High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Every fusor and fusion system seems to need a vacuum. This area is for detailed discussion of vacuum systems, materials, gauging, etc. related to fusor or fusion research.
User avatar
Dennis P Brown
Posts: 3159
Joined: Sun May 20, 2012 10:46 am
Real name: Dennis Brown

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Dennis P Brown »

When you say:
Additional high and ultra-high vacuum pumps can also be employed, such as ion pumps and titanium sublimation pumps with relative ease and low cost
These are very expensive items new, and not very easy to obtained used; and more often, not in the best condition when used (especially the electronics.) Also, adapters for these components and extra ports can also lead to not insignificant costs. So, might want to modify that part of the post. Newbie's here do not know cost issues very well and I assume you are posting here for all levels of readers.

While your maximum deuterium gas flow for a given system is interesting, you do realize that isn't what most people that build fusors are concerned about at all. Rather, since deuterium gas is difficult to obtain and not inexpensive or if created from heavy water, both time intensive to make and even more expensive, one wants to conserve this gas. Hence, most people are only concerned with using as little as possible. As such, maybe in the future determining minimum flow rates that provide best neutron rates for a fusor would be a more valuable project, I'd think. Your discussions on water vapor issues is a topic that could use a lot more discussion since this is a bane to fusion for fusors. Maybe getting a rough idea of what does adhere for a standard humid atmosphere and typical volume system/chamber with pump down rates (not baked) to get this below an acceptable level would certainly be of use here.

Your calculations and work are rigorous and interesting and frankly, one of the few examples I've seen on this subject. Looking at this subject specifically for fusors usage is relevant (which you do), so focusing on that a bit more would make your articles even more useful. Not that high vacuum is not important since some here do require that for ion accelerator fusion.
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

Dennis P. Brown,

Thank you very much for your suggestions and insight. It has given me a lot of new ideas to think about that I hadn't originally planned to dive into, but you make very good points.

When I was typing out the last section, I knew I would probably have to specify more about the ion and titanium sublimation pumps. Thank you for bringing that up, I will go back and revise that last statement a bit. They are definitely very expensive pieces of equipment new. I don't even think I have come across titanium sublimation pumps used on eBay. However, in the past couple months of eBay browsing, I have come across a decent number of used ion pumps that appear to be in quite good shape physically. The nice thing is that a lot of pumps sized for 20-25 L/s are fitted with 2.75" Conflat flange hardware, which makes it very convenient to integrate into these smaller systems, and at ultra-high vacuum levels even 20 L/s is a huge pumping boost. These pumps that I have come across so far are in the $100-$250 range, which is quite reasonably priced for ultra-high vacuum pumps. Of course functionality is another matter, but generally ion pumps don't require much maintenance, especially if they are used properly. Old controllers are also not terribly expensive, and if someone can get a standard fusor working with proper instrumentation, then building an ion pump supply should not be too difficult or expensive as well.

I am quite interested in the titanium sublimation pump however because it would be very easy to construct a simple one with little effort and components, and the controller for such would be dirt simple (low voltage high current supply cycled on and off a few minutes a few times an hour). I found this guide that actually shows how to make a simple one based off of 2.75" conflat hardware that I am going to follow and modify myself:

https://www.rbdinstruments.com/blog/hom ... tion-pump/

All it would consist of is a nipple section, an insulated low voltage feedthrough with at least 2 feedthroughs, and some titanium wire. For my own system, I was able to obtain a 2.75" nipple section and an insulated low voltage feedthrough with 4 inputs for free, so all that is needed is the titanium wire and a simple controller. Interestingly, it is very easy to calculate the pumping speeds for both cryopumps and sorbtion pumps, but I have not seen much specifically on titanium sublimation pumping. I am working on a CAD model of the simple design, which I will post about, along with thermal modelling later. I would like to explore and implement a small one for my system to see how well it works.

In regards to the rest of the calculations, all of this work was originally and primarily done for myself for my own system. However, I figured that since I am doing all this work, others might be able to learn and benefit from the process as well, since a lot of people might not have access to a lot of the high vacuum engineering texts that I can obtain very easily, and I have found these texts have vastly more knowledge and insight than what I have found available online. I do try to tie this all back to fusor related operation as best I can, although the fusor is only one operating mode for my system, working at the highest pressures calculated.

It was motivated by two major driving points: 1.) since I will be spending a lot of time and money on this system, I had better make sure that it can do near everything that I want, with room for growth, and I wanted to estimate how reasonable my goals would be for this system, and 2.) how do I know that my system can support the process I aim to achieve? Can it actually support everything at high vacuum with reasonable gas flows? The answer to the second question is currently it can support all my goals except for one. I wanted to experiment with argon-fueled micro electric space propulsion engines at 10^-7 Torr. This was the most stringent limiting factor on my system, which helped push the optimization of speed for my system. Currently, as is it won't be able to support this project, but everything else for now can be run.

Another thing I was interested in, relating to the fusor side, is how much gas flow can my system support? Browsing through prior posts, it appears that the common sccm flow rate for deuterium for fusors operating in the 10^-2 Torr level use anywhere between a few sccm to a couple of tens of sccm of deuterium. How do I really know that my system can support this? I didn't want to spend all of this money and time to build a system to find out it couldn't do what I wanted. Based on all of the above math, I believe I have successfully shown that it is reasonable to support up to several tens of sccm of deuterium in the system at 10^-2 Torr. It also allows me to get a feel for how much gas flow I can support for deuterium beam systems in the high vacuum region, which is even more stringent and challenging. While the system can handle a lot of deuterium flow at low vacuum, it does not mean I will run the system at full flow - the manual gate valve will allow me to change conductance if needed, and I would put much less into the system anyway since it is very expensive. However, it is best to plan for extra room than not enough. Another unseen positive taken from these calculations is with the above calculations for max flow rate for air - if I have a leak that is preventing me from reaching some vacuum level, I at least roughly know how much load from air due to the leak is acting on the system, which could be very useful for troubleshooting and experimental run planning.

I think at this point, it apparent and demonstrated that anyone with a bit of motivation, determination, and some money can build a working fusor capable of fusion, without any rigorous design work or calculation. In comparison with other fusion capable devices, a fusor is incredibly simple, and it has been shown many times that a working system can be slapped together with moderate effort and scrounged components. I am in no way whatsoever downplaying the challenge and accomplishment of getting a running system both operational and proven, which still requires a lot of work and personal investment to achieve. However, there is a major difference between producing neutrons and maximizing neutron production efficiency, as many experienced fusor enthusiasts are well aware of by now. Especially with the new developments and push towards very small fusor systems, we may have reached the point where in order to make that next leap in improvement, more rigorous planning and engineering needs to be taken to maximize its potential, and to better understand the underlying principles of the device. For example, as you mentioned, water vapor is definitely a major bane for fusor operation, and certainly would effect efficiency. Yet a lot of systems presented generally do not appear to go through enough conditioning to fully drive off or reduce the water vapor loads. While running a plasma would certainly help drive off water vapor, and easily burn it out of the grid, for such short runs that the general fusor enthusiast operates at, there is not enough thermal energy acting on the entirety of the system surfaces or long enough pumpdown times at high enough vacuum to effectively drive out all of the water vapor of the whole system. Plasma cleaning definitely helps in bombarding the immediate surfaces, but the process still takes time. Even for systems that are baked to several hundreds of degrees C and pumped at high vacuum levels, it still takes many, many hours of continuous pumping and baking to effectively eliminate or reduce water vapor in the system. Once the system is admitted back to atmosphere, new layers of water vapor will immediately start adsorbing on the surface. Back-filling the chamber with nitrogen or other inert gas would help reduce re-conditioning times.

As you also mentioned, it would be very interesting to start exploring more experiments involving the relationships of water vapor and other contaminants in the system and fusion efficiency. Based on the work I have done so far for my system for example, it should be do-able to estimate for example the total number of water molecules present in the system, and work on minimizing this number. Maybe there is a certain point where decreased water vapor has no more effect on fusion yields. I do not know these answers. And recent developments in small fusor operation does show a higher efficiency for small devices over large devices, operating at higher pressures. Therefore it may be beneficial to see exactly how far a system can be pushed and how much deuterium it can handle at a given operating vacuum level.

There is a lot of info presented here so far, and a lot more that I will be posting and documenting. I apologize for the incredibly long posts, and applaud anyone who has the patience to trudge through them. Hopefully something useful can be gleaned from these efforts for all experience levels. I know when I first started I was eager to try and build this thing as fast as possible and start doing experiments immediately - however, I have come to thoroughly enjoy and savor the design process, slowing it down, and breaking it down to the most fundamental levels, then building on top of it. I do have almost all the parts I need now, and will start to actually assemble the system. I just have a one more post at this point for the main calculations, as well as some thermal modelling work. I hope this weekend to at least assemble the low vacuum roughing side and maybe qualify how well sealed it is and the ultimate vacuum of the roughing pump. I will post more about this as I get to it, as well as other developments such as the titanium sublimation pump build.
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Richard Hull »

Just a little note....(Let's make that a long note).... My current fusor, fusor IV, uses a larger 6" sphere and has been around since 2004. It has had a terrible leak ever since I assembled it. I have a very accurate 0.1 torr heated Baratron gauge on the chamber. I use a standard 5 CFM Precision belt drive fore pump that can pump down the foreline, the cold diff pump and the fusor chamber to 10 microns in about 3 minutes. With the Diff pump on, I can hit the high 10e-5 torr level, at best, in the fusor chamber.

If I valve off the chamber at 10e-4 torr it takes only 4 minutes to leak to 2X10-1 torr, (200 microns). This is considered a terrible leak by any standard.

I rarely run the fusor now, save for special events and demos. I have, for years, taken, 2-3 days of 2-3 hour run-ins to condition this leaky fusor to produce 2 million fusions per second. (1 million neutrons/sec). The conditioning time is a lot easier for me to work with than tearing a fully functional fusor apart to locate and seal the leak. Whether this leak, over time, has created a water issue inside the chamber or whether the conditioning has to do with burying D2 into the chamber walls, is unimportant to me. I will say that the leak is in the fusor chamber itself. I have done a halfhearted effort on two occasions to locate the leak with the classic acetone/alcohol spray and have seen no clear sudden increase in pressure. I just gave up.

I realize that a true, dyed-in-the-wool vacuumist and "vacuum-head" demands an absolutely sealed system. Water loads and certain tramp gases are the bane of those demanding 10e-8 torr final vacuums. However, the person doing fusion in a fusor need not worry about water or a little leak, even as bad as mine. As noted before, the average newbie is a vacuum dunce and all his issues revolve around raw beginners failure of technique and failure to read a few basics in good texts or the vacuum forum FAQs here. Another tangle to their feet is in purchasing pumps that are shot or that, by their very nature, are not up to the task even when new. Finally, these same folk seem to never acquire good vacuum metering to even know where they are in a pump down scenario!! In the end, even a fairly rotten sealed fusor will do good fusion if the pumps are great and the operational technique rises to the occasion.

I am in no way condoning sloppy work. One should strive to do the best they can with what they have. There is a point, however, where pushing work to perfection where perfection bears no additional fruit for a specific task at hand.

Believe me... I have been here since 1998, at the very start of all of this, and virtually 100% of all vacuum systems ever assembled here are dismantled and stored or sold off the instant an abject failure to do fusion is at hand and even if there is a fusion win, the systems are doomed to being torn apart and stored or sold off. The folks here are for one thing....Doing fusion! A vacuum system is just a horrible and expensive briar-filled path they must begrudgingly trudge along on their way to the super highway of fusion. Many are at an age where all of this is an entertainment just prior to college or discovering that girls are warm and soft and nice..... fusion and all of its entangling vacuum and high voltage stuff, at this juncture in their lives, will be as dead as yesterday's egg salad.

This post, while not quite as long as others here, does tell the tale of the average vacuum here at fusor.net.

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
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

Richard Hull,

Thank you very much for sharing your insight and experience. As the first one who has really started and pushed the whole endeavor for amateur fusion, you certainly have some of the most experience and run-time of anyone here. While calculations are good for determining how the system might behave, ultimately it does still come down to practice and running it yourself. I plan on continuing my efforts and optimizing the system, and running new experiments long after I achieve first neutrons (if money allows). Since I have so many other projects lined up that are also non-neutron producing, this will ensure that my system won't be broken down and sold off after all this effort and time (I am also fortunately past college and married now, so I don't have the same distractions as the younger fusioneers will encounter.)

For a small system, maybe bad leaks and water vapor have more detrimental effects on neutron output than a large system like the one you are running, or maybe it will be similar to your experience. It will have to come down to experiment at the end, I couldn't possibly say one way or another. Hopefully if minor effects in the vacuum system itself such as residual water vapor can be proven that they do not interfere with neutron output for small systems, then full efforts can be mounted to maximize and improve its efficiency even more by other means. Chances are your observations are more than valid and applicable for small systems, but it would also be beneficial if someone could experimentally validate this on a small system to remove it from the list of ways to improve efficiency. I suspect that a super clean vacuum would probably give almost non-noticeable improvements, but you never know. My focus will be more on the experiments after, including implementing ion guns and other experiments. Right now I'm just taking it slow, one step at a time and documenting the process as I go and just focusing on the vacuum while I do not have the funds to support neutron experiments yet. By the time I can start doing it, hopefully my system will be well prepared to start pumping out results immediately (no pun intended.)
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

6.) Foreline Parameters and Pumpdown Time from Atmosphere to Rough Vacuum

From the previous five sections and numerous subsections, the major and critical vacuum parameters were explored and calculated for my V4 design. This section will cover the final main calculations for this system. Originally, I was going to include calculation estimates for pumpdown from rough vacuum to high vacuum in addition from atmosphere to rough vacuum. While the roughing times are much simpler, calculating the pumpdown time for high vacuum is much more complex and imprecise, relying on a large amount of simplifications and assumptions that dive into many of the complexities that makes high vacuum engineering a challenge. Because there are several ways of going about these calculations, and due to the assumptions needed and the very rough estimation and wide range of times, I will omit these calculations here for now and focus just on the foreline and roughing parts. Since fusors generally operate in the tens of micron range, which falls into the roughing pumpdown region, this may be of much more use and interest for the majority of people here on the forums.

Below is the PDF for the various foreline parameters and figuring out the pumpdown time from atmosphere to rough vacuum:


A few initial considerations should be noted for the roughing line. For the range explored, from atmosphere to about 10 microns, the load due to outgassing is negligible, and can be omitted. In addition, a well designed foreline should not have any effect on the speed of the foreline pump. Therefore, the effective speed of the system should be equal to the speed at the inlet of the pump. The foreline for this system was kept as short and minimal as possible, incorporating a necessary pressure gauge, isolation valve, and a replaceable molecular sieve foreline trap. While a foreline trap is not required, it can be greatly beneficial from keeping both the diffusion pump oil free of contamination from the roughing pump, and keep the roughing pump oil cleaner by absorbing moisture during initial pumpdown. A used trap on eBay was purchased at more than five times lower than its original cost new, and in excellent condition. Cost needed to be minimized, as well as size to make the system more compact and portable on a single mobile cart.

The first set of calculations in Section 1 of the PDF goes over figuring out some initial key parameters of the roughing line. First, the roughing flow regime is determined. This can fall into one of three categories: purely viscous flow, purely molecular flow, or transitional flow. This is established by the flow factor which is represented by Dp, where D is the diameter of the roughing line, and p is the mean gas pressure. Units here are cm and Torr, respectively. For my system, the diameter of the roughing line is 2.210 cm. The mean pressure is found between the starting and ending pressures – in this case, it is assumed that the starting pressure P1 is 760 Torr (atmosphere), and the ending pressure P2 in the foreline at the backing inlet of the diffusion pump is 2x10^-2 Torr, or 20 microns. Based on the boundary conditions for the Dp factor, it is verified that the roughing flow is indeed purely viscous.

Now that the flow regime is established, the equation for conductance of a pipe in viscous flow can be used to find the conductance of the roughing line. The general equation for viscous flow conductance is given in the PDF – since air at 20C is considered as the gas, this equation can be reduced to a simplified form for this gas input. The line is based off of KF25 hardware, and the diameter is assumed the same for all parts. The line is also treated as a single length line with the total length of all parts, including bends for simplicity. The conductance of the line is found to be 41,487.062 L/s. Note that this number is massively larger than other conductances previously calculated, but again, this is for an entirely different flow regime between atmosphere and very rough vacuum, where gas flow principles are much different than in molecular flows.

Once the conductance is found, then the effective speed can be calculated. First, the maximum speed of the forepump is determined. The flow rate of the selected pump is 6CFM, which can be calculated to 2.830 L/s. The pump speed was determined based off the minimum pumping requirements for maximum throughput of the diffusion pump from the datasheet. From the original diff pump manual, the minimum displacement of the backing pump is 80 L/m, or 1.333 L/s. The manual also specifies that a single stage rotary pump capable of 120 L/m (2 L/s) and an ultimate vacuum of 0.1 Torr is recommended. The backing pump selected is a two-stage pump capable of 0.015 Torr and has about 53% higher throughput than the recommended min displacement, which was also in part selected due to pump availability and cost, and should have plenty of overhead for the gas loads that will be seen in the system or the effects of the foreline trap (in viscous flow). Based on the equation to calculate effective speed with calculated conductance and established max speed, the effective speed was found to be almost exactly 2.830 L/s. Since the foreline conductance is so massive compared to the maximum speed of the pump, there is virtually no change in the calculated effective speed. From this, it can be shown that the conductance of the roughing pipeline in viscous flow has a negligible effect on the maximum pumping speed of the backing pump, and is indeed well planned and designed.

Once the effective speed has been determined, the pumpdown time can be calculated. One of the factors required in the equation is the volume that needs to be pumped. Therefore, the total volume of the system is calculated. Because the roughing pump will pump out the entire volume of the whole system, all parts must be factored. This includes the volume of the foreline, the total internal volume of the diffusion pump, the volume of the high vacuum pipeline, and the volume of the chamber. This was determined from CAD models of the system. The total volume of all above parts was found to be about 5.464 L. Based on this volume, with a effective pump speed of 2.830 L/s, a starting pressure of 760 Torr, a final pressure of 0.020 Torr, and the ultimate pressure attainable by the forepump (from the pump datasheet) of 0.015 Torr, the total pumpdown time is found to be about 23 seconds. Based on the small size of the system, and minimized length of the foreline and high vacuum pipeline, as well as the speed of the pump, this would seem like a very reasonable number to expect.

This section effectively concludes the major engineering design calculations for this system. More advanced calculations may be applied later as the system is tested, or new components are introduced, but at this point all the major required parameters have been well explored and estimated. From here on out, further posts will be dedicated to the actual build and testing of the V4 design as presented and calculated. Other parameters such as thermal modelling of the system and various components, such as the diffusion pump, titanium sublimation pump, and basic fusor grid will also be presented as they are completed as well.
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

BUILD LOG UPDATE #01

Roughing Line Assembly

Now that all of the major calculations have been completed, besides some additional CAD work, organizing drawing sheets for the full system, and thermal modelling, the system can finally start being pieced together. While the calculations are very enlightening, actually building the system is where the real fun begins. It really is thrilling to see a high tech system slowly come together after months of planning, especially when it comes to life exactly as designed in CAD.

Below are all of the parts for the roughing line, laid out on a sheet of clean, coating free aluminum foil. All of the parts were handled with clean nitrile gloves and thoroughly wiped down with 99.9% pure anhydrous isopropyl alcohol. While this is not really needed for the low vacuum side, it never hurts to get into the habit of proper practices and procedures for handling high vacuum equipment.

20180210_185125.jpg

The foreline trap needed to be disassembled and cleaned prior to use. The trap material, some form of zeolite or alumina pellets, were already well contaminated and spent. They can be seen as small purple pellets in a plastic bag at the top of the picture below. The inside of the trap along with the stainless steel basket was thoroughly cleaned spotless with the alcohol. I did not refill the trap yet with new adsorbent since I would be qualifying the roughing line first and calibrating the thermocouple gauge on a different oil-free pumping station, so opening up the container and absorbing moisture during qualification and calibration runs is unnecessary for now.

20180210_192855.jpg

After all of the parts were cleaned, the roughing line was assembled. The side on the left with the 90 degree manual valve goes to the diffusion pump roughing inlet, while the side on the right with the foreline trap goes to the backing pump.

20180210_195510.jpg

Finally, the adapters for the roughing pump were assembled to the pump. The pump inlet is a 3/8" male flare fitting. Several adapters are needed to go from this to the KF25 to 1/2" female NPT adapter. I decided to keep the adapters equal to or larger than the original adapter on the pump to keep pumping speed high. A 3/8" female flare to 1/2" male flare adapter connects directly to the pump fitting, followed by a 1/2" female flare to 1/2" male NPT, which goes directly to the KF25 adapter. Teflon tape was used for the 1/2" male NPT, but not for any of the flare fittings. The pump also includes an additional small 1/4" male flare fitting off of the side of the main 3/8" flare, which I replaced the provided plastic cap with the appropriate 1/4" brass flare cap.

20180213_202427.jpg
20180213_202433.jpg
20180213_203731.jpg

With the foreline components assembled, the next phase would be to qualify the vacuum integrity of the line and calibrate the thermocouple gauge sensor. This has already been accomplished in the past couple of days, and I will go over the calibration procedure and data in the following posts.
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Richard Hull »

Looks like my system on fusor IV. Short connections, pump to TC gauge, fore line valve and Diff pump. You bought well.

I wish others would start out this way. Pump down of the fore line should be near instantaneous. The diff pump will also be quick.

My system can hit fore line plus the diff pump to 20 torr in about 3 minutes. Once the fusor chamber valve is opened it takes about another 5 minutes to hit 15 microns throughout the entire vacuum system. I turn on the diff pump heater once the system crosses the 50 torr mark. The boiler takes over 15 minutes to start pumping. By that time, the system is at about 12 microns due to the mechanical pump.

Your stuff and assembly shows proper fore thought gained through study. Most newbies here have to blunder into these realizations.

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
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

Richard Hull,

Thank you for your comments and sharing your experience with your system. I ran the final qualifying test today with the roughing system, and I am very pleased with the results, and even more excited to continue now. Having gone through many prior posts and seeing how many issues others have had pulling a vacuum on their system with KF and even CF hardware (which is designed to not leak even remotely close to fusor pressures), I was honestly worried that this would be a massive struggle. However, so far everything has progressed very smoothly without any issues yet and has been significantly better than I was expecting. I would say that doing thorough research ahead of time, working very slowly and methodically, planning each step, and breaking the whole thing down into smaller subsystems has already proven to be massively beneficial. Troubleshooting during qualifying runs took about a minute or two for the roughing side.

For those who are starting out, I would absolutely recommend to qualify each subsystem in order. Don't bother putting together your high vacuum side until you build the roughing side and qualify the vacuum integrity of the line, calibrate your roughing sensor, and establish your minimum forepressure. It's very tempting to rush ahead and start bolting nice and shiny high vacuum hardware together and try running a fusor as soon as you get all the parts, but there is really not much point until you can properly back the system first. This will eliminate issues with a major part of the whole system, and make troubleshooting easier as you progress. Even better if you can CAD your system out first, as it will help figure out the optimum orientation and allow you to plan, build, and change things without the need to spend money.

After today's test, I was able to qualify that the roughing line and the roughing pump were able to achieve a minimum stable pressure of about 12.5 microns, which is much better than I was anticipating, as my initial goal was to shoot for 20 microns. The pressure shot down to the several tens of micron range very fast, and leveled out after about 5-10 minutes of pumping to its ultimate pressure. I vented the system and re-pumped it back several times, and I got the same numbers consistently. The foreline itself was qualified to even lower pressures prior during gauge calibration. One thing that I will need to improve however is minimizing vibrations of the roughing line due to it being so short and directly connected to the pump, which may damage the thermocouple gauge in the long run, but I have located the largest source of physical vibration in the line (the foreline trap) and have a plan for reducing system vibrations. I will post the calibration procedure and curves for my gauge sometime soon.
Last edited by Michael Bretti on Sat Feb 17, 2018 2:42 pm, edited 1 time in total.
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

BUILD LOG UPDATE #02

Roughing Line Testing, Thermocouple Gauge Calibration, and Roughing Pump Qualification

With the roughing line complete, I was able to run tests on its vacuum integrity as well as calibrate the thermocouple gauge sensor. I performed these tests earlier in the week, and finished the last bit of testing today.

The roughing line and thermocouple gauge were qualified using a scroll-pump backed calibrated Varian helium leak checker pumping station. First, the roughing line was tested. The roughing isolation valve was closed, and the line connected directly to the system's KF25 inlet, right to the foreline trap, where the backing pump would normally be connected. The system was turned on and allowed to go through its start-up sequence. The line was then pumped down, and achieved an ultimate vacuum of 2 x 10^-3 Torr, or 2 microns, which is about where the pumping station is rated at. No adjustment or tightening of the components in the line was needed. The foreline trap was also not filled with fresh molecular sieve during this phase of testing since the scroll pump is a dry, oil-free pump, so contamination of the thermocouple gauge from back-streamed oil is not present.

Next, the thermocouple gauge was calibrated against the leak checker. The thermocouple gauge is a VGT-1504 from LDS Vacuum. The gauge was connected to a very simple 0.22VAC source (several step down transformers controlled by a variac), and the output was measured directly by a sensitive enough voltmeter that could read down into the sub-millivolt levels. The line was first pumped to its ultimate vacuum of 2 microns with the thermocouple gauge off. Next, the gauge filament was slowly brought up to full voltage and allowed to stabilize for about 5-10 minutes. Using the isolation valve at the end of the roughing line, I slowly introduced very steady and controlled leaks into the line. This took a bit of fine control and practice, very slightly tapping the handle to introduce ever increasing wisps of air into the system, allowing the leak checker and thermocouple gauge to stabilize at each reading. The gauge was calibrated against the leak checker from 2 microns up to 700 microns (7 x 10^-1 Torr), where the signal from the gauge cut out. Below is a PDF of the calibration and plots of the calibration curve across its measured range:


Note that this curve is valid only for this particular gauge, since equivalent gauges will still vary. However, since thermocouple gauges are not the most accurate to begin with, these numbers may provide a rough guideline for others using equivalent gauges, though the gauge should still be properly calibrated against a known measurement source for more correct accuracy.

Once the gauge was tested, the system was evacuated and pumped down several times to reconfirm the numbers, which seemed to track with good repeat-ability. Next was to qualify the whole line attached to the vacuum pump. The roughing line was connected to the pumping station at the isolation valve, and the roughing pump was connected to its proper location under the foreline trap. The roughing pump has a built in isolation valve, which was closed during this next test. The pump also was not filled with oil. With this test, I could qualify the ultimate vacuum of the entire line, the brass adapters, and the isolation valve of the roughing pump. At first, I pumped the system down and got a pressure of 160 microns. However, since I already qualified the whole vacuum line prior, I knew that the problem could only lie either in the brass fittings or the roughing pump isolation valve. Sure enough, as soon as I tightened the extra 1/4" flare cap port first, the pressure dropped down almost immediately to 16 microns. I then tightened each of the brass fittings further on the main line, which dropped the ultimate achievable vacuum to 6 microns. The thermocouple gauge was also turned on and read to check the calibration numbers against the new readings.

Once everything from the roughing pump isolation valve to the roughing line isolation valve were qualified, I could finally test the ultimate vacuum of the roughing line using the actual roughing pump. The line was disconnected from the leak checker pumping station and tightly closed. The roughing pump was then filled with oil, and the trap was filled with fresh sieve material. The pump was turned on, and pulled down to its minimum pressure shown by the built in dial gauge immediately. The thermocouple gauge was turned on and allowed to stabilize for another 5 minutes, then readings were taken. After about 5-10 minutes of pumping, the pump was able to achieve an ultimate pressure of about 12.5 microns, which corresponds to an output voltage of this sensor at 12.1mV. Since my initial target was 20 microns, I was extremely pleased with the results. This ended up being much better than I was anticipating. The line was vented and re-pumped several times to verify the results.

Now that the roughing system is well qualified, I can set this aside and work on the next subsystem, which will be the cooling system. This will consist of three separate cooling loops - a large, air cooled heat exchanger and water based line for the diffusion pump, and two smaller loops chilled with an insulated peltier cooling array for the water cooled baffle and the titanium sublimation pump, using some sort of coolant such as propylene glycol. The target temperature of the diff loop is 20C or better, while the two smaller loops for the baffle and sublimation pump are 0C or lower. Cooling flow, pump speeds, temperature, and interlocking will all be automatically controlled and monitored. The goal is to integrate this and other systems into a semi-automated pumpdown sequence and develop a full user interface for all of the subsystems for very precise monitoring and control of the whole system
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

DOCUMENTATION UPDATE AND NEW THERMAL ANALYSIS

Just wanted to post a quick update regarding the efforts on the high vacuum system build. The efforts presented here so far are not just for a single project, but an ongoing research effort over multiple projects that will span years. As a result, there will be an incredible amount of additional documentation to cover. As is, from just my walkthrough posts alone, I have generated the equivalent of about 30 pages of Word documentation on explaining the calculations alone, in addition to close to 100 pages of total calculations. This is not including CAD models, drawing sheets, and simulation data. This is also only for a single vacuum system, which has not yet touched power supply design, cooling systems, instrumentation, and control. There will be even more when I start getting into ion beam design and other vacuum setups. Because of this, I have been working on migrating and expanding these efforts for a while now beyond just the initial scope presented, and will be focusing on a full initiative for creating open-source engineering and documentation for high vacuum and ion beam systems, centered around a hobbyist-budget approach. I have been slowly organizing and setting up these resources on a new site, which can be found here:

http://appliedionsystems.com/

I will still be posting updates on the progress of builds here, but the posts will be much briefer than prior posts in this thread due to the large amount of data and resources associated with each project. In the long run, having a dedicated site for documentation and resources will make it much easier to manage all of this information. The focus is also more on high vacuum and ion beam system engineering rather than strictly fusor efforts, although the fusor will be covered as I get to it.

As an additional update directly related to prior posts in this thread, I have just completed the first thermal modeling simulations for the system. The first series of simulations goes into modeling the steady-state thermal characteristics of the Edwards EO4 diffusion pump. This is a simplified model to give me initial data to start designing the temperature-controlled cooling system for the diffusion pump. The vapor jet stack is omitted from the model for simplicity. However, the proper fill of 175mL of DC-705 oil was included into the model with the proper thermal characteristics of the oil. The simulations are also run using convection coefficients for the oil as well as the internal surfaces for rough vacuum conditions to compensate for the model being static instead of simulating the dynamic properties of the oil spraying against the walls, condensing, and contributing to oil cooling. Based on the comparison between the uncooled vs. cooled diffusion pump models, the discrepancy between average oil temperature from simulated to expected is accounted for. The full summary can be found here:

http://appliedionsystems.com/portfolio/ ... sion-pump/

The second simulation was to look at the steady-state thermal characteristics of the water-cooled baffle and adapter-plate subsystem to be used in conjunction with the diffusion pump. This looks at a comparison of the thermal analysis of an uncooled vs. cooled water baffle while the pump is operating with proper cooling to determine what will be needed for the temperature controlled chilled baffle cooling loop. Details of the study are presented here:

http://appliedionsystems.com/portfolio/ ... ed-baffle/
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

Another update on the status of the build progress so far. Since all of the vacuum calculations have been complete I have been focusing primarily on the thermal management system planning, and recently in the past two weeks I have been preparing to run simulations looking at the sputtering damage and ion implantation effects of a diffuse plasma on stainless steel surfaces during glow discharge cleaning, which I will post more information later in the advanced topics section.

As mentioned above, thermal modeling simulations have been completed for the two major thermal loads of the V4 system: the EO4 diffusion pump, and water cooled baffle. These simulations are used in the planning of the PID controlled closed loop cooling system. Since I have the manual for the diff pump, as well as the diff pump itself, I know just about all of the major parameters needed to model and simulate the thermal effects of running the pump and cooling. An extensive search through literature yielded the necessary thermal parameters needed to input into the thermal simulation software for various heat loads in the system, both for outside atmospheric conditions as well as vacuum conditions, and average values are used. The model was simulated without the vapor stack for simplicity, but the heater, internal structure, and proper fill of 175mL of DC-705 oil was modeled, with all of its thermal and physical characteristics. Since the thermal modeling software used can only simulate steady-state static models, the pump was modeled both without cooling and with proper cooling to determine the proper thermal gradients and effects of cooling. Fusion 360 was used for the modeling. Below are a couple of the resulting images:

Edwards EO4 Diffusion Pump Thermal Modeling - DC 705, No Cooling, Internal View, Legend.JPG
Uncooled Diffusion Pump

Edwards EO4 Diffusion Pump Thermal Modeling - DC 705, Water Cooling, Internal View, Legend.JPG
Diffusion Pump Cooled with 25C Water


All of the input parameters and full details can be found here: http://appliedionsystems.com/portfolio/ ... sion-pump/

A few interesting things to note:
1.) The heat shield, despite its simplicity, contributes to a large degree of thermal shielding, and illustrates its importance.
2.) The body of the diffusion pump, both in uncooled and cooled conditions, does not directly affect the oil temperature itself through thermal conduction through the bulk mass of the walls down to the oil pool. Based on the results, for the 850W power input of applied heating to the heater element results in an average temperature of the DC-705 of 291C, which is higher than the target operating temperature of 245C. The average temperature of the casing of the body, with the pump uncooled, rises to a steady state value of 118.5C. With the recommended max temperature of 25C of water cooling, the average temperature for the oil stays the same - however, the body drops to an average temperature of 50C. Due to the effects of the oil being cyclically evaporated, sprayed, and cooled along the surface of the casing, condensing back down into the oil pool and restarting the cycle, oil temperature is therefore controlled. The difference in temperature between the body casing of uncooled vs. cooled conditions and average oil temperature between simulation and expected nominal temperature therefore accounts for this effect.

The next thermal simulation looks at the effect of cooling on the baffle above the pump, accounting for thermal conduction loads while the pump is running. The simulation was run with the above diffusion pump input parameters, in a steady state cooled condition, looking at both the thermal gradients of the uncooled baffle and adapter plates, and cooling the baffle with a constant flow of 15C chilled water. Below are cross-sectional views of the baffle and planned aluminum adapter plates:

Water Cooled Baffle Thermal Modeling - Uncooled Cross-Sectional View.JPG
Uncooled Baffle Assembly

Water Cooled Baffle Thermal Modeling - Cooled 15C Cross-Sectional View.JPG
Baffle Assembly Cooled with 15C Water


The normalized results comparing the two scenarios is shown below:

Water Cooled Baffle Thermal Modeling - Uncooled vs Cooled 15C, Side by Side, Normalized Gradient.JPG

Full details of the simulation and all input parameters can be found here: http://appliedionsystems.com/portfolio/ ... ed-baffle/

Without any cooling on the baffle, with the diffusion pump running at nominal conditions, the average baffle temperature rises to 32.68C. With a flow of 15C water in the baffle cooling channels, the average temperature is reduced to 20.05C. The ideal goal is to run the baffle with coolant chilled to at least 0C, preferably lower, with 15C as the maximum coolant temperature. These simulations establish that cooling of the baffle with at least 15C water will allow for an average temperature of 20C, plenty enough for the baffle to be effective. The manual for the diff pump states that with 15C water cooling of the baffle and pump, and all metal gaskets, with the proper oil, the system can reach an ultimate pressure in the 10^-10 Torr region, without the need for cryo trapping. Even with o-rings in the system, which I am looking to add differential pumping along with the chilled baffle, ultimate pressures based on prior calculations in the 10^-7 Torr range and potentially lower are reasonable to expect.

As described before, there will be two separate cooling loops for the system: one for the pump, and one for the baffle. Each system will be a triple-loop, closed loop PID controlled chiller based off of Peltier cooled heat exchangers. Primary fan cooled heat exchangers will remove all heat from the loads before re-entering the coolant tank, which will be continuously chilled with recirculating coolant between the Peltier exchangers and the tank. The system will be automatically monitored and controlled during runs to give precise system control, and interlocked to the main and diffusion pump power as well, and will be part of the (hopefully) automated pumpdown sequence. The peltier chillers, pumps, chiller exchangers, and power supplies have been acquired. The tanks and main exchangers still need to be obtained. All cold areas will be thermally insulated with proper material and keeping lines as short as possible to minimize heat transfer from the cold sections to ambient air. The above simulations will be verified using a high-sensitivity biomedical thermal imaging camera that I acquired several years ago for free. More details to follow as the build progresses.
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

After a couple of months of additional project documentation, CAD, thermal simulations, and just recently, advanced plasma simulations using TRIM, I have had the opportunity to work on some physical assembly. I started to assemble the main chamber and pipeline of the V4 system. An overview with a brief write-up as well as resources for where parts were obtained can be found here:

http://appliedionsystems.com/high-vacuu ... -assembly/

Below are some pictures of the initial assembly. Everything was assembled on a clean aluminum foil surface, with ports cleaned with 99.9% isopropyl alcohol and Kimwipes, and everything handled with clean nitrile gloves:

2,75in conflat blanks.jpg

High Vacuum System V4 Build Pic 1.jpg

High Vacuum System V4 Build Pic 2.jpg

High Vacuum System V4 Build Pic 3.jpg

High Vacuum System V4 Build Pic 4.jpg

High Vacuum System V4 Build Pic 6.jpg

High Vacuum System V4 Build Pic 5.jpg

High Vacuum System V4 Build Pic 7.jpg

The only deviation from the original CAD model so far was the decision to have the manual gate-valve be angled towards the back as opposed to the front. A few of the ports will be kept covered with the aluminum foil while I decide whether or not I have the resources and funding available to build and install an anticipated Faraday cup and high-power pulsed electron gun. These will be installed in the two upper ports on the main 5-way cross chamber for my first experiments (I won't need a gas handling supply for this setup, and have several experiments planned for the intense pulsed e-beam and associated x-ray production.) Chances are however I will just blank these ports off and focus on baking, conditioning, pumpdown, and the automated control system programming for now. I will wait to install the HPT-100 wide range vacuum transducer on the lower 4-way cross port until the full system is mounted to the pump and support structure to prevent damage to the sensor.

Next on the list will be to get the adapter plates machined, bolt the chamber to the diff pump, design the housing, and build/mount the chamber and pump to the housing. In parallel with these efforts, I will be working on finalizing the design and starting the build for the cooling system. I already have most of the parts for the diff pump chiller. Efforts will also be made to start writing code for automated pumpdown control and remote monitoring. I am also considering working on developing automated control and monitoring via wireless through a phone app. However, remote computer control and monitoring will be the primary focus for this subsystem for now.
User avatar
Dennis P Brown
Posts: 3159
Joined: Sun May 20, 2012 10:46 am
Real name: Dennis Brown

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Dennis P Brown »

Outstanding work to date; I really look forward to your neutron detection system build - or will you just buy a turn key system?
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

Dennis P Brown,

Thank you for your compliment. While this system is certainly not the fanciest setup here, my goal is to utilize a simple and compact topology to create a fully engineered and modular test platform with precise control over every parameter within reason. It takes much longer and requires a large investment in time, planning, and research, but the end results should be well worth the effort. I would rather have precise experimental control than large neutron yields.

I am also particularly looking forward to when I get to neutron detection and instrumentation. At this point I would say that I would be looking to build my own detection system as opposed to purchasing a full turn key solution. I already have a bunch of ideas I would like to implement for multi-detector arrays, as well as potentially playing with neutron collimation. I will be taking the same engineering design and approach towards that build as well, though I don't foresee it happening for a quite a while unless I can secure some additional sources of external funding to speed up the process. Since I will also be working primarily with fast-pulsed systems (high peak power pulsing in the nanosecond time scale), detection will provide some interesting challenges to tackle as well.

My main focus will be pulsed beam-on-target, but I will also dabble with pulsed fusors. In a textbook I have on ion beam systems, there is a design for an ion gun that only requires a kW of input power and generates D2 beams up to several tens of mA continuous, which could certainly yield some high counts, or be excellent for target pre-loading. My current system should allow for two ion gun inputs with a single target. Pulsing a fusor at voltages in excess of 50-100kV should also be relatively straightforward in terms of the pulser design (as pulsed power is one of my areas of expertise in engineering), and should be able to generate peak currents in the range of amps to 10s of amps for short durations, (tens to hundreds of nanoseconds), with only hundreds of watts of input (one of the great benefits of energy compression in pulsed power).
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Richard Hull »

I am sure you are aware of the pitfalls related to pulsed fusion detection. I call this putt-putt boat fusion. Those multi-ampere pulses are massive noise generators making electronic neutron detection nightmarish. Activation analysis may have to be the sole arbiter in such systems.

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
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

Richard Hull,

Yes, noise shielding is certainly a major issue with high peak-power pulsed systems. Fortunately I have directly worked with active shielding and noise suppression for electronics and instrumentation around these types of pulsers. It is a challenge, but I have dealt with it before very successfully. One of the systems I actually built for work was a 90kv Marx generator to simulate electron gun arcing faults specifically for EMP and noise shielding for sensitive electronics around the fault. My goal is to develop the pulsers to work successfully with my pulsed e-gun setup, then migrate it over towards pulsed neutron work later (the e-gun setup would be less costly due to the fact that no expensive gas or gas handling is needed.)

Activation analysis was actually the main method of detection I was planning on implementing for the high peak power fast-pulsed systems, and I look forward to the design challenges and getting to that stage of testing.
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Richard Hull »

The one thing about activation; it is the only warranted noiseless, absolute indication that thermal neutrons had been there. As you note, calibration can be tricky, but with good math and the knowledge to apply it, calibration can be very accurate. You will just need a good averaged flux over time. From this, based on the pulse peak voltage and current, the flux per pulse should be easily computed.

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
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

Richard Hull,

Thank you for your advice. I also just finished reading that new PDF you posted on activation - really fantastic introductory paper on activation, I greatly enjoyed it and it was very helpful, especially the section at the end explaining different target materials for activation. I was planning on starting with silver, and seeing this list of additional elements to explore has given me some new ideas for experimental test stand development that I think would be very exciting to build and share with the fusor community. Although I won't get to activation for a long time, I can certainly start working on gathering the necessary information and designing the various detector systems.
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Richard Hull »

Silver might have been #1 in my list and it certainly is the most common activated material here at fusor.net over the years.

Rhodium was #1 simply due to the raw data related to it, my own experience with it, and the fact that Enrico Fermi used it all through his work that lead to his Nobel Prize. He was forced to use made-up sources for his neutrons,(radium-beryllium), to calibrate and quantify them, he needed a fast activation element to 100%. Rhodium was ideal.

I had a friend who wisely purchased a 1 ounce bar of Rhodium, a short while back, when its price plunged to $700 per ounce. I contemplated it back then, but waited too long and the price soared again. As noted 1/10 ounce bars are available.

In my activation of my friend's bar, I was pleased that it exceeded silver in activation and that it took only 4 minutes to fully activate to 100% of its attainable radioactivity with a given flux!

As an ideal activation element it will forever remain #1 on any and all lists with only its crippling purchase price being its sole impediment.

Again........

100% of its atoms are ready to activate as it is a single isotope element. Its capture cross section is high and the finished radioactive product has a short half life, meaning weak and limited neutron sources can be rapidly checked out and quantified.

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
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

New update on the V4 System build. After additional planning and design, I have settled on a table I will be using for the system. Below is a CAD model of the proposed design:

System V4 Dual Purpose Table Assembly - V4 Chamber.png

The experimental table consists of a 2'x2'x3' structure made from 80/20 10 series extrusion. The 80/20 was purchased off ebay for half its original price won during surplus lot auctions. I ended up getting both black anodized and silver 80/20 since that was what was available for cheapest in bidding lots, and arranged it to have at least a symmetric and pretty cool look color-wise. I was originally going to go with a 2'x2'x2' table, however after discussing this project and another vacuum system project with one of my friends, he suggested to save money by combining the infrastructure of the two systems to support both. After some careful planning and design, I have been able to come up with a low-cost solution to support both systems simultaneously. Only the 2.75" conflat based system is currently shown in the model. The roughing pump is in the center, laying on the floor to minimize vibrations transferred to the rest of the structure, but still keeping it within the build volume. There will be ample space for the power supplies, diffusion pump chiller, baffle chiller, and other control electronics. An additional 6" KF25 bellows line will be needed to connect the roughing pump to the main roughing line under the foreline trap.

The second system is based around a 6" conflat tee that I obtained for free. This system will be used for ion and plasma engine testing. Last year I had originally bought a very large 6" throat gate valve/butterfly valve combo for the 21 port custom conflat system I posted about a while ago, which after looking at the logistics to get operational, was recently sold. I also obtained an 8" water cooled baffle for the original system as well. Both the baffle and the valve were $100 each, and in excellent condition. Since I already have another Edwards EO4 diffusion pump, as well as the baffle, gate valve, chamber, and basic feedthroughs, all that is really needed are the two aluminum adapter plates to fit everything together. I was going to build a completely separate system for this setup, but since the diffusion pump is the same as the V4 system, it was decided to extend the table to accommodate this test chamber as well. I will end up splitting the roughing line symmetrically to feed into both diffusion pumps, as well as route cooling for both diffusion pumps and baffles to run off the same cooling system. I will only run one system at a time, never both simultaneously, so in my control architecture I will make accommodations to switch between pumping and cooling for both systems. Initially, this second system was planned for operation years down the road, however using this shared setup, I will be able to deploy it and start testing it in a much shorter time-span, since the infrastructure will already be built and shared from the first system. The cost savings implementing this shared topology is on the order of $2k or so. I will post an updated CAD render of both systems mounted with the proposed roughing line upgrade shortly.

In terms of the actual physical build, the 80/20 has already been ordered and arrived. The remaining hardware is coming in today. I have also already started modeling the chiller block subsystem of the diffusion pump cooling system. The main heat exchangers will be arriving within a week. Once this chiller is qualified, then I will proceed in building a second chiller for the baffles. I will also post design specs of the chiller as I progress further, but I have most of the parts already for the chiller block.
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

Bit of a sudden and unforeseen setback to the above plan. Upon completing the 80/20 table, I decided to open up one of the two hermetically sealed boxes that, when I received them along with the open diffusion pump, was told they were new, unopened pumps. The boxes were the same dimensional size and weight as the EO4 diffusion pump. However, it turns out that they were nothing more than very large circuit breaker boxes. They were such old pieces of equipment from decades ago it was forgotten what was in them and assumed to be pumps. Turns out I only have one instead of three diff pumps that I initially thought I had. Rather disappointing considering a large portion of future systems relied on the requirement of having multiple of the same diff pumps. However I am still quite happy that I have even one diff pump in new condition. I could still proceed to mounting the second chamber since I have most of the parts anyway, but running it will have to wait for now, since the small chamber is priority.
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

Quick build update. Most of the 80/20 table has been completed over a week ago. Still leaving some of the center support beams off until I actually start mounting equipment:

20180519_214150-1.jpg

The design of the chiller system has just about been finalized, and almost all of the remaining parts have been ordered. Many of the parts were ordered on eBay from overseas suppliers to save cost, so they will take over a month to get in, but that gives me plenty of time to work on other system designs. I am currently working on the cad model for the final design of the chiller subsection, which will be posted when completed. Its current specs are rated so far for around 1200W cooling capacity (1000W air-liquid heat exchanger capacity, and an additional 200W liquid-liquid heat peltier chiller capacity.) As described before, it is a triple-loop, closed loop peltier based chiller. The EO4 diffusion pump works optimally with cooling water at the inlet in the range 15C-25C (max), up to 850W of power. The entire chiller will be mounted to a 2'x2' MDF panel that will be mounted vertically to the back section of the table. The system will utilize 4 thermocouple gauges, mounted at the diff pump inlet, outlet, in the main water storage tank, and the secondary storage tank for the peltier heat exchanger subsection, in addition to a water flow meter. All three pumps used in the system will be PWM driven and modulated based on temperature readings to keep precise temperature control of the system, and controlled through an Arduino Mega.

I also have left space in the design for future upgrades to be able to double the entire cooling capacity of the chiller to around 2400W to be able to support my second larger high vacuum system for ion engine testing. Both of these systems will share the same cooling and roughing infrastructure to significantly reduce costs of having two operational systems. I have been able to secure an even larger diffusion pump in practically new and unused condition for free for the ion engine test stand, and as a result, this pump will require much more cooling capacity (with the temperature range at the inlet the same as the EO4). I will be posting details of this second build in a new post specifically for that system.

One final note. In addition to the new diffusion pump, I have recently acquired an incredible piece of physics equipment that will allow me to experiment with charged particle beams at energies and peak power levels unlike anything presented on these forums here so far. If I am successful with this new system design and rebuild, it will be, as far as I am aware, the highest energy and peak power charged particle beam system currently built and operated at the hobbyist level. In addition to this component, I also obtained supporting vacuum flanges and pumps to run it. I will post more details of this new, third system in a different forum topic, but its potential can bring about an exciting new level of physics research at the hobbyist level.
User avatar
Dennis P Brown
Posts: 3159
Joined: Sun May 20, 2012 10:46 am
Real name: Dennis Brown

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Dennis P Brown »

If your high energy charge particle device is so powerful, then you had better be aware of the x-ray hazard and be certain of your shielding.
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

Yes, I am fully aware of the hazards, particularly with xrays. This device will not be operational for quite a while, so it will be a long term endeavor. I also work at a nuclear research facility so I have plenty of experience with radiation hazards and training. I also directly work with these devices regularly, and am well versed in their operation. I spent over a year almost exclusively working on various pulse power drivers for one, so this is by no means a typical amateur project.

I will be rebuilding it for a different mode of operation than its current configuration. It will only operate at very low repetition rates of a few hz at most, at only several nanoseconds in pulse widths. However I am looking at ultimately reaching the 1MeV barrier, with peak power levels of hundreds of MW at the nanosecond timescale. Not sure if I can drive 1MV across it yet, but I should be able to very readily achieve hundreds of kV without issue. I am already working on the preliminary specs for the pulse power driver for the rebuild. I will release more details later in a different post.
Michael Bretti
Posts: 175
Joined: Tue Aug 01, 2017 12:58 pm
Real name: Michael Bretti

Re: High Vacuum Engineering Design, Analysis, and Build of a Small-Scale Multipurpose System

Post by Michael Bretti »

New update on the cooling system build. I have finally finalized the CAD designs and layout for the system. I have posted a full overview description as well as system specs, which can be found in the link below. I plan on eventually adding the engineering drawing sheets for all of the components, as well as datasheets for the parts used in the build:

http://appliedionsystems.com/portfolio/ ... p-cooling/

Just wanted to share a few of the CAD models for the proposed design. The first is the actual peltier chiller block. This is the second loop of the triple-loop system, where water is continuously flowed through the chiller from and to the tank. The cooled water mixes with the incoming water at the end of the first loop from the heat exchangers after the diffusion pump, which will help for better temperature control. The chiller block consists of five TEC1-12708 peltier modules sandwiched between two aluminum water cooling blocks - one for the hot side, and one for the cold side. The hot side features three solid copper heat sinks with integral cooling fans for initial heat extraction, where an additional 500W air-to-liquid heat sink removes the rest of the heat before entering the secondary storage tank:

Peltier Chiller Block - Resized.png

The next picture shows the full assembly of the chiller block. This includes 1" thick XPS insulation foam on the cold side, and HPDE plates to hold the assembly together. All thermal interfaces between heat sinks and peltier devices will use Arctic Silver 5:

Peltier Chiller Block Assembly - Resized.png

The next renders show the full assembly positioned on the 24" x 36" x 0.5" MDF board. Unfortunately a lot of the components have varying inlet sizes, so several different tube sizes and adapters were needed to mate everything together. The tubing consists of various sizes of opaque black EPDM, rated for -40F to 300F. This model took quite a while to finalize, as there were a significant number of hoses and connections to position and mate. However, it has let me fully plan out the system before the build. As mentioned above, the system has space and contingency built in for a future upgrade to double its heat handling capacity from 1200W to 2400W. Right now the estimated power consumption at nominal operation is expected to be around 450W:

Cooling System Assembly - Top View.png
Cooling System Assembly - Isometric View.png

Finally, the full cooling assembly is mounted to the 80/20 experimental stand with the V4 high vacuum system and roughing line:

System V4 Dual Purpose Table Assembly - V4 Chamber with Cooling System.png

So far, everything is falling into line well with the design. It should make for a compact but highly controlled and modular system. The cooling system has admittedly cost more than I had initially wanted, but it is best to plan and invest in a solid system now that can handle all of my needs for any vacuum projects in the foreseeable future. Since the roughing line and cooling system will work for both diffusion pumps I have, with a couple of aluminum adapter plates I can quickly and easily change between the three high vacuum chambers I have for testing, which saves a significant amount of money for the future, and provides me a robust experimental test stand design that will allow me to run all of my physics experiments for years to come. I have obtained all of the hardware to start the build, so I will be working on the actual cooling system in the next few weeks. I hope to actually do some test runs, as well as do thermal imaging to analyze the response and performance of the system. I will also be starting to work on the control software for the cooling system, which will be used to set several interlocks for the main diffusion pump power based on the system temperature and flow rates.
Post Reply

Return to “Vacuum Technology (& FAQs)”