as for soldering:

Depopulating a whole board would take ages, but it's not uncommon to only rework one large component. The large BGAs are also the hardest to rework with hand tools.

Especially interesting is their top-line product, the VP6000, which has a vacuum system. I assume it is used to recover more of the fluid, or to make the boards dry faster. They have options for filtering flux residue and for convection coolers to enhance board drying.
It also shows that they can calibrate the process using a live device inserted into the vapor bath.

[jeremy]

There are some interesting videos from the manufacturers of these machines.
This company has several videos: This one demonstrates a jig that can desolder a BGA using the same machine and process as for soldering:

Depopulating a whole board would take ages, but it's not uncommon to only rework one large component. The large BGAs are also the hardest to rework with hand tools.

Especially interesting is their top-line product, the VP6000, which has a vacuum system. I assume it is used to recover more of the fluid, or to make the boards dry faster. They have options for filtering flux residue and for convection coolers to enhance board drying.
It also shows that they can calibrate the process using a live device inserted into the vapor bath.

Nice, thanks for sharing.

Asscon (related to siemens?) actually has some recent patents on the vacuum. It outgasses the solder to extract any flux pockets caught inside the joints that would cause solder voids. See http://www.smtnet.com/library/files/upload/Vapor-Phase-Soldering.pdf

[mrpackethead]

1) yes, otherwise you will risk tombstoning or wicking. It's actually more of a problem with smaller components apparently because they will heat much faster than the pads. 150C preheat makes this problem basically go away. Source is in one of my previous posts in this thread. Usually an different batch oven is used, but that's only because in production everything needs to go super fast.

Tombstones are just painful.  We get the odd ones happening in our reflow oven from time to time.   For a small batch / prototype system,   I wonder if it will be sufficient to heat the board in a convection oven, and then drop on the tray ( gently! )

2) depends on the amount of heating. Seems like as long as the board is in the vapour and not the fluid it will be ok. If you're talking about density, it is heavier than air.

I was more thinking about physically how high the vapour cloud will be, and how big the "zone" between 'air' and the vapour will be.  ( there will be some 'transition zone' )

[/quote]
3) yes, or maybe some of those watercooling blocks from overclocking kits too? coppice: you want the fluid to condense on the lid so that it doesn't escape.
[/quote]

Now that sounds like quite a good idea.   Though i have some good pelters and some fans, that would be very effective as well, and i save yet more water pumping around.

4) I don't see why you couldn't just wash the board with distilled water in place after everything had cooled down, then just run the thing on 120C for a bit to evaporate the water. When I finally get hold of some galden or fluorinert (I'm trying to get 500mL, but the cheapest I've got so far is about AU$1200 for 3.8L), I think it would be good to do some experiments to see just how much loss you get when you do something like this.

It may not even be a problem. but looking at another machine it seems they have implemented some fans that blow air over the board once its out of the vapour.  My guess is that this helps dry them. ( a bit like squiring compressed air on a board, that you've washed in IPA.. ).. 

[mrpackethead]

Does anyone have any contacts at 3M/dupont/solvay that they can put me in touch with? I've already tried contacting 3M AU but they (understandably) haven't got back to me yet.

Been in touch with solvay, and am waiting for an answer.

[ElektroQuark]

Quote from: jeremy on Today at 09:47:20 PM

Asscon (related to siemens?)

I looks like the use a SIEMENS PLC to control the system (logo on LCD screen).

[jeremy]

2) depends on the amount of heating. Seems like as long as the board is in the vapour and not the fluid it will be ok. If you're talking about density, it is heavier than air.

I was more thinking about physically how high the vapour cloud will be, and how big the "zone" between 'air' and the vapour will be.  ( there will be some 'transition zone' )

The vapour cloud height should actually be a function of the heating amount. My simple understanding of chemistry says that you will get more diffusion by injecting more energy into the system, hence more height. In the 3M video, they clearly solder 2 LGA1155 cpus in height plus some, so that's 37.5mm * 2 + a bit >= 75mm of "soldering zone". Sure, its a lower boiling point fluid, but it's a ballpark at least.

Some interesting historical patents (I found these on the derwent innovation index which I have access to, but they are all on google patents for free):

US 4394802 A - Counter-convection vapor control system; similar to free_electron's pic
US 4549686 A - Vapor phase soldering using perfluorotetradecahydrophenanthrene (C14 F24); this seems like the original commercial patent, but uses CFCs
US 4628616 A - Vapor tank; Hitachi patent on a vapour phase soldering tank
US 4996781 A - Vapor reflow type soldering apparatus with an improved flux separating unit; a later Hitachi patent
US 4681249 A - Vapor phase soldering apparatus; a continuous soldering system
US 4806662 A - Fluids having an oxetane structure and improved characteristics for special applications; the original "galden" patent

The original galden was created by a company known as Ausimont in Italy. Since galden lists Italy as the contact on the msds I became a little suspicious, so I did a bit of poking around and sure enough I also discovered that they merged with solvay later on: http://ec.europa.eu/competition/mergers/cases/decisions/m2690_en.pdf

Will post if I find any more interesting stuff.

[coppice]

US 4549686 A - Vapor phase soldering using perfluorotetradecahydrophenanthrene (C14 F24); this seems like the original commercial patent, but uses CFCs

The fluids we used in the early 80s for leaded soldering were CFCs. The electronics industry used to put enormous quantities of CFCs into the air back then. Boards were mostly cleaned with CFC113, commonly referred to as "trik", and the stuff was cheap enough that it was allowed to evaporate quite freely.

[jeremy]

A few more (not sure if these are available on google patents):

DE20300375-U1 - Cooler for workpieces or components under cooled vacuum dome, especially in vapor phase vacuum soldering installation; Asscon patent from 2003
DE20300374-U1 - Assembly for temp. treatment of workpieces or components in controllably heated vacuum dome; similar to above, also from asscon

US 20070194083 A1 - Process and device for soldering in the vapor phase; IBL patent on using a vacuum to stop void forming inside joints. Still valid! This is the company that sells the US$6k/AU$26k ( ) machine mentioned earlier

US 4871109 A - Vapor phase soldering using certain perfluorinated polyethers; chemistry patent from everyone's favourite chemical company: Monsanto

http://www.emeraldinsight.com/doi/abs/10.1108/SSMT-10-2013-0028 - A very recent paper on solder voids due to vapour phase soldering
I read the paper but its not free access so I'll give you all the short version: yes you get voids without a vacuum. RoHS or not doesn't seem to make a difference.

http://www.emeraldinsight.com/doi/pdfplus/10.1108/eb037911 - Voids in BGAs.
Again, not free, but the summary is:
- convection oven with changeable air/nitrogen supply
- all BGAs originally came with no voids (verified by xray), Sn63 balls with ~1.5mm pitch
- no vacuum was used
- going to higher temps means more % voids
- flux/paste solvents with lower boiling points tend to cause more voiding (solder paste solvents are worse)
- recommended profile with worst solder paste caused about 1-2% voiding, best was 0.5%
- mesh size of the paste balls does show a minor trend towards smaller being better (200-300um instead of 500um), but it is so slight that I'm not buying it
- reflow atmosphere doesn't matter (air or nitrogen) with respect to void formation
- flux activity doesn't matter

Hopefully someone else is finding this as interesting as I am 

[ElektroQuark]

How do they use the vacuum?
If you apply vacuum the boiling point of liquid drops highly, so the temperature necesary for the soldering can't be reach.

[jeremy]

How do they use the vacuum?
If you apply vacuum the boiling point of liquid drops highly, so the temperature necesary for the soldering can't be reach.

If you look in the patent, it has a separate chamber which is pulled under a vacuum. So while the solder is still liquid, it the board moves from the vapour blanket to the chamber and the vacuum does its work. Cooling isn't really a problem, because it's a vacuum!

[ElektroQuark]

So they remove gas bubbles and liquid from the board, avoiding gaps.
Cooling is important for the solder (Pb, Sn,etc), not for the soldering liquid, of course.

[helius]

Asscon (related to siemens?)

They are both German; but I get the sense that they use all COTS PLC equipment, and Siemens is a big player in PLCs.

I haven't seen it mentioned here yet, but vapor phase is an oxygen-free process. As the board is lowered into the vapor cloud, the condensing liquid forms a film all over it and so there is no oxygen at the joint during reflow. Separate pre-heating is not needed because the ramp rate is controlled by the heat input to the liquid.

[jeremy]

One last interesting patent: US 4523039 A - Method for forming perfluorocarbon ethers. Check out the number of patent references, mostly by Solvay.

Basically get a polyether, add fluorine and heat up like crazy. I think this one is best left to the experts...

[mrpackethead]

The Wenesco Machine looks not too bad, and actually demonstrates how simple the system probably is.   There is a lot of potential here to engineer a solution that is a lot more complex than it needs to be, i suspect.





Its a bucket with a lifter in it!!

I'm seeing a project coming up here.    I'm thinking that it would be wise to start with building an experimental system.  I had thought about using the deep fryer, but the issue is that the element will need to be completely covered, and that in itself will require a lot of the expensive juice.     time to rethink the heating, just a little bit.   

What seems apparent to me, is that a pretty decently tall system will be your friend here.   I'm thinking you probably want a tank that is a minimum of 600-700mm deep.

[SeanB]

Like Vincent said use an external heater. Either a simple solid plate that you hard solder to the bottom of the chamber, and insulate with a lot of high temperature insulation, or use the Free electron style of an induction heater coil coupling energy directly into the base plate.

[ajb]

How do they use the vacuum?
If you apply vacuum the boiling point of liquid drops highly, so the temperature necesary for the soldering can't be reach.

If you look in the patent, it has a separate chamber which is pulled under a vacuum. So while the solder is still liquid, it the board moves from the vapour blanket to the chamber and the vacuum does its work. Cooling isn't really a problem, because it's a vacuum!

If you look at the video for the VP6000 (), it looks like the process is this:

- Load carrier into entry chamber.
- Transfer carrier into process chamber.
- Lower carrier into soldering zone.
- Lift carrier out of soldering zone.
- Activate drying heat and vacuum.
- Transfer carrier into entry chamber.
- Cool.
- Unload.

The drying heat appears to be a halogen tube lamp.  Presumably the drying heat also keeps the solder molten long enough for the vacuum to properly degas it, and also controls the cooling speed to avoid quenching.  The vacuum system (which is very nicely labeled!  Too bad more of the labels aren't legible in even the HD video) appears to use filters to recover the process fluid.  I'm kind of surprised they didn't go for a cold trap, since the system already has a water cooling loop, but I guess the filters are effective enough.

[gxti]

no isolation of the wires here: http://www.ibrtses.com/projects/vapourphasesoldering.html

I like the "bigass beaker" approach. Instead of immersion heating, throw a steel puck in the bottom and stick the whole thing on an induction heater. Maybe even find a metal with a lower curie point for the puck so it can't overheat if it "runs dry" -- once it hits that point, it should just stop being heated by the magnetic field.

[jeremy]

...

Its a bucket with a lifter in it!!

I'm seeing a project coming up here.

you and me both. now to find some of this pesky galden...

[IconicPCB]

Yes.. I thought Wenesco had a good approach in their batch oven.