Diffusion Pumps 101
-
- Posts: 262
- Joined: Tue Aug 15, 2006 8:51 pm
- Real name: George Schmermund
- Location: Carlsbad, CA
Diffusion Pumps 101
Well, Jon R has thrown down the gauntlet and I suppose that I've got no choice at this point but to try to pull the rabbit out of the hat. Yesterday I went shopping at Home Depot, Target, and the hobby shop for parts. I suppose that all the stuff is obtainable from Amazon.com, too, but I like to look at the items in real life if I can. Rotating the parts in my hands makes it easy to figure out how well they'll fit together.
My mission will be to build a small, high performance, metal diffusion pump that requires only simple tools, easily acquired parts, and takes no more than a day to assemble. Easy construction doesn't account for much if the parts aren't readily available.
Also, the glass pump projects that I've posted about in the past aren't in the running because they're inaccessible to the average Fusor builder due to the special talent needed to do the glassblowing and the high cost.
The most difficult part about building a diffusion pump is understanding how they work. I mean REALLY work. There are no moving parts in a DP except for the high velocity motive vapor which is working in a very low pressure environment. There are areas of pumping operation that are very unintuitive even to those who would claim to know everything about how they work. I've had experienced vacuum system builders check out one of my glass pumps in operation and they confessed to being stunned at what they were able to observe. It changed their whole mental model of DP's.
So, rather than belabor you guys with another one of my protracted tales, I'll try to march a straight course and just show some pictures of the parts and how simple it is to put them together. I'll start with what I've got on hand now and progress to the finished device in a later post. The parts should be self explanatory. The total out of pocket cost sofar is less than $60.
For scale, the cocktail shaker is 9" tall
[attachment=0]8814.DSCF1606.JPG[/attachment]
My mission will be to build a small, high performance, metal diffusion pump that requires only simple tools, easily acquired parts, and takes no more than a day to assemble. Easy construction doesn't account for much if the parts aren't readily available.
Also, the glass pump projects that I've posted about in the past aren't in the running because they're inaccessible to the average Fusor builder due to the special talent needed to do the glassblowing and the high cost.
The most difficult part about building a diffusion pump is understanding how they work. I mean REALLY work. There are no moving parts in a DP except for the high velocity motive vapor which is working in a very low pressure environment. There are areas of pumping operation that are very unintuitive even to those who would claim to know everything about how they work. I've had experienced vacuum system builders check out one of my glass pumps in operation and they confessed to being stunned at what they were able to observe. It changed their whole mental model of DP's.
So, rather than belabor you guys with another one of my protracted tales, I'll try to march a straight course and just show some pictures of the parts and how simple it is to put them together. I'll start with what I've got on hand now and progress to the finished device in a later post. The parts should be self explanatory. The total out of pocket cost sofar is less than $60.
For scale, the cocktail shaker is 9" tall
[attachment=0]8814.DSCF1606.JPG[/attachment]
Anything obvious in high vacuum is probably wrong.
-
- Posts: 9
- Joined: Wed Jan 05, 2011 6:43 am
- Real name:
Re: Diffusion Pumps 101
Hi George,
Have you made a glass diffusion pump before? I'd be very interested to hear your results and experience as I'd like to go the glass direction myself.
I also appreciate you posting this thread. I will be following it closely. Thanks again.
-James
Have you made a glass diffusion pump before? I'd be very interested to hear your results and experience as I'd like to go the glass direction myself.
I also appreciate you posting this thread. I will be following it closely. Thanks again.
-James
-
- Posts: 9
- Joined: Wed Jan 05, 2011 6:43 am
- Real name:
Re: Diffusion Pumps 101
Hi again,
I'd actually be curious of where you're located and if you'd be up to collaborate on a glass diffusion project? Feel free to contact me anytime.
-James
I'd actually be curious of where you're located and if you'd be up to collaborate on a glass diffusion project? Feel free to contact me anytime.
-James
Re: Diffusion Pumps 101
I will be very interested to see this built. From my own disassembly of a diffusion pump, all the parts make sense.
I did a quick search on youtube for any videos of glass diffusion pumps in operation and did not see any. It would be very cool if you could post a video of one of your glass pumps in operation.
I did a quick search on youtube for any videos of glass diffusion pumps in operation and did not see any. It would be very cool if you could post a video of one of your glass pumps in operation.
- Richard Hull
- Moderator
- Posts: 14992
- Joined: Fri Jun 15, 2001 9:44 am
- Real name: Richard Hull
Re: Diffusion Pumps 101
All the best of luck on the DIY diff pump. Sounds like both fun and quite a challenge.
I bought my current 2.5" throat Veeco diff pump currently used on Fusor IV at my HEAS gathering fleamarket about 10 years ago from an old friend selling out of his trunk for $30.00! It is air cooled with high CFM blower. Doesn't get much cheaper than that.
FYI for the newbs..............
If you are building a fusor and the chamber is less than 10 liters, (most always are), there is little need for a diff pump having much over a 3" throat. Air cooled is always the best if you can find it.
Used, air cooled diff pumps of this size command a slight premium due to their being among the most desirable of all diff pumps. Expect to pay about $100.00 on e-bay or the internet as a reasonable sum.
You may, however, locate a 6" throated, water cooled diff pump for free or even $10.00, but remember the adapter might cost $100.00 and then there is that flowing water plumbing.
No need paying $20.00 for an enitre, functional 8" throat diff pump, then, $250.00 for the adapter network, $70.00 for water plumbing hookups and $75.00 for the diff pump oil to mechanize such a big monster, drawing down a 3 liter chamber.
caveat emptor!
Richard Hull
I bought my current 2.5" throat Veeco diff pump currently used on Fusor IV at my HEAS gathering fleamarket about 10 years ago from an old friend selling out of his trunk for $30.00! It is air cooled with high CFM blower. Doesn't get much cheaper than that.
FYI for the newbs..............
If you are building a fusor and the chamber is less than 10 liters, (most always are), there is little need for a diff pump having much over a 3" throat. Air cooled is always the best if you can find it.
Used, air cooled diff pumps of this size command a slight premium due to their being among the most desirable of all diff pumps. Expect to pay about $100.00 on e-bay or the internet as a reasonable sum.
You may, however, locate a 6" throated, water cooled diff pump for free or even $10.00, but remember the adapter might cost $100.00 and then there is that flowing water plumbing.
No need paying $20.00 for an enitre, functional 8" throat diff pump, then, $250.00 for the adapter network, $70.00 for water plumbing hookups and $75.00 for the diff pump oil to mechanize such a big monster, drawing down a 3 liter chamber.
caveat emptor!
Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
- Carl Willis
- Posts: 2841
- Joined: Thu Jul 26, 2001 7:33 pm
- Real name: Carl Willis
- Location: Albuquerque, New Mexico, USA
- Contact:
Re: Diffusion Pumps 101
Hi George,
Some creative shopping there.
I wonder if Mike Kan, our memorable vacuum-gear Marxist (go read his old posts for a laugh, and I mean that in a good-natured way), will be around to witness your effort's success.
Are you planning to go the solder or brazing route with all this stuff?
-Carl
Some creative shopping there.
I wonder if Mike Kan, our memorable vacuum-gear Marxist (go read his old posts for a laugh, and I mean that in a good-natured way), will be around to witness your effort's success.
Are you planning to go the solder or brazing route with all this stuff?
-Carl
-
- Posts: 262
- Joined: Tue Aug 15, 2006 8:51 pm
- Real name: George Schmermund
- Location: Carlsbad, CA
Re: Diffusion Pumps 101
Carl - The plan so far is to braze the foreline joints and epoxy the top colder joints. This will be another chance to play with some JB Weld.
The jet assembly will be compressed together using the threaded rod and nuts & washers to form a single unit.
The top most piece will be the KF 40 half nipple which allows direct connection to standard manifold hardware.
A small string heater of a couple hundred watts will heat the boiler.
The cooling will be done with a miniature muffin fan attached to a simple cowling made from aluminum flashing.
The jet assembly will be compressed together using the threaded rod and nuts & washers to form a single unit.
The top most piece will be the KF 40 half nipple which allows direct connection to standard manifold hardware.
A small string heater of a couple hundred watts will heat the boiler.
The cooling will be done with a miniature muffin fan attached to a simple cowling made from aluminum flashing.
Anything obvious in high vacuum is probably wrong.
Re: Diffusion Pumps 101
I, for one, have no doubt that George's effort will be successful.
And the liberating effect of the DIY effort..... is reward enough.
Thanks for picking up the challenge, George...
Dave Cooper
And the liberating effect of the DIY effort..... is reward enough.
Thanks for picking up the challenge, George...
Dave Cooper
- John Taylor
- Posts: 97
- Joined: Wed Jun 22, 2011 1:43 pm
- Real name:
- Location: Dardanelle, Arkansas
Re: Diffusion Pumps 101
George, I have a new 4" muffin fan with connector and matching guard I'll give you if you promise to keep us informed of the details. I, for one, would rather build my own equipment than buy it if at all possible. This seems like it would be a rather fun project.
-
- Posts: 262
- Joined: Tue Aug 15, 2006 8:51 pm
- Real name: George Schmermund
- Location: Carlsbad, CA
Re: Diffusion Pumps 101: Jet Assembly
I had some time this week to play with the DP project and decided to do some brazing. The setup and results are in the photos. The tools needed for the project to this point are: tin snips, hacksaw, file, sand paper, Scotch-Brite, torch, beer.
The brazing alloy of choice was Harris Safety-Silv 45 in a 1/16" wire and Stay-Silv paste flux. This will give a good match for the copper to stainless joints and has the advantage of a reasonably low melting point. The only torch needed is a propane / air type which should be easy for all to use. The parts are small and light weight and won't over heat with this approach. Some practice on scrap pieces is recommended.
For prep and post braze clean-up, I have not found anything easier to work with and more effective than Lysol Power Toilet Bowl Cleaner and Scotch-Brite.
The pictures should be self explanatory as to how the jet assembly goes together. What's pictured is a fit test and the final alignment will be made when the assembly is compressed with the threaded rod.
More to follow.
[attachment=2]8838.DSCF1624.jpg[/attachment]
[attachment=1]8840.DSCF1629.jpg[/attachment]
[attachment=0]8839.DSCF1636.jpg[/attachment]
The brazing alloy of choice was Harris Safety-Silv 45 in a 1/16" wire and Stay-Silv paste flux. This will give a good match for the copper to stainless joints and has the advantage of a reasonably low melting point. The only torch needed is a propane / air type which should be easy for all to use. The parts are small and light weight and won't over heat with this approach. Some practice on scrap pieces is recommended.
For prep and post braze clean-up, I have not found anything easier to work with and more effective than Lysol Power Toilet Bowl Cleaner and Scotch-Brite.
The pictures should be self explanatory as to how the jet assembly goes together. What's pictured is a fit test and the final alignment will be made when the assembly is compressed with the threaded rod.
More to follow.
[attachment=2]8838.DSCF1624.jpg[/attachment]
[attachment=1]8840.DSCF1629.jpg[/attachment]
[attachment=0]8839.DSCF1636.jpg[/attachment]
Anything obvious in high vacuum is probably wrong.
-
- Posts: 107
- Joined: Tue Apr 28, 2009 10:45 pm
- Real name: Chad Ramey
- Location: Georgia
- Contact:
Re: Diffusion Pumps 101: Jet Assembly
Wow....just simply fantastic. The craftsmanship exemplified thus far is astonishing! I look forward to seeing how it works.
-Chad
-Chad
-
- Posts: 262
- Joined: Tue Aug 15, 2006 8:51 pm
- Real name: George Schmermund
- Location: Carlsbad, CA
Re: Diffusion Pumps 101
The boiler section of the pump has been the easiest part so far. It took about 5 minutes to setup and braze a 1/4" SS bolt to the bottom of the pump body and another 10 minutes to cut up a vegetable can from the pantry and then put some holes and grommets in it. The trimmed can was used like a cookie cutter to make the Fiberfrax insulation pieces. The rope heater was wrapped flat against the bottom of the pump body and compressed into place by the lid from the can, which also had a hole in the center. This formed a sort of spool arrangement to keep the heater coils tight and flat. The leads were then fed through the grommets and the whole boiler assembly tightened up with the bottom nut.
More to come.
[attachment=1]8842.DSCF1642.jpg[/attachment]
[attachment=0]8841.DSCF1641.jpg[/attachment]
More to come.
[attachment=1]8842.DSCF1642.jpg[/attachment]
[attachment=0]8841.DSCF1641.jpg[/attachment]
Anything obvious in high vacuum is probably wrong.
-
- Posts: 307
- Joined: Sat Aug 27, 2005 8:30 am
- Real name: rich normand
Re: Diffusion Pumps 101
Very, very nice George!
Can’t wait to see how well this will work out.
Most of the stuff seems to be stainless steel and such. Is the tin can part of the heater a possible problem? They (tin cans) tend to have coatings inside (plastics or flash metals) and soft welds at the seams that could be an issue even if it is not in the vacuum part of the pump? I assume your temp will be low so that is not such an issue.
Still, the eye candy value alone has me glued to the photos.
Looking at the inventiveness, particularly using common utensils, is in the best traditions of Rutherford's work and "Building scientific apparatus" first edition....
Can’t wait to see how well this will work out.
Most of the stuff seems to be stainless steel and such. Is the tin can part of the heater a possible problem? They (tin cans) tend to have coatings inside (plastics or flash metals) and soft welds at the seams that could be an issue even if it is not in the vacuum part of the pump? I assume your temp will be low so that is not such an issue.
Still, the eye candy value alone has me glued to the photos.
Looking at the inventiveness, particularly using common utensils, is in the best traditions of Rutherford's work and "Building scientific apparatus" first edition....
- Carl Willis
- Posts: 2841
- Joined: Thu Jul 26, 2001 7:33 pm
- Real name: Carl Willis
- Location: Albuquerque, New Mexico, USA
- Contact:
Re: Diffusion Pumps 101: Jet Assembly
Modern art with "found objects"! Nice work, George.
I can't wait to see the results of your trials. What will be your pumping fluid? Extra-virgin olive oil?
-Carl
I can't wait to see the results of your trials. What will be your pumping fluid? Extra-virgin olive oil?
-Carl
-
- Posts: 262
- Joined: Tue Aug 15, 2006 8:51 pm
- Real name: George Schmermund
- Location: Carlsbad, CA
Re: Diffusion Pumps 101
Thanks for all the encouragement, guys! If I get some time today I'll finish up the foreline and braze it into the pump body. The final effort will be to shape and install the cowling and cooling fan(s).
As far as the can used for the boiler heat shield goes, there is no seam, but there is a very thin coating on the inside surface. I'll carbonize the coating with the torch and then clean up the can with some Scotch-Brite and Lysol PTBC. I'll also probably add a thermocouple to the bottom of the boiler. This will allow me to run some temperature/ pressure curves in the future.
I do remember something about distilling olive oil and using it in a desperate act as a diffusion pump fluid. I think it was in The Amateur Scientist column back in about 1960 or so. Anyway, I'll stick to DC 704 for this project.
As far as the can used for the boiler heat shield goes, there is no seam, but there is a very thin coating on the inside surface. I'll carbonize the coating with the torch and then clean up the can with some Scotch-Brite and Lysol PTBC. I'll also probably add a thermocouple to the bottom of the boiler. This will allow me to run some temperature/ pressure curves in the future.
I do remember something about distilling olive oil and using it in a desperate act as a diffusion pump fluid. I think it was in The Amateur Scientist column back in about 1960 or so. Anyway, I'll stick to DC 704 for this project.
Anything obvious in high vacuum is probably wrong.
- Richard Hull
- Moderator
- Posts: 14992
- Joined: Fri Jun 15, 2001 9:44 am
- Real name: Richard Hull
Re: Diffusion Pumps 101
Looks like it ought to work. You have poured the work into it. If this works out it will offer a cheap way for the DIYs to make a small air cooled pump more suited to a small fusor rather than have to pay the premium often seen on these small manufactured units.
Richard Hull
Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
- John Taylor
- Posts: 97
- Joined: Wed Jun 22, 2011 1:43 pm
- Real name:
- Location: Dardanelle, Arkansas
Re: Diffusion Pumps 101
Very nice workmanship! I can't wait to hear how well it works, I would like to emulate this as it looks like a great project.
-
- Posts: 262
- Joined: Tue Aug 15, 2006 8:51 pm
- Real name: George Schmermund
- Location: Carlsbad, CA
Re: Diffusion Pumps 101
The pump body is basically finished at this point. Next is the cooling system. I've been thinking about different configurations and they all seem to have their own merits. I'll put a few thermocouples at the points where I'm interested in seeing the effects of the different methods.
I've added another pic of the boiler so that it's clear how the rope heater is installed and kept under control. It's important that the heater remain in intimate contact with the bottom of the boiler.
After torching the can to red heat to remove the plastic film on the inside and then Scotch-Brite scrubbing it clean, I decided to retorch it to get that 'gun metal blue' look. This is my concession to steampunking and Carl's comment about "art".
[attachment=1]8853.DSCF1650.jpg[/attachment]
[attachment=0]8852.DSCF1648.jpg[/attachment]
I've added another pic of the boiler so that it's clear how the rope heater is installed and kept under control. It's important that the heater remain in intimate contact with the bottom of the boiler.
After torching the can to red heat to remove the plastic film on the inside and then Scotch-Brite scrubbing it clean, I decided to retorch it to get that 'gun metal blue' look. This is my concession to steampunking and Carl's comment about "art".
[attachment=1]8853.DSCF1650.jpg[/attachment]
[attachment=0]8852.DSCF1648.jpg[/attachment]
Anything obvious in high vacuum is probably wrong.
-
- Posts: 307
- Joined: Sat Aug 27, 2005 8:30 am
- Real name: rich normand
Re: Diffusion Pumps 101
Would one or two computer CPU fans with a metal shroud following the curve of the main body be adequate? If not a ThermalTake liquid CPU cooler might do, with a small copper tube spiral on the outside.. It comes with its own pump, radiator liquid-to-air heat exchanger and all the trimmings. I used them for cooled pmts, peltier cooled ccd detectors and ccd camera. The old models are available on ebay for cheap and can displace lots of heat in a small turn-key package.
Re: Diffusion Pumps 101
George - beautiful!
You combine, art, science, practical engineering and a great sense of humor. Impressive work.
(Lots of real talent amongst those on this board.)
Dave Cooper
You combine, art, science, practical engineering and a great sense of humor. Impressive work.
(Lots of real talent amongst those on this board.)
Dave Cooper
- Richard Hull
- Moderator
- Posts: 14992
- Joined: Fri Jun 15, 2001 9:44 am
- Real name: Richard Hull
Re: Diffusion Pumps 101
Traditionally, All air cooled pumps have a series of 6 to 10, 1/8" or thicker pure copper rings of at least 1" width brazed to the stack's exterior and a special shroud that funnels air from a very high CFM blower through the plenum. The air cooled pumps operation and ultimate depth of vacuum demands that the oil vapor be condensed before it can back stream.
Stainless steel is not a very good conductor of heat,but with a thin wall and good copper fins brazed to on, it should do fine.
There is a reason you don't see diff pumps with over a 3" throat that are air cooled. Naturally, Santovac or Pentovac fluids would help, but they are about $100.00 for 50cc.
Certainly, any suitably small DIY diff pump should be air cooled.
Richard Hull
Stainless steel is not a very good conductor of heat,but with a thin wall and good copper fins brazed to on, it should do fine.
There is a reason you don't see diff pumps with over a 3" throat that are air cooled. Naturally, Santovac or Pentovac fluids would help, but they are about $100.00 for 50cc.
Certainly, any suitably small DIY diff pump should be air cooled.
Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
-
- Posts: 262
- Joined: Tue Aug 15, 2006 8:51 pm
- Real name: George Schmermund
- Location: Carlsbad, CA
Re: Diffusion Pumps 101
Rich N - Thanks for the tip on the Thermaltake stuff. I was thinking about building some heat pipe devices to play with and then found a large selection of them on eBay. I bought a couple of new 125 watt units for $8 each. Plug and play. I couldn't even start the project for that price.
I'm also looking at Peltier coolers as a possibility. They're real cheap on eBay. Lots of methods to choose from.
I'm also looking at Peltier coolers as a possibility. They're real cheap on eBay. Lots of methods to choose from.
Anything obvious in high vacuum is probably wrong.
-
- Posts: 262
- Joined: Tue Aug 15, 2006 8:51 pm
- Real name: George Schmermund
- Location: Carlsbad, CA
Re: Diffusion Pumps 101
After considering all of the alternatives for cooling the pump, the best strategy seems to be the method that I used on the glass pumps. This would be a hybrid water/air arrangement as pointed out by Rich N. It could obviously be done with air alone, but the amount of work needed to make good fins and fit and braze them to a tapered body would defeat the propose of making a simple pump. The other downsides to just air are excessive noise for above bench-top level work and the heat transfer in close proximity to the experiment area. Also, it's a lot of wind where you generally don't want it.
Being the cheapskate that I am, I'll use a Harbor Fright fountain pump and a container of water. The water will circulate through the 1/4" copper tubing that will be brazed onto the pump body and then through a simple mini radiator and back to the reservoir. This will allow placing the water and fan at any reasonable distance from the work/experiment area.
The pump is now basically finished and the next thread will cover how to setup and test the performance of this Bad Boy!
Being the cheapskate that I am, I'll use a Harbor Fright fountain pump and a container of water. The water will circulate through the 1/4" copper tubing that will be brazed onto the pump body and then through a simple mini radiator and back to the reservoir. This will allow placing the water and fan at any reasonable distance from the work/experiment area.
The pump is now basically finished and the next thread will cover how to setup and test the performance of this Bad Boy!
Anything obvious in high vacuum is probably wrong.
-
- Posts: 2
- Joined: Sun Jul 24, 2011 10:18 am
- Real name:
Re: Diffusion Pumps 101
If the jet stack is mostly made of one pipe cut into sections, what's going to keep the forepump from sucking all the vapor out of the lower jet before it can successfully reach the top jet?
Chase
Chase
-
- Posts: 148
- Joined: Thu May 12, 2011 11:25 pm
- Real name: Bruce Meagher
- Location: San Diego
Re: Diffusion Pumps 101
Hi Chase,
I believe to answer your question you need to understand the kinetic theory of gases, laws of flow of gases (particularly through small capillaries), and the accommodation coefficient of the oil vapor. I've been reading Vacuum Techniques by Saul Dushman and he has a very detailed discussion of the theory and operation of many vacuum pumping techniques (including diffusion pumps).
One of the key fundamental concepts that allows a diffusion pump to work is the fact that the oil vapor condenses when it hits a metal surface, and as long as the temperature of the metal surface is low the oil will not re-evaporate. Once condensed, with the help of gravity, the oil will return to the boiler. So long as the oil vapor touches a cold surface (cooled pump wall, foreline wall, foreline baffle) no oil vapor will leave through the roughing pump. Of course this same concept is what makes a diffusion pump failure cause such a mess. If the roughing pump fails or the vacuum vessel is opened to atmosphere then the oil vapor molecules coming out of the jets collide with many molecules causing them to lose their momentum towards the wall. Once their momentum towards the wall is lost the direction becomes random. Due to the small mean free path because of the failure (or maybe more correctly because of the higher pressure), the oil vapor molecules make there way out everywhere until they eventually come in contact with a wall causing an oily disaster.
I'm a newbie here, haven't fired up my pumping station yet, but this is my current understanding from my reading.
Bruce
I believe to answer your question you need to understand the kinetic theory of gases, laws of flow of gases (particularly through small capillaries), and the accommodation coefficient of the oil vapor. I've been reading Vacuum Techniques by Saul Dushman and he has a very detailed discussion of the theory and operation of many vacuum pumping techniques (including diffusion pumps).
One of the key fundamental concepts that allows a diffusion pump to work is the fact that the oil vapor condenses when it hits a metal surface, and as long as the temperature of the metal surface is low the oil will not re-evaporate. Once condensed, with the help of gravity, the oil will return to the boiler. So long as the oil vapor touches a cold surface (cooled pump wall, foreline wall, foreline baffle) no oil vapor will leave through the roughing pump. Of course this same concept is what makes a diffusion pump failure cause such a mess. If the roughing pump fails or the vacuum vessel is opened to atmosphere then the oil vapor molecules coming out of the jets collide with many molecules causing them to lose their momentum towards the wall. Once their momentum towards the wall is lost the direction becomes random. Due to the small mean free path because of the failure (or maybe more correctly because of the higher pressure), the oil vapor molecules make there way out everywhere until they eventually come in contact with a wall causing an oily disaster.
I'm a newbie here, haven't fired up my pumping station yet, but this is my current understanding from my reading.
Bruce