Connecting a Pump to High Vac System

[bk8509a]

I've been checking out the Vacuum FAQs and I cant find anything about how people are connecting mechanical pumps such as a two stage rotary pump to a steel vacuum chamber. My chamber is going to be pretty standard in design, CF hemispheres with a bunch of 2.75 inch half nipples. What kind of port should I have welded on my chamber when I purchase it in order to connect a pump to it? Do different pumps have different ways of connecting to a vacuum chamber?

Thanks in advance,

BK

PS: Sorry if this is explained in detail somewhere, I've been searching FAQ's and haven't found much of anything

[Tyler Christensen]

Typically you would either run a metal bellows hose to the pump or a rubber reinforced tube. The rubber tube could be hooked into a 2.75 or KF - hose adapter and clamped on with a hose clamp (http://vacuumshopper.stores.yahoo.net/stainhosad.html). A metal bellows would require plumbing at the pump end to most likely convert conflat to NPT depending on the pump specifics

[Nicker]

Pretty easy.

First step, figure out what fitting is on the roughing pump (KF16,KF25).
-page 1-60 of Lesker 9th edt. for KF sizing if you do not know.
( viewtopic.php?f=10&t=3809#p24448 )
Second step, KF(size) connected to bellows hose KF(size) to KF(size). (page 1-133)
-Don't forget the KF Centering rings and clamps for sealing. (page 1-69)

Third step, KF(size) to CF 2.75, mounted to your chamber. (page 1-105)




But if I were you I would do Roughing pump to Diffusion pump, Diffusion to chamber.

[John Futter]

Nicker
If you are going to advise please get the terminology right!
A mechanical pump behind whatever high vacuum pump is a backing pump ie it is backing up the high vacuum pump

Roughing pumps are what they say--- they are there to produce a rough vacuum that is within the range of a high vacuum pump to operate to continue to a higher vacuum.

Some systems use separate mechanical pumps to do both functions-- but most use valving to isolate the backing line for a short time so the same mechanical pump can be used to rough out the space above the high vacuum pump and then be returned via the valving to the usual backing function.
High vacuum pumps (turbos and diffusion) require a backing pressure of around 1 by ten to the minus ten millibar --- above this level they stall with internal contamination and contamination of the high vacuum space a real possibility.
I realise that you are saying to rough through the high vacuum pump but you forgot to say that if you use this technique the high vacuum pump must not have been enabled yet

ie diff pump cold / turbo not started.

one problem of roughing through the system is backflow of mechanical oil pump vapour into the high vacuum pump innards and also the chamber being evacuated. If you are going to use the same pump for roughing and backing then a foreline trap of activated alumina in the line to the mechanical pump is a must to prevent contamination of the rest of the system.

Brian the web is font of knowledge on vacuum technology as is this site, not all is covered in the FAQ's a proper read of the Vacuum technology forum should answer all your questions. If the web scares you there are countless texts on vacuum technology --search for " high vacuum techniques" Most here on this site have done this before asking vacuum 101 questions.
A good place to start with knowledge is the commercial vacuum suppliers such as Edwards, Varian, Phieffer, Leybold, Alcatel, and further research from system suppliers such as AMAT , and vacuum parts suppliers Kurt Lesker, (sorry Cindy Mental block), MDC, Vacuum generators, and a few others

[Chris Bradley]

John Futter wrote:
> High vacuum pumps (turbos and diffusion) require a backing pressure of around 1 by ten to the minus ten millibar --- above this level they stall with internal contamination and contamination of the high vacuum space a real possibility.
I hope you mean 10^-1 mbar, for a backing pressure!?



> most use valving to isolate the backing line for a short time so the same mechanical pump can be used to rough out the space above the high vacuum pump and then be returned via the valving to the usual backing function.
I was under the impression that you merely had to connect the things up in serial as a turbo and/or diffusion pump are just "through-volumes" when not running? I guess this is very relevant if one needs ultra-clean UHV turbo internals to stay that way. I'll take the advise as I have a few electromechanical valves I can rig up to be "A OR B". But if you run two valves like that, why does it make much of a different when backstreaming up the 'other' pipe can simply then run on into the 'top' of the turbo pump, rather than the 'bottom'?

This has become a bit more relevant to me as I appear to have acquired some turbomolecular bits to help push my tufnol/glass-cylinder arrangmement into unit microns. (I'm currently able to get 20microns out of it with an E2M2, which I think is doing pretty good seeing as the chamber is 0.5cu.ft. while the pump only has a 1.6CFM flow rate. But it takes a long time to pump down once it's into 10's of microns and I'm presuming a quick blast of a turbo pump will help out here.)

[Richard Hull]

Chris is correct. The bypass valving is just not needed at all in fusor work as we never go to high scientific vacuums. They can simply be serial volumes. There is zero need to hit 10e-6 torr in any fusion chamber. 10e-4 torr is more than adequate and 10e-5 is super. You are going to back fill at flowing 10e-2 toor, anyway.

Only two valves are needed in any fusor system. Both must act to simply isolate the diff or turbo pump.

I start by closing all valves and mechanically pump the foreline to about 10 microns. (takes about 1 minute on my system. I then open the valve to the diff pump and due to its fine valving and seals, it pumps to 10 micons in under 1 minute. I next open the valve to the chamber and it pumps to about 10 microns in 4 minutes, but once it crosses 50 microns (less that 40 seconds), I turn on the diff pump heater and cooling fan.

Within a total of 20 minutes from air, I am at a fusor pressure of 5X10e-5 torr and can let in Deuterium against a restricted valving off of the diff pump to a flowing D2 pressure of 10e-2 torr.

I have operated like this for 10 years now. I have had the same fusor IV online without tear down for 5 years. So the above works just great.

For the purest or for anyone with only one vacuum system that is used for truly HV work that also services a fusor, a foreline trap would be a must.

For most here, they have never used or seen a vacuum system. What they assemble will be solely used for a fusor system. Thus, there is no need for a foreline trap.

In my earliest systems through Fusor III, I used the micromaze foreline trap as my high vacuum pump!!! Old posts will show that fusion is quite possible exhausting to only slightly sub micron vacuums (6x10e-4 torr), afforded by a rigorously maintained micromaze. It is a bit wasteful of D2 gas, but so is normal running with flowing D2 and a diff or turbo pump.

All the above has been expounded upon many times over the years by me in many postings.

Richard Hull

[Doug Coulter]

Chris,

I believe e-1 mbar is the right number there, some things will take a little more, perhaps 10x e-1, probably a finger fumble. At least those are the numbers I and all my vacuum books use. Turbo-drag combo pumps are so tolerant the roughing pump can be a diaphragm pump, and that's what's in Pfeiffer's little leak tester pump station! Kind of cute, completely oil free, and quite reliable in my experience. I have a big piston two stage dry pump on my big chamber, because, hey, it's really big and otherwise it would take too long to rough.

For a turbo, which isn't spinning, you don't need to have the extra valving, as yes, you can just "pull through" it while bringing the chamber down. When it's time to bring it back up to room pressure, you'd like to slow the turbo gently by venting between it and the roughing pump, so as not to burn off the rotor tips by letting air, or whatever purer gas mix, on the high vac side while it's supersonic.
And if your turbos are anything like mine, you will want to have a way to "brake" them as they take forever to spin down from bearing friction alone - hours in HV. They spin up pretty fast, going the other way, however, and my controller, which also controls the roughing pump, sequences fine simply by limiting the drive power to the turbo -- it doesn't start getting fast until the roughing pump has done most of the roughing and so stays out of supersonic shock waves on the rotor tips during that process.
Eg, the motor alone can't burn out a turbo rotor, but once spun up, the flywheel energy in one surely can and it can make an expensive death rattle you'll never forget. There can be a pretty dangerous amount of mechanical energy stored in a flywheel going 800-900 rotations per second!

For a diff pump, the issues are very different. When they are hot, you just can't let that hot oil be exposed to much air at all -- fire! Even if not that extreme, most oils will decompose under those conditions. And, they take quite a while to cool the boiler when turned off -- so much so that the fancier ones have separate water cooling plumbing to do that if quick shutoff is needed. That part of course needs some valves so as not to water cool the boiler when running.

If you want to be able to cycle from air to vac to air to vac more than about once a day, you need a way to valve off the diff pump entirely (eg valves on both ends of it) while you bring your chamber up to air to do things in it, then rough the chamber with the roughing pump while the diff pump is still valved out of the system, then finally open the diff pump valve to the chamber and to the roughing pump - that way, you can leave the diff pump hot the whole time and be ready for it run right away again. This has to be done with some care, as a quick air inrush through a diff pump will deposit its expensive oil in the roughing pump -- and since I've not yet seen one with a sight glass, I suspect what happens next will be confusing and waste some time as well as that money. So a diff pump system needs at a minimum 3 valves to work quickly when you are testing things and need to cycle the chamber a lot. One is roughing pump to chamber, one is roughing pump to diff pump output, and the last is diff pump to chamber. The first two can be fairly cheap ball valves, even, that last needs to be a big one, the kind of thing you try to find on ebay rather than buy new. We are using a gate valve there, full bore size, in our diff pump system. This, along with a light dimmer on the boiler, can be used to "throttle" the diff pump if you want to let gas flow through the system, but not waste a ton of gas keeping a certain pressure in there. Even smart guys make mistakes, and they are costly here if you're using good diff pump oil, so we made checklists on the operations and their orders to keep air off hot diff pumps, and avoid quick rush throughs. So far, we've been careful enough and not had accidents.

The turbo is better that way, just turn it on and off, and manipulate a leak valve (we use a tiny needle valve there, cheap) to bring the system up and down. It's faster and more convenient to cycle often, which is probably its main attraction in fusor kinds of work -- ours also give much better ultimate vacuum, which does help any outgassing go a little quicker when cycling, and gives us more confidence we're running only in the gas we let in on purpose. Most turbos can throttle by setting a lower rotor speed in the controller. For a fusor, running it helps the tank outgass better than almost anything else ever could -- those hot ions and electrons hitting the tank walls eject adsorbed gasses with alacrity. It does a faster and more thorough job than a couple of kw of quartz heaters in there unless those are on for longer than my patience allows. In a non fusor system, it'd almost be worth it to put a small fusor like thing in there for helping with that.

When I'm on a vicious cycling loop, like trying several things a day, I use welding argon to bring up the tank, only have the door open a little while for whatever change I make, then right back down. This speeds up the next pumpdown by a factor of several, as the argon doesn't stick to the tank walls like the more reactive shop air molecules and water do, and the tank then outgasses quick.

[David Rosignoli]

Richard,

So, when you first open the valve to the main chamber, the inner chamber is at 1atm, and the foreline is in the microns level? Doesn't that expose the diffusion pump to a higher pressure than intended? Is there an advantage to doing it in your sequence versus connecting the mechanical pump to the main chamber and pumping it down first, then valving in the diff pump, by connecting it to the foreline? Or have I misunderstood you.

Dave

[Doug Coulter]

Richard's method does and will work just dandy for the operational type of thing he's doing -- he's not in and out of his fusor several times a day -- he just runs it, and his system is fine for that, and simpler and cheaper in the bargain.

If you want to cycle a lot, to try a lot of things, then the more valve system is better with diff pumps.
Diff pumps don't mind atmosphere when they are cool, some mind a heck of a lot when hot.
So, if you don't mind waiting, you don't need the valves.

[Richard Hull]

I just posted a full FAQ in this forum on the startup-operation and shutdown of a fusor using either a diff or turbo pump. All is made manifest there.

Dave did misunderstand. The entire system is pumped to 10 microns with the mechanical pump before any turbo or diff pump is turned on. Again, refer to the FAQ.

Richard Hull

[Chris Bradley]

Rather than 'pollute' the FAQ, I thought I'd put this question here:

A turbo pump doesn't work until its working pressure is backed off to under 0.1mbar. This we know.

My question is about 'safety' matters in terms of ensuring a turbo pump survives. If a pump is full speed in a free molecular flow regime but is exposed to the higher pressure of a viscous flow, then clearly it will suffer damage.

But if a turbo pump is turned on, from stationary, at atmospheric pressure then is there any particular consequence? I am presuming that the inductive motor parts, which are actually quite weedy but work by slowly winding the rotor up in vacuo, would simply struggle to push the rotor around. In this state, I can't really see what damage would be caused to the rotor as I presume it wouldn't spin up very quick, though I would not 'try it out' because of the excess *electrical* load that the electronics and windings might experience.

So, in summary, if you accidentally switch a turbo on at >0.1mbar, is there any real consequence?

[Carl Willis]

No.

[Chris Bradley]

So if you're impatient, then it sounds safe to turn it on around 1 mbar and wait until it 'catches' the right pressure as the backing pump draws the foreline down.

[Doug Coulter]

Carl, that's a new record for concise! I must learn that.

I have a couple of fairly new Pfeiffer turbo systems, and that is what you do -- just turn the whole mess on.
The motor just won't put out enough power to make the turbine spin up much until the roughing is mostly done with. This is by design in that brand, and the electronics are designed to limit motor power gracefully (you can even set that limit in the software in the controller if you like).

It works fantastic, as near the end of roughing, most roughing pumps get kind of slow (see their performance graphs at their low pressure limits) but as this lets the turbo start spinning up for real, they start to have a "compression ratio" at that point and even the roughing part gets sped up significantly. As nothing is ever perfectly matched, you can actually hear it in the exhaust note of the roughing pump when the turbo starts actually helping. I have set up my systems so that the roughing pump is turned on and off by sensing the turbo drive power -- it only runs when it would help, which saves power and life on that pump. This allows me (on solar power!) to leave them on overnight etc, so I spend a lot less time waiting for tanks to outgas. Even a big turbo may only draw 20-30 watts to stay running in a good vacuum, which I can tolerate here.

That number for what a turbo needs as backing "pressure" is *extremely* variable depending on whether you have a straight turbo, or a compound (usually a turbo-drag). In the latter case, it can tolerate 10 millibars and still make ultimate vacuum, or very nearly! This is covered in the tech notes in the Lesker catalog fairly well.

I therefore propose a little clarification when we talk of these, say straight or compound turbo?

Because the numbers really are way-way different for those. A compound turbo can be backed by a diaphragm pump, that looks like what you'd see in an aquarium store, a straight turbo needs "real" roughing. You almost cannot buy a straight turbo new these days, but a lot are on the surplus market.

Basically, the compound type costs so little extra to make and it is so much better (and saves on system costs as you no longer need such a good roughing pump), no one bothers with the straight ones anymore unless they need an exact replacement for an existing system.

[Chris Bradley]

I seem to recall reading that Varian once made a turbo[like]pump that worked pumping straight from-and-into atmosphere. Not sure how true that is but I can't see any real reason why you can't have a mix of free molecular flow-type rotors at the top, and viscous axial compressors at the bottom.

Doug, what is the difference, compound/straight?

[DaveC]

There's no real damage done to a turbo turning it on at atmospheric pressure. Most system controllers have a "low speed" mode, which is intended for use when a system is leaky or has high foreline pressures.

The contamination issue is a greater concern for applications where a very very clean vacuum... no HC's or water is required. It takes a while to get the rotors scuffed free of their contamination. Flushing with a measured "ballasting" dry gas is one recommended method to clean up.


Dave Cooper

[Carl Willis]

I recommend reading your pump's manual.

Starting at high pressure accidentally will not kill the pump. Whether or not it makes sense to do this is another issue.

On many pumps, if you start up at high pressure, the controller will shut down automatically as it senses the load at low RPMs and so there is no point in trying to get a head start just to be expeditious.

On other pumps, coolant must be supplied to the motor and / or bearings if the pump is going to be started up or run at a high load.

The real reason to have isolation around the turbo (or diff pump) is for when you are returning the chamber to air. You can leave the high vacuum pump running, do your business in the chamber, and then rough it down and return it to the high vacuum pump at once. Otherwise, you have to go through the slow process of venting the turbo / allowing the diff pump to cool down (the only exceptions to the slowness are on giant pumps that have emergency cooling or braking systems). For small systems isolation of the pump matters less or not at all. On big systems it's often worth the added complexity and cost.

-Carl

[Chris Bradley]

Carl Willis wrote:
> I recommend reading your pump's manual.
Manuals... Indeed, a rare treat with ebay purchases! Well worth reading if you get one: Probability of getting one..?

[Carl Willis]

Did you try getting a manual and have trouble? What pump are you thinking about running in the manner you described?

[Chris Bradley]

Yes.

I'm looking at acquiring a Balzers/Pfeiffer as the controllers and pumps seem more commonly available than other. I came across a website [http://www.ptb-sales.com/manuals/balzers/] which has quite a few, but is not comprehensive and I found no other sites as openly helpful. If you have links, please say them.

Anyhow, the instructions aren't that helpful. They seem to rather presume you have an integrated system and it simply says 'turn everything on and the turbo will start automatically'. I am rather presuming that this is an instruction set for a well-installed system and not for some dodgy-bodge job by an amatuer (with some dodgy-bodged parts from ebay) where not everything might be connected in the system.

Have I done enough prior searching to satisfy you, Carl, or perhaps I shoud've done a two year night-school on vacuum technologies before justifiably asking the questions? (Not sure if that is an over-reaction, I can't read your in-between-the-lines words these days.)

[Richard Hull]

I find it stunning that a simple mechanical pump down is just not done out of hand prior to turning on a turbo. It is all too easy to do. You'll not come on line any sooner.

The sluggard turbo is not going to do anything real until the mechanical pump has done its job.

More much ado about truly nothing.

If I were ever to put my large Temescal turbo online here, in my system, I would be very paranoid. They aren't cheap and a mistake is a pump life ender. With a diff pump slip up, you just have a nasty clean up job on your hands. (zero dollars to fix).

Richard Hull

[Doug Coulter]

Richard,
It's just the default in both my Pfeiffer integrated systems, and having tried it both ways, it is a little faster, nothing to write home about -- takes a 6 minute process to about 4 min in the big tank. Doing it that way means the turbo starts helping as soon as it actually can, rather than at some imaginary number you think you have to reach first, and which is different for every turbo design anyway. There's no hard and fast number you can tell people on that one, it differs with turbo design details, relative sizes of the pumps to each other and the tank etc. It's different for my two systems, one a 65l/sec turbo, one 512L/s. So they just take care of that for you. It makes more difference on my small pump/tank system because the roughing pump on that one is so lame -- only gets to a couple mbar by itself anyway, and is tiny, so it needs all the help it can get from the turbo.

Yes, running a turbo can be scary, I've almost lost one due to inrush accident. You do run screens over the intake so if something should shatter in the tank (had that happen) it can't eat big pieces and die from that. The owners manual equivalent of an MSDS is fairly scary -- one of those rotors can store enough mechanical energy to be fairly dangerous -- even shear off all the mounting bolts and throw the thing, according to them.

Worst I've had so far was a small inrush (a couple cu in at stp before I stopped it, operator error) and the thing made the gawdawefullest noise I've ever heard as all the rotor tips were hypersonic in the new pressure for a little while, then it recovered as it pumped the gas while slowing down fast. The whole apparatus jumped a bit, and it's never been quite as perfect since, but still way good.

I would slightly dispute the diff pump zero dollar accident cost if you were running really good diff pump oil and it either went into the mech pump or mech pump oil got into it. Oil ain't free!

[Richard Hull]

Compare 50ml of KJLC704 oil (~$5.00) to a $3000.00 turbo. It is effectively zero dollars.

Oh! By the way... We are forced to purchase our first fill for the pump in a 500CC bottle and suffer the $84.00 price. That was up front money to just use the pump on day one!

After that, I can blow up my diff in 9 more stupid accidents, complete system teardowns and associated cleanings without spending another dollar. I clean and reuse all my copper gaskets as I do not crush them out of existence on first use. I can get about 3 re-applications on the same gasket.

With a turbo, you would need $30,000 worth of up front spending to not suffer out of pocket new costs in 9 destructive events.

Turbos are massive dollar losses in waiting to all but the hyper vigilant. You can't perpetually expect to get a fully functional turbo with controller and cabling that will bolt right onto your present system off e-bay on the cheap following an accident.

I can have my exact, same diff pump back on line the same day at no additional out of pocket expense.

Again, I have a 10" throat turbo that works perfectly with all cabling and controller in storage here and refuse to use it until the right project comes along that might demand its use. The fusor is not one of those projects. I also have a 3.5" and a 5" diff pump on hand, but they are water cooled and over kill on a fusor as well. If you have a 2.5" air cooled diff pump on hand that is the one you need.

Turbo's are for people who are in and out of their systems constantly. This vastly hazards the continued existence of same due to far more operational related errors in the probability mix per unit operating time.

All things have their place and their price points. Most applicants here are very poor neophytes and vacuum based "babes in the woods".

Richard Hull

[Chris Bradley]

Richard Hull wrote:
> I find it stunning that a simple mechanical pump down is just not done out of hand prior to turning on a turbo. It is all too easy to do. You'll not come on line any sooner.
> The sluggard turbo is not going to do anything real until the mechanical pump has done its job.
> More much ado about truly nothing.
Speaking for myself, this stuff is relevant and germane. As you say, why bother? But discussing the full enevelope of what is and what is not permissible and possible gives [in any situation] a deeper understanding. If we operate within the confines of manuals and procedures all the time that are given by others and we never seek to ask or wonder whether some other modus operandi might work, or why it might not, or why it might be damaging or unsafe, then we are just thoughtless automatons repeating something to a pattern. Some gain amusement and value from doing exactly that, but my own interests don't lie in doing anything in 'the standard' way; I want to learn *why* it is done in the standard way [which often involves asking "why not do it ..this.. way?"] and then to thereafter stretch the envelope by doing something *not* done. This is a trivial example of that tenet, but is served by it.

Doug's comments regarding the higher start-up pressures and greater flexibility for some turbo designs is particularly interesting, from an understanding of what might be practically built up into a working system.

[Carl Willis]

If I have a piece of equipment that costs more than my car, and I want to use it properly, I read the manual.

Chris's original question about killing a turbo by turning it on accidentally at high pressure has a simple answer: no, you won't kill it.

Chris's extrapolation from this about turning a turbo on at high pressure if you're impatient just doesn't follow. At all. Most of these pumps (and I have no choice but to defer to the specific pump's manual for details) won't abide that condition, either because of an overload trip or an overtemperature trip. Again, you won't kill the pump! But this is not a recipe to save effort or simplify your job. It's a recipe to have to reboot your pump a bunch of times, possibly apply water cooling during spin up, or get shut down for longer because of an overtemp trip.

The whole episode here is just illustrative of a certain tendency to extrapolate information from a response that just isn't implied. And it annoys me that this happens. Chris, when and if you do get a turbo and get to know it, I'm confident (in spite of your comments here) that you will have no problem treating it right and finding the right balance between expediency and good operating practices. I'm also confident that you will locate its manual.

-Carl