Page 2 of 3

Re: Question About Design Trade-offs for Backing Pumps

Posted: Thu Jan 11, 2018 3:58 am
by Jerry Biehler
10^-6 is doable but you are going to get some pretty good backstreaming without at least a water cooled baffle. That's why i am converting all my DP based systems to turbos. I dont want to deal with backstreaming.

I had one system with a 18" x 18" bell jar on it and a 6" diff pump on it and left it running for a couple days and had a nice film coating the inside of the jar. On the JEOL FESEMs I have worked on they have a bypass line to pump the ion pumps down to a low enough pressure to get them to start. On the DP pumped systems you can only leave this bypass open for about 20 minutes before you have to worry about DP oil contaminating the column from back streaming.

You will probably want to make sure the ion pump is not in line of sight with the DP or it could get contaminated by oil. You could just use an elbow or something.

Re: Question About Design Trade-offs for Backing Pumps

Posted: Thu Jan 11, 2018 8:02 pm
by Michael Bretti
Jerry Biehler,

Thank you for your input on your experience with diffusion pumps. Reading more into it, while the oils themselves may be rated for much lower torr, I failed to account for the fact the diffusion pumps can experience considerable back-streaming during pump-down, especially when the foreline pressure is at a certain level, mainly at the level that any backing pump would be able to achieve that I could afford at this point. Right now I wouldn't be able to afford a pre-built water cooled or refrigerated baffle for my system, and at this point I'd rather not try to integrate it into my current design with a homemade approach going from atmosphere to vacuum with water line connections and risk introducing massive leaks due to poor construction. I do have another idea though that may help my problem.

Currently, the ion pump is directly in the line of sight of the diffusion pump, since the pipeline and chamber is completely straight through and very short. This arrangement has allowed me to double my molecular flow conductance from my initial design which didn't have a straight line of sight to the top of the chamber. However contamination would be an issue without a baffle as you mentioned. I do have an idea though that would cost nothing to implement and probably be helpful in reducing this back-streaming without effecting my conductance with the system as is. I am thinking about making a very simple baffle mounted underneath my adapter plate above the diffusion pump inlet. This way I could suspend the baffle over the pump without having it in direct contact with any potential hot parts (such as if I were to rest it on top of the diff pump stack.) Since the throat of the diff pump is getting choked effectively from 5" down to 1.5", this should also help reduce back streaming somewhat as well. Eventually what I could even do is integrate cooling on the adapter plate itself (for example, having an array of peltier coolers mounted on the plate around the adapter inlet) so that the entire plate could act effectively as a cooled plane. If I design the baffle correctly with enough thermal transfer from the plate, I could also use this to help indirectly cool the baffle as well. I could very easily run thermal simulations with my model in fusion360 to verify if this would work. Whenever I get around to this I will post pictures of my design concept and the thermal modelling results. The good part about this approach (besides its low cost and ease of fabrication) is that it won't really affect my final conductance and speeds. Even if the baffle reduces the main pump speed by a significant amount, since the pumping speed is still so much greater than the molecular flow conductance, the resulting effective speed would essentially stay the same.

Also, doing some very preliminary calculations for system pump-down, I can definitely get a feel now for how much of a struggle it will be to cross past the 10^-7 torr range. Assuming no back-streaming or leaks, the out-gassing load alone is enough from preventing me from reaching 10^-7 torr, even though my system is quite small, under normal pumping. It appears that I really will only be able to achieve this baking the system and pumping for periods longer than 24 hours. This is obviously common sense for high vacuum systems, but it is nice to have some numbers now to back this up, and specifically for my system as well. I'm also in the process of figuring out my pump-down times from atmosphere to high vacuum as well, so hopefully it will give me some rough orders of magnitude what to expect for the various stages of pumping.

Re: Question About Design Trade-offs for Backing Pumps

Posted: Fri Jan 12, 2018 2:34 am
by Michael Bretti
So I went back to my diffusion pump just to check out where the stack ends at the top to try and gauge how I could implement my baffle idea. It turns out the stack is secured at the top with a cap that has two adjustment screws to clamp it to the wall:

20180111_211314.jpg

20180111_211333.jpg

Looking at it from the side and taking a quick measurement, it appears that the cap wider than the inlet hole of the adapter plate, which sits directly on top of the diffusion pump, and is a zero-clearance transition from 5" to 1.5". The cap is a bit bigger, about 1.6875" wide, and would sit only 0.25" from the entrance of hole. In effect, is it possible for the entire combination of the larger diameter cap sitting so close to the smaller diameter hole in a massively reduced zero-clearance throat create kind of a pseudo "self-baffle"? If the inlet to the chamber was much bigger, on the order of several inches or even closer to the throat diameter of 5" of the diffusion pump, then I would most definitely need some kind of baffle, since there is a very clear and unobstructed line of sight. However, in this case, its not as open or direct. If not, I could probably cut out some disks and stack them on the top cap clamp in a way to effectively make an optically dense baffle, though I am not sure if it would still be effective since it is not water cooled.

Re: Question About Design Trade-offs for Backing Pumps

Posted: Fri Jan 12, 2018 3:13 am
by Jerry Biehler
What the dimensions of the inlet on the diff pump? Outer o-ring diameter? I might be able to scrounge up a water cooled baffle that will fit that pump.

Re: Question About Design Trade-offs for Backing Pumps

Posted: Fri Jan 12, 2018 4:10 am
by Michael Bretti
Jerry Biehler,

Thank you a ton for your response and help looking into this, I greatly appreciate your efforts, as well as everyone's help in their responses so far. I made up a quick drawing sheet of the top flange of the diffusion pump from my model:

Edwards EO4 Diffusion Pump Drawing - TOP.pdf
(167.02 KiB) Downloaded 41 times

To summerize the dimensions:
Inlet Diameter: 5"
Flange Outer Diameter: 9"
O-ring Inner Diameter: 5.25"
O-ring Outer Diameter: 5.75"
Mounting Holes Diameter: 0.75"
Radius of Mounting Holes to Center: 3.75"
Number of Mounting Holes: 8

Re: Question About Design Trade-offs for Backing Pumps

Posted: Fri Jan 12, 2018 8:18 am
by Jerry Biehler
OK, let me look around. I think I know of one you though you might have to make an adapter plate. it just clamps between flanges.

Re: Question About Design Trade-offs for Backing Pumps

Posted: Sat Jan 13, 2018 10:31 am
by Jerry Biehler
Here's what I found, You will have to make a plate plate that goes between this and the DP since both have o-ring grooves.

https://flic.kr/s/aHsmckNkwf

It's $20 plus shipping.

Re: Question About Design Trade-offs for Backing Pumps

Posted: Sat Jan 13, 2018 5:17 pm
by Michael Bretti
Jerry Biehler,

Thank you so much for your help on this, I truly appreciate it. I have just been CADing up some quick models based on your pictures and dimensions to see how I could fit the baffle to my current system with a simple ring adapter, but I think it could work. The overall dimensions should also let me use my original plate adapter on the top, and allow for me to clamp down the baffle and ring adapter for the bottom of it to the diffusion pump all together with just some longer bolts. Does the orientation of the baffle matter? I think the simplest way to adapt it to my current design would be to mount it so that the o-ring groove on the baffle is facing up. This would seal against the bottom of my original plate adapter which would go to the 2.75" conflat hardware on top. The ring adapter I would need machined would have an o-ring groove on the top face, so it seals against the flat un-grooved plane of the baffle. I think this would be the simplest and cheapest way for me to put this whole system together with the baffle. I have attached a PDF below with a quick cross-sectional view of how the whole system would stack together. You can see all of the major plates and components, with the appropriate o-ring grooves. Let me know what you think:


If you think that my above arrangement is acceptable and would work, then I think I would definitely be interested in purchasing this baffle from you. Should I contact you through your gmail so we can arrange the details?

Re: Question About Design Trade-offs for Backing Pumps

Posted: Sun Jan 14, 2018 6:30 am
by Jerry Biehler
Yeah, send me a message and we will get this thing going.

Re: Question About Design Trade-offs for Backing Pumps

Posted: Sun Jan 14, 2018 6:31 am
by Jerry Biehler
Oh, and yes, your adapter should work fine.