Vapour phase Soldering

[IanB]

Why wouldn't one just use a hot air convection oven?

[jeremy]

Why wouldn't one just use a hot air convection oven?

When one BGA is worth $50, you don't have to make many toaster oven mistakes to cover the cost of the fluid

[IanB]

I'm not necessarily talking about home brew solutions, though it might apply here too.

If we are trying to solve the problem of uneven heating due to radiant paths, then a circulating hot fluid can help. But why does the hot fluid need to be some expensive chemical? If you circulate hot air around the board it seems you should get even heating in a similar manner at lower cost.

[wraper]

I'm not necessarily talking about home brew solutions, though it might apply here too.

If we are trying to solve the problem of uneven heating due to radiant paths, then a circulating hot fluid can help. But why does the hot fluid need to be some expensive chemical? If you circulate hot air around the board it seems you should get even heating in a similar manner at lower cost.

You won't get even heating with hot air. The main "thing" in this fluid is boiling temperature, it will condense at coldest parts (read heat them most), therefore even heating is guaranteed. Also you cannot overheat the parts.

[mikeselectricstuff]

I'm not necessarily talking about home brew solutions, though it might apply here too.

If we are trying to solve the problem of uneven heating due to radiant paths, then a circulating hot fluid can help. But why does the hot fluid need to be some expensive chemical? If you circulate hot air around the board it seems you should get even heating in a similar manner at lower cost.

The difference is latent heat of vaporisation.
The heater vaporises the liquid at its boiling point. Additional energy just turns more liquid to vapour without significantly increasing the temperature of the vapour. When vapour condenses on the PCB, it transfers heat efficiently, and will keep condensing as long as the PCB temp is below the vapour temp.
 
With air, there is no inherent control of temperature, and this will vary significantly due to convection etc. And you will need to start with air at a higher temp to compensate for this to get the temp you want at the surface, and this will vary with different PCBs

It also helps to be using an atmosphere that's chemically inert to avoid oxidation - this is why some conventional reflow setups use nitrogen, especially for higher temp leadfree prcesses.

[wraper]

[mrpackethead]

Main problem with vapour phase is tombstoning due to an imbalance of thermal properties of the design.

This is assisted by slightly more inert atmosphere in the vapour cloud compared to IR/ convection oven.

I am presently looking at a new  USD6000 vapour phase unit  versus a laser soldering robot for those pesky double sided jobs.

What machine is $6k, i'm curious.  A Laser won't do a BGA very well though.

[IanB]

Granted that the condensing inert vapour might be the "best" solution, but in other circumstances ordinary forced convection ovens are found to provide more even heating than radiant heating or natural convection. So I am wondering if a fan convection oven using temperature controlled hot air might be a less expensive and less complex compromise?

Sometimes the best can be the enemy of the good...

[mrpackethead]

Granted that the condensing inert vapour might be the "best" solution, but in other circumstances ordinary forced convection ovens are found to provide more even heating than radiant heating or natural convection. So I am wondering if a fan convection oven using temperature controlled hot air might be a less expensive and less complex compromise?

Sometimes the best can be the enemy of the good...

Yup, but thats going off topic.

[jeremy]

Granted that the condensing inert vapour might be the "best" solution, but in other circumstances ordinary forced convection ovens are found to provide more even heating than radiant heating or natural convection. So I am wondering if a fan convection oven using temperature controlled hot air might be a less expensive and less complex compromise?

Sometimes the best can be the enemy of the good...

I disagree that it is more complex. As you can see from the various links in the thread, all you need is a big test tube, an electric kettle and a variac. No PID, nothing. The only problem is the cost of Galden.

I am of course talking about homebrew style though.

Also, I just got a quote back for the IBL minilab. > $25k with no Galden included

[IconicPCB]

Jeremy,

Thats a Hawker Richardson price.

I have been quoted USD6K plus freight and importation.
Again fluid free price.

I am considering laser soldering robot for those pesky double sided boards.

Dont need much more than a 10W laser.

[jeremy]

Jeremy,

Thats a Hawker Richardson price.

I have been quoted USD6K plus freight and importation.
Again fluid free price.

I am considering laser soldering robot for those pesky double sided boards.

Dont need much more than a 10W laser.

Can you share your source?

Also, 10W at what wavelength? I can't imagine it will be a gas laser because that will just be reflected, so you'll probably still pay a pretty penny for the source itself.

[mrpackethead]

It just never ceases to amaze me how companies down here think that they can add redicouly large amounts of margins to things like PNP,  and other equipment. they seem to think that we don't read the internet..
Is the 6k for the IDL Minilab.   

Am getting The Galden LS230 priced up now out of Singapore.   

at 6k,    DIY might not be worth it.     

Jeremy,

Thats a Hawker Richardson price.

I have been quoted USD6K plus freight and importation.
Again fluid free price.

I am considering laser soldering robot for those pesky double sided boards.

Dont need much more than a 10W laser.

[helius]

I found this video demonstrating vapor phase soldering:

[jeremy]

I found this video demonstrating vapor phase soldering:
...

Very interesting. According to this, that is a prototype "oven" from multicore in the early 90s. It's still kicking I guess!

And that guy appears to work for 3M, hence the fluorinert instead of galden.

[SeanB]

If you are wanting to recycle the fluid ( it would be good at the price) You will need to ensure that the top cold zone is as cold as possible, preferably higher than the deep fryer, and with a good source of cooling. A water jacket with a pump to circulate the water and a chiller to make it cold will work best, using an old draught chiller as the cold water source. You need to keep the board in this zone for a while after reflow, to ensure that all the liquid condenses on it and then drips down into the bottom, and keep it running after you turn the heat off till it is at room temperature and you want to dispense it back into a sealable pressure vessel to store it.

Had that where I used to work, though there it was an ultrasonic cleaner using TCE, and it used the vapour phase to clean the fluid, as you could collect the drip off condensed liquid into a bucket and concentrate the dirty fluid down to a sludge for dirt removal. That would take the dirtiest castings you could find and leave them as bare clean white castings. Would strip a PCB of the resin ( except for the one conformal coat that survived no matter what) and all printed and painted parts except the copper and solder in around 3 minutes.

[jeremy]

If you are wanting to recycle the fluid ( it would be good at the price) You will need to ensure that the top cold zone is as cold as possible, preferably higher than the deep fryer, and with a good source of cooling. A water jacket with a pump to circulate the water and a chiller to make it cold will work best, using an old draught chiller as the cold water source. You need to keep the board in this zone for a while after reflow, to ensure that all the liquid condenses on it and then drips down into the bottom, and keep it running after you turn the heat off till it is at room temperature and you want to dispense it back into a sealable pressure vessel to store it.

Had that where I used to work, though there it was an ultrasonic cleaner using TCE, and it used the vapour phase to clean the fluid, as you could collect the drip off condensed liquid into a bucket and concentrate the dirty fluid down to a sludge for dirt removal. That would take the dirtiest castings you could find and leave them as bare clean white castings. Would strip a PCB of the resin ( except for the one conformal coat that survived no matter what) and all printed and painted parts except the copper and solder in around 3 minutes.

I'm actually very strongly considering trying this. I was thinking that the best approach would be one like you are describing; basically just go really slowly with everything. I can live with more than a 6 minute cycle if it means less wasted fluid.

[IconicPCB]

Vapour phase soldering machine from Wenesco USD5500 on their website.

http://www.wenesco.com/vapor.htm


A solid state diode laser source is sufficient for the task. New source and fiber under 2000 Euro new or under USD300 on ebay of questionable provenance.

[helius]

that configuration is very widely used for a cleaning technique called vapor degreasing. they usually can be set up with two tanks and when the working tank is full of dirt, it is "boiled down" to distill the solvent into the clean tank.

the solvents used for vapor degreasing are quite aggressive (high KB figures) compared to the inert fluids used for vapor soldering (should have KB at or near zero). They can sometimes cause leaching of plastic resins though, so this is something to test.

in the video from 3M, they are doing selective soldering using a low melting point alloy for attaching heat sinks. So they need to use a fluid with a boiling point between the melting point of their solder (140° C) and the solder in the processor package (probably SAC305: 217° C). The Fluorinert FC-40 has a bp of 155° C, so it is perfect in this application. Galden fluids have a boiling point of 230° C because they are designed for primary soldering.

[jeremy]

that configuration is very widely used for a cleaning technique called vapor degreasing. they usually can be set up with two tanks and when the working tank is full of dirt, it is "boiled down" to distill the solvent into the clean tank.

the solvents used for vapor degreasing are quite aggressive (high KB figures) compared to the inert fluids used for vapor soldering (should have KB at or near zero). They can sometimes cause leaching of plastic resins though, so this is something to test.

in the video from 3M, they are doing selective soldering using a low melting point alloy for attaching heat sinks. So they need to use a fluid with a boiling point between the melting point of their solder (140° C) and the solder in the processor package (probably SAC305: 217° C). The Fluorinert FC-40 has a bp of 155° C, so it is perfect in this application. Galden fluids have a boiling point of 230° C because they are designed for primary soldering.

This is true, but you can get Fluorinert FC-70 with a boiling point of 215C. I'm thinking they designed this for the pre-RoHS era and never updated their plans.

[free_electron]

vapor phase can be done reasonably cheap at home.

get an inductive single plate heater ( 50$ , amazon ) why ? no need to drill holes in the holding tank to shove a heating element in

get a nice high cooking pot compatible with inductive heaters. cooking pot needs a metal lid. if you can find a 'turkey fryer' style pot ( those are really deep and come with a basket )

put half an inch of galden in the pot and switch on.

use a fryer basket : put in the board on pegs so it does not contact the basket.

hang the basket so the board hangs in the cloud of vapor.

put on the lid. around the top of the pot : run metal tubing a couple of times and run cold water ( a bucket of water with some ice in it and a 12 volts dc pump will do the trick.
the same for the lid : attach a spiral of metal tubing and run cold water there as well.

you may want to attach an inverted cone to the lid so the condensation droplets fall down on the cone and then roll towards the side.

hmm.. difficult in words. lemme draw it

hang on

[free_electron]

there ya go.

the cooled cone attached to the lid condenses the vapor and makes it run down towards its edges where the droplets fall back down , away from the pcb.

there is a swiss dude that built one on this principle. works well aparently

[IconicPCB]

What happens to the working fluid in presence of flux  in solder paste?

[mikeselectricstuff]

What happens to the working fluid in presence of flux  in solder paste?

nothing - it's inert

[IconicPCB]

Does it get contaminated?
If so how is contamination handled? distillation?

Is contaminated how is the equipment affected by contaminants?