Practical DIY or modest cost Vapor Phase - does it exist yet?

[mrpackethead]

Would something like a 20q/l roasting oven be a useful starting place? 

too shallow unless you are willing to loose Galden, which is much more expensive than that oven.

Actually, maybe not.  My work with using the frying pan says that you can have a shallow pan.  The galden is actually reletively contronable.   that is not the problem that you solve with using a deep container.    Using the deep container lets you build MUCH better profiles, becuase you can build the elevator platform. That is the game changer.   I had avoided it, because I could not come up with a mechanically simple way to build it.  What Lukas05 has done is come up with a mechanically simple way to make it work. That is the game changer.

I scoured Aliexpress looking for a deep Dish, and have failed. :-(

[pisoiu]

I scoured Aliexpress looking for a deep Dish, and have failed. :-(

Same here with aliexpress but there may be another way. I made some tests with that kind of rectangular containers which are used to serve icecream and food in malls. Don't know what they are called. But that format has maximum 200mm depth. I wanted deeper and I was able to find at a local dealer of stainless steel supplies. Among other things (tubes, bars, various containers), they sell containers for dish washing in restaurants. I find those suitable, they are quite big, and they have another feature, the bottom is slightly sloped, which in my opinion is a good thing to use the deepest part for heater.
Check this pdf: http://www.italinox.ro/images/pdf/gastronorm.pdf , at page 4 and 5 you will find them. Texts are in Romanian but images and drawings are self explanatory. Walls are 1mm thick and 330mm depth seem ok for this application.

[tautech]

I scoured Aliexpress looking for a deep Dish, and have failed. :-(

Really ? 

Plenty here:
https://www.aliexpress.com/wholesale?spm=a2g0o.productlist.0.0.44464cbeo8t7Ji&SearchText=stainless%20steel%20pot%20with%20lid

275L one big enough ?
https://www.aliexpress.com/item/4000070766006.html?spm=a2g0o.productlist.0.0.44464cbeo8t7Ji&algo_pvid=b151105f-39f8-4a2d-8913-e8c848d8aecb&algo_expid=b151105f-39f8-4a2d-8913-e8c848d8aecb-4&btsid=95242914-9696-4d11-9c60-e6e31836fb72&ws_ab_test=searchweb0_0,searchweb201602_9,searchweb201603_52

[Kjelt]

rectangular containers which are used to serve icecream and food in malls. Don't know what they are called. But that format has maximum 200mm depth.

those are standard for the horeca and called Gastronorm.

You can use it as the base that contains the fluid and get a stainless steel plate chassis welded upon it as high as you want.

[mrpackethead]

I scoured Aliexpress looking for a deep Dish, and have failed. :-(

Really ? 

Dish as opposed to Pot.  Being round, it makes it difficult to attached flat plate water coolers ( as used for GPU's etc ) which are readily avaialble and cost effective.     It seems that Gastronorms max out at 200mm deep.

[mrpackethead]

I scoured Aliexpress looking for a deep Dish, and have failed. :-(

Same here with aliexpress but there may be another way. I made some tests with that kind of rectangular containers which are used to serve icecream and food in malls. Don't know what they are called. But that format has maximum 200mm depth. I wanted deeper and I was able to find at a local dealer of stainless steel supplies. Among other things (tubes, bars, various containers), they sell containers for dish washing in restaurants. I find those suitable, they are quite big, and they have another feature, the bottom is slightly sloped, which in my opinion is a good thing to use the deepest part for heater.
Check this pdf: http://www.italinox.ro/images/pdf/gastronorm.pdf , at page 4 and 5 you will find them. Texts are in Romanian but images and drawings are self explanatory. Walls are 1mm thick and 330mm depth seem ok for this application.

Yes, it seems that the 'Gastronorm' maxes out at 200mm.  They all come in standard sizes.

[Lukas05]

Hi,

since some of you liked the project, I thought I would continue documenting the progress

A problem occured in the last run of the vapor phase oven. The container is sealed pretty good at the moment. There is only a small hole in the side wall for the heating element cables.
This poses a problem when it comes to pressure equalisation. The relatively cold PCB is lowered into the vapor phase. A lot of the gas condenses onto the cold surface in the first few seconds.
This causes the pressure in the system to drop, and air from the outside rushes into the chamber. This by itself isn´t a problem. But it gets problematic when lifting the hot PCB out of the vapor phase.
By lifting up the hot board into cooler parts of the chamber, the pressure rises. This causes galden vapor to get pushed out of the chamber. This of course means unwanted galden loss over time.
Currently I´m designing sorf of a reflux condenser. This will probably just be a 6mm copper pipe bent into an air cooled coil. The galden condenses in the pipe, and flows back into the container, while also providing a controlled pressure equalization.

I´m goint to verify this design in a FEM simulation. I´ve already startet so simulate parts of the project, to get some ideas of the material stresses and also transient behaviour of the system.
This will also be the foundation along with the measurements for the temperature profile regulator. I will post the results once I have calculated them through.

The post process of the coupled cfd simulation looks beautiful already I think
Here you can see the convection of the galden until 30 seconds after turning on the heating element.

https://www.flickr.com/photos/136915492@N02/49070612068/in/dateposted/

[Kjelt]

Wow you are really taking it a step further 
Just trying to think along could it be that your base carrier plate is to massive, taking too much Galden up?
If I look at this pro machine the boards carrier is only a wire mesh
https://youtu.be/jVQTcViEaUk

[mrpackethead]

Wow you are really taking it a step further 
Just trying to think along could it be that your base carrier plate is to massive, taking too much Galden up?
If I look at this pro machine the boards carrier is only a wire mesh
https://youtu.be/jVQTcViEaUk

You'd get the same issue, if you have a pcb that is large ( its also solid ).     

[mrpackethead]

A problem occured in the last run of the vapor phase oven. The container is sealed pretty good at the moment. There is only a small hole in the side wall for the heating element cables.
This poses a problem when it comes to pressure equalisation. The relatively cold PCB is lowered into the vapor phase. A lot of the gas condenses onto the cold surface in the first few seconds.

This would absolutely be correct.  This raises another problem. If you lower the PCB into the Galden too fast, your re flow profile is going to be way too steep on the way up.    With the 'static' frying pan method, i was effectivel 'preheating' the pcb. ( giving it a soak period ) by regulating the amount of energy that was going into the system.    If you were to monitor the temp of the PCB,  ( you'd need a thermocouple attached it ), you' would be able to control the rate at which it heated up quite well.  Above the 'thick' galden cloud, is a 'soft' cloud, and that vapour provides a pretty useful way of getting teh board up to a temp 10 lower than the melt point of the solder paste.

Galden vapour is easy to disturb. you only need a tiny draft and you'll blow it out of the tank.  However, If its 'still', it really does sit nicely, and unless you put LOTs of energy into it, its easy to keep it in the tank.     Moving PCB's in and out of the vapor does create a lot of movement.


This causes the pressure in the system to drop, and air from the outside rushes into the chamber. This by itself isn´t a problem. But it gets problematic when lifting the hot PCB out of the vapor phase.
By lifting up the hot board into cooler parts of the chamber, the pressure rises. This causes galden vapor to get pushed out of the chamber. This of course means unwanted galden loss over time.

Currently I´m designing sorf of a reflux condenser. This will probably just be a 6mm copper pipe bent into an air cooled coil. The galden condenses in the pipe, and flows back into the container, while also providing a controlled pressure equalization.

That has proved to be an effective way of keeping galden in teh tank as well.  I tryed using it to cool my frying pan project, and it worked.     The big problem with teh frying pan project is that while i can get really good results,  its a very long cycle time..  as the galden goes from cold to nearly cold, and thats the best part of 15 minutes.

Your design is really very clever using the cable lifts.    Some of the commerical VP machines bring the board out of the vapour into the top of tank, and then a door closes between teh board and the bottom of the tank.   I think thats doable with another cable lift.   I'll try and draw a picture tommorrow of what i mean. 
If that could work, you'd be able to get great cycle times, becuase you'd probably nto even need to cool the galden down.

[Lukas05]

Hi,

If you were to monitor the temp of the PCB,  ( you'd need a thermocouple attached it ), you' would be able to control the rate at which it heated up quite well.

I think this is the route I will go with this. The mass of the PCB varys a lot potentially from small PCBs with 0.8mm thickness, to 2.4mm multilayer boards with huge inductors.
This of course influences the time constant of the thermal low-pass wich ist formed with the heat capacity of the PCB. I´m thinking of an easy mechanism that presses a thermocouple on the PCB.... (not so sure how I would tackle this right now)
Another approch I had in mind would be to fixate a thermocouple on a small PCB with a ground plane that is installed permanently somewhere where it doesn´t take up much space.

Some of the commerical VP machines bring the board out of the vapour into the top of tank, and then a door closes between teh board and the bottom of the tank.

Yeah the two chamber design is very common it seems. I´m not sure if I want to install something similar, since it adds quiet a bit complexity an potentially BOM cost to the project.
Since I don´t need very low cycle times I think I will first evaluate the quick cool concept in detail. If I´m not satisfied with the result I will investigate the two chamber hybrid.

Just trying to think along could it be that your base carrier plate is to massive, taking too much Galden up?

Yes I think I will also swap that base out for something that has a higher "hole-to-metal" ratio.

[Koen]

I tried Gastronorm and the walls were too thin. I needed a bigger area for bigger panels and pots + glass lids + 5cm high spacers won again. 30 EUR for each set.

[Lukas05]

Hi everyone,

the project progresses fast which means it is time for an update 
(I think I´ll start a thread in the "projects, designs" sub forum next time)

The prototype got a temporary casing, which will be used for thermal performance testing of the system. The Airflow of the axial fans get´s hindered quite a bit by the casing, which means the steady state temperatures will also move a bit. I wanted do make a thermal simulation at fist, but the meshing process of this intricate geometry with all the cables and tubes would have taken a lot of time.
Thermal performance measurements are coming up soon.

https://www.flickr.com/photos/136915492@N02/49148444541/in/dateposted/

The display doesn´t function at the moment. I still have to throw the controller PCB together.
For quick testing I´ve decided to use the classic perfboard approach. The Board is located directly under the water pump  Which is kind of okay since it is only a 12V system...
The final PCB of course will be placed somewhere else in the case.

https://www.flickr.com/photos/136915492@N02/49148444301/in/dateposted/

As mentioned earlier I also wanted to swap out the carrier plate. The new plate has a much lower thermal mass, and also allows much more vapor to come through.
I used this opportunity to max out the available space, and increased the overall area of the carrier plate.

https://www.flickr.com/photos/136915492@N02/49148653877/in/photostream/

On the run I´ve installed a bigger window, and equipped the lid with some LEDs for a more pleasant PCB stalking experience during the solder process.

https://www.flickr.com/photos/136915492@N02/49147953273/in/photostream/

https://www.flickr.com/photos/136915492@N02/49147953628/in/dateposted/

[mrpackethead]

The prototype got a temporary casing, which will be used for thermal performance testing of the system. The Airflow of the axial fans get´s hindered quite a bit by the casing, which means the steady state temperatures will also move a bit. I wanted do make a thermal simulation at fist, but the meshing process of this intricate geometry with all the cables and tubes would have taken a lot of time.
Thermal performance measurements are coming up soon.

Do you even need to do encase this?  I found that it really impacted the cooling phase a Lot.   I dramatically increased my airflow, to get a resonable cool down.       The other thign was to not place the fans just above a bench. ( like another brand i wont' mention ) did.. It almost rednered them useless, and only slightly better than passively cooling.

[Lukas05]

Do you even need to do encase this?

Hmm if I were the only one to use this oven, I wouldn´t encase ist. My colleagues will probabely use the oven for prototyping as well... It feels much saver this way

The other thign was to not place the fans just above a bench

This is something I noticed as well. The inlet of axial fans in general is really sensitive to obstructions of the air flow. This effect seems to be highly dependent on the fan blade design, the hydraulic diameter of the fan itself, the rotational speed of the fan, and the pressure differential across the fan. The fans I use need at least 4cm - 5cm space to deliver acceptable performance.

[cgroen]

Any updates of your nice project Lukas ?

I have been looking at the IMDES VP "ovens", I soon stopped again when I saw the pictures of the internals.....
So, the more I think about it plus the amount of advice here on the forum by all those that have tried the vapor phase, I'm beginning to realize that this might be a possible solution for me (although I'm pressed on time). I have been lurking around first for a decent reflow oven, didn't find "the perfect one".
If one were to go with vapor phase, what is the "minimum" height of the container for the Galden ? I see that 300mm is mentioned, is it so that this is the minimum, or should f.ex 200mm also work fine without too much loss of galden ? (Gastronorm containers are usually max 200 mm). Also regarding cooling, as I understand it, you will need active cooling ?
I'm looking at something in the size of 320 x 260 mm PCB area (at least)

[Koen]

Regarding height, just close the lid.

Gastronorms I tried were too thin and bowed or popped during the process.

[cgroen]

Regarding height, just close the lid.

Gastronorms I tried were too thin and bowed or popped during the process.

Thanks Koen! I found some 0.7mm stainless steel ones, but guess they are no good then, pity as they are easily available (and claim they are good to 300 degC) !
I wonder what type the container Lukas uses, didn't find any hints so far

[Koen]

His lid looks like this : https://www.nisbets.be/fr/couvercle-hermtique-inox-et-silicone-vogue-gn-12/cp269

I found I lost a lot of space in Gastronorms due to the rounded corners (versus my expectations) and I hated how unsafe it felt. I didn't want to deal with a 230C issue nor nanny-watch it. Regular pots don't cause me concerns and my panels fit so I thought "k, good enough".

[KaneTW]

Yeah, the IMDES ovens are horrifying.

When I was engineering a vapor oven design, I arrived at the conclusion that you pretty much need active cooling to get the dT/dt required.

Gastronorms will warp under the temperature--whether that's acceptable for you or not is another matter. Since the requirements for the container aren't super strict, I'd just weld together a container of acceptable thickness stainless if a Gastronorm doesn't work out.

Deep friers just use Gastronorm afaik.

[cgroen]

Thanks guys!
I was looking at these: https://www.expondo.dk/royal-catering-gastrobakke-gn-1-2-200-mm-10011039
They claim +300 degC working. Guess its time for experiment....

[KaneTW]

I've done a quick simulation for thermal stresses when going from 300K to 600K. A (slightly simplified) GN 2/3 0.7mm container with fixed constraints on the rim will buckle about 2mm outwards on the sides and 1mm inwards on the bottom. The stresses on the body are acceptable (<80MPa). The constrained rim gets hit with a lot of stress but that's a modelling artifact. Allow for about +- 1mm of warping around the rim.

[cgroen]

I've done a quick simulation for thermal stresses when going from 300K to 600K. A (slightly simplified) GN 2/3 0.7mm container with fixed constraints on the rim will buckle about 2mm outwards on the sides and 1mm inwards on the bottom. The stresses on the body are acceptable (<80MPa). The constrained rim gets hit with a lot of stress but that's a modelling artifact. Allow for about +- 1mm of warping around the rim.

Thanks Kane,
thats not too bad I think ?
Just trying to figure out if I go with the GN 1/1 or the GN 1/2.
I found a 2000W heating element https://horecatiger.eu/en-eu/shop/heating-element-2000w-230v-heating-circuits-1-415669
that would fit the GN 1/1 quite nicely, but I'm afraid it wont heat the Galden too well as the heating element is right at the edge of the tank (the tank is 530 x 325 mm externally, the heater is 452 x 256 mm) ?

In the pciture below I have placed a 8 mm thick alu plate (the heater is 6.3 mm and will sit around 1.5mm above the bottom because of the size and the corners in the tank) in the center (gray area) to consume some volume (and lower the amount of Galden, like on Lukas design), if I put Galden in it so that the alu plate has 5 mm Galden on top of it I need a little short of 1 liter Galden.
But again, I'm not sure this will work with the large size of the tank (and the heater at the perimeter)....
Maybe the GN 1/2 is easier to get to work in this way, although I could really use the size the GN 1/1 gives me.....

[KaneTW]

Quote from Solvay's Galden FAQ:

For Galden® PFPE grades, a general recommendation is not to install a heater watt density higher than 4 ÷ 5 W/cm2, in order to avoid instability in heater surface temperature for low flow units (i.e. fluid reservoir in the TCU). Having the heater section where fluid flow is present near the outlet of the pump, it will allow higher watt density heaters to be used. The critical breakdown is ~ 14 W/cm2. We recommend 4 W/cm2. The decision to go beyond this level should be made by the customer.

I would not use a insertion heating element due to the low surface area. That heater will give you 9 W/cm2, which is fairly high.
Like you mentioned, it's also at the edge of the tank, so heating will be fairly localized.

My recommendation is to heat the entire bottom of the tank. Get a bunch of mica heating strips so that the bottom of the tank has a watt density of 4-5W/cm2 (remember steel is fairly poor thermal conductor). Epoxy them to the bottom of the tank with a heat-resistant epoxy. Same with aluminum water cooling heatsinks. Mount the container with a flexible high-temperature adhesive (e.g. Loctite 596).

Unless you have the money to spend, I strongly recommend sourcing the elements and heatsinks from China. The prices I saw for those at Western sites are absurd (>200EUR for a heatsink etc)

[cgroen]

Quote from Solvay's Galden FAQ:

For Galden® PFPE grades, a general recommendation is not to install a heater watt density higher than 4 ÷ 5 W/cm2, in order to avoid instability in heater surface temperature for low flow units (i.e. fluid reservoir in the TCU). Having the heater section where fluid flow is present near the outlet of the pump, it will allow higher watt density heaters to be used. The critical breakdown is ~ 14 W/cm2. We recommend 4 W/cm2. The decision to go beyond this level should be made by the customer.

I would not use a insertion heating element due to the low surface area. That heater will give you 9 W/cm2, which is fairly high.
Like you mentioned, it's also at the edge of the tank, so heating will be fairly localized.

My recommendation is to heat the entire bottom of the tank. Get a bunch of mica heating strips so that the bottom of the tank has a watt density of 4-5W/cm2 (remember steel is fairly poor thermal conductor). Epoxy them to the bottom of the tank with a heat-resistant epoxy. Same with aluminum water cooling heatsinks. Mount the container with a flexible high-temperature adhesive (e.g. Loctite 596).

Unless you have the money to spend, I strongly recommend sourcing the elements and heatsinks from China. The prices I saw for those at Western sites are absurd (>200EUR for a heatsink etc)

Thanks for the hint with the heating density! Just to be clear, the mica heating strips you are tlking about, are they "solid" ie encapsulated in some metal like these: https://au.rs-online.com/web/p/general-heating-elements/9210035/
or is it something else ? If not, I would be worried that there will be problems if these are bonded to the tank with epoxy and the tank bulges under heat (they might slip/crack) ?
I wonder what heating method IMDES is using on their/his tanks ?