Some Diffusion pump Questions

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.
Post Reply
Lukas Bochtler
Posts: 9
Joined: Sun Jun 25, 2017 3:13 pm
Real name: Lukas Bochtler

Some Diffusion pump Questions

Post by Lukas Bochtler »

Hello all,

So as the title states i have a few questions about diffusion pumps. I have most of an electron microscope, that i would like to get working again. some of the things missing so far are the roughing pumps, and the diffusion pump chiller (though i do have a manual for a very over the top chiller that original ran the microscope, but seeing as the original one is specked at 25KW cooling power, i suspect it is massive overkill, seeing as the microscope has a peak power consumption of 5.5 KW).

So, the pump is a 4" diffusion pump, with a 500W heater, model should be (i need to check the actual model installed, but based on another copy of the microscope i have, thats the one that should be in there) DPF-4Z by DAIA.

Based on the label, it has a 500W heater, and based on the manual of the microscope, it needs 1.5 L/min at 10 to 20°C for the water supply.

So, the first question, how much cooling power do i need for the chiller, since i dont have the exit temperature, i cant see a way to calculate the power the pump dissipates into the cooling water. I really dont want oil backflow into the column, so id rather go overboard in terms of cooling, as apposed to having to clean the column. (and possibly replacing the gun filament).

So far i intend to use a 790W aquarium / PC cooling chiller to cool the pump, but im not sure if its overkill or not.

The next question is, what size roughing pump do i need for the pump to work normally, and is there a possibility of using too large of a roughing pump. The diffusion pump is specked at 570 L/ sec pump speed.

The original vacuum pumps (which im still trying to track down, but as time marches ever onwards, the hopes of finding them intact dwindles.), where 145 L/sec pumps that go down to 10^(-2) mbar (if i recall correctly).

If i cant find the original pumps, i am out of luck, since the only ones of the same model (based on guesswork mind you), are not up to spec anymore, hence i would not buy them. So, i was thinking of using a pair (yes it needs two pumps), Leybold Trivac D16B pumps, one for the diffusion pump, and one for the sample exchange chamber. There is the problem of higher pump speed, since the trivac pumps about twice as much gas per second as the originals, as well as going down deeper then the originals.

So, my question is basically, will the D16B's work instead of the originals, or will this create problems with the diffusion pump?

Another thing, i am wondering weather or not i need to change the oil in the pump, the microscope has sat unused for a decade (apparently) before i got it. So, i dont really trust the pumps oil anymore. I was thinking of using Leybold's diffusion pump oil, but am unsure weather this is the best idea or not. (heck i don't even know how to change the oil in this pump, but im working on tracking down a service manual for the microscope, which might help).

Also note: I have a Leybold Trivac S4A pump that needs an oil change, but i highly doubt it will be enough to run the microscope. (also i need it for other projects)

Thank you in advance for any help i receive here.
John Futter
Posts: 1848
Joined: Wed Apr 21, 2004 10:29 pm
Real name: John Futter
Contact:

Re: Some Diffusion pump Questions

Post by John Futter »

Lukas
500Watt heater = 500w heat load so all is calculable from there. It also depends on what the electromagnetic lenes in the column require in input power usually these are water cooled as well for stability (mechanically) of the column
There are only a few Diff pump fluids that I would use
Dow corning DC704 or DC705 or Santovac
Santovac is extremely good but eyewateringly expensive while the DC fluids are cheap especially from China

Better mechanical pumps are totally OK the system will pump down faster
ian_krase
Posts: 636
Joined: Mon Nov 28, 2016 2:48 am
Real name: Ian Krase

Re: Some Diffusion pump Questions

Post by ian_krase »

You might need more cooling than that. The water flow rate is the immediately important part, and big reserviors can buy time, possibly lots of time. You can see if the temperature rises though at 20C you will need active cooling. (is slightly warmer water OK?)

AFAIK there is no too-big forepump. Just make sure you can develop the good enough vacuum at the gas load - possibly a problem if big tube manifolds are needed to adapt pumps.

I don't see why you couldn't use a single pump with valves and if the loadlock isn't diffpumped you might get away with much smaller pump.


All three popular diffpump oils: Octoil, DC704, and Santovac 5, have lots and lots of equivalents/offbrands/oem versions.
John Futter
Posts: 1848
Joined: Wed Apr 21, 2004 10:29 pm
Real name: John Futter
Contact:

Re: Some Diffusion pump Questions

Post by John Futter »

I do not like Octoil.............

a disaster if you get something wrong
Lukas Bochtler
Posts: 9
Joined: Sun Jun 25, 2017 3:13 pm
Real name: Lukas Bochtler

Re: Some Diffusion pump Questions

Post by Lukas Bochtler »

Thanks for your replys sofar.

I have taken a look at the manual, and i dont see any mention of water cooling the electromagnetic lenses. Ill have to check the actual unit to be make sure i have not missed something.

As for the oil, the price of the Santovac oil, it is indeed eye wateringly expensive (more then i had expected). Id take the plunge if i can get only 150ml (which the pump needs) but its still expensive in comparison to Dow Corning.

Ill have to see the price for the leybold stuff, though im not sure if it would be better then Dow Corning ECO-705 (which im currently considering).

Also, what are the differences between Santovac 5 (im guessing) and Dow Corning ECO-705? Also dose anyone have experience with Leybonol 5xx oil?

While im on the subject of oil, what oil would you guys recommend for Leybold Trivac pumps (since at this rate ill have 3 in the microscope room, 2 for the microscope and 1 for the sputtering machine). So far i found the Leybonol 300 oil to be the one recommended, are other oils recommended?

As for the chiller, im currently thinking of using the Hailea Ultra Titan 1500 chiller. According to Aquatuning it has 790 Watt cooling power, so it should cool the pump and trap.

for those wanting to research my microscope (good luck with that) i have a Hitachi S800 FEG SEM
Jerry Biehler
Posts: 975
Joined: Tue Nov 24, 2009 3:08 am
Real name:
Location: Beaverton, OR

Re: Some Diffusion pump Questions

Post by Jerry Biehler »

Usually you use something around the size of a Edwards RV8 on a SEM.

Dont worry about the oil, it is fine.

Simplest thing if you dont plan on using it all the time is to just use tap water and just dump the water down the drain or something, 1.5lpm is nothing... Do put a regulator on to heel the pressure down if you have high pressure locally.

My SEM only needed water for the DP, one of the reasons I ditched the DP and went with a turbo. Some SEMs use water to cool sections of the column as well as the sweep electronics so that might not be a viable option for everyone.

For the roughing pumps just use generic rough pump oil from duniway or whoever.

BUT! Here is the main thing, you have a field emission unit which means this has ion pumps in the column because field emitters need really high vacuum. You dont want to spend a lot of money on this machine until you get the column pumped down and the ion pumps running. It may take several days for the column to get to pressure before you can find out if you emitter is shot or not. If the emitter is shot you are looking at a few thousand dollars for an emitter. Thats one of the advantages of a thermionic emission SEM.

So, get or borrow one roughing pump for the DP, ignore the load lock. Run a water line to the DP and get it running and pumping down. Now you get to try and get the column down, there are probably bypass valves to open the gun chamber to the diff pump, you cant leave this open too long otherwise you will get oil contamination in the gun section and really mess things up. You basically want to run this just long enough to get the ion pumps running. Once thats going you can let the machine pump down. You may have to bake the gun section too. I am not sure.
Lukas Bochtler
Posts: 9
Joined: Sun Jun 25, 2017 3:13 pm
Real name: Lukas Bochtler

Re: Some Diffusion pump Questions

Post by Lukas Bochtler »

Sadly i don't have the luxury of tap water where the SEM is located, hence why im thinking of using a chiller right of the bat.

I do need to bake the machine, the startup (well most of it) is described in the manual. It requires 8 hours of baking (the gun chamber as well), the vacuum system is fully automated as well, though the mentioned bypass valves are fully manual, Annoyingly the exact procedure is not documented in the manual (if i recall) but it might be in the service manual, which i am trying to track down.

As for the emitter, well i don't even know where to get a new one from, other then getting the electron gun off another S-800 or similar microscope. But, according to the previous owner, the microscope worked prior to its removal (and subsequent damage to the cables). I am quite certain the gun chamber is at atmospheric pressure now, since the bypass valves where slightly open, with the load lock also being open.

Spending about a grand on getting this thing even to the point of doing what you states is unavoidable, due to all cables from the control console to the column where severed (something the original owner only found out after the fact, and was furious about). I am currently also looking into building a new ion pump power supply, since i have yet to locate the original, the only expensive part in that, are the half dozen custom transformers needed to make it work.

But, on the bright side, i was going to build one from scratch anyway, so i might as well restore an old one first. (the old one will probably out preform my own design).
Jerry Biehler
Posts: 975
Joined: Tue Nov 24, 2009 3:08 am
Real name:
Location: Beaverton, OR

Re: Some Diffusion pump Questions

Post by Jerry Biehler »

The automatic cycle is for actual use of the machine when everything is up and running. You have to get the ion pump down to about -4 to -5 before it will start. Hopefully it will.

Don't try to build an ion pump supply. There are ion pump supplies all over eBay. There's nothing too special about them, helpfully there was no communication between the supply and console that can't be overridden.

Emitters can be gotten, it's just they are VERY expensive. There is a guy that has been trying to make his own and having so-so luck with it. You're making a tungsten tip with a tip size of about an atom. Real easy to mess that up.
Lukas Bochtler
Posts: 9
Joined: Sun Jun 25, 2017 3:13 pm
Real name: Lukas Bochtler

Re: Some Diffusion pump Questions

Post by Lukas Bochtler »

Well there is a small problem with the use of a non S800 ion pump power supply, the power supply also contains the baking power supply, as well as controller / timer. Also the control console seems to send some voltage references to the ion pump power supply controllers (just 9 op amps). (though i need to do a proper trace threw the schematic book to be sure)

So, here is my plan:

- Try and find the original power supply (it vanished together with the roughing pumps and air compressor)
- Fix the cables (ill be going to my microscope tomorrow to get the high voltage connectors and the little bit of cable left between them, and will be sending a inquiry to a German company specialized in fixing such things. (though they usually fix X-Ray equipment) I will also note down the cable specifications and connectors to make new ones.
- Verify the console fires up once cables are fixed (not sure if this is a good idea without the roughing pumps)
- Get a roughing pump (will probably go for a Trivac D16B, as i like Leybold gear)
- Get the above mentioned chiller (or anyone that will dissipate 500W or greater), as well as a small pump and some plumbing to get the flow rate to 1.5L/min (im thinking of a bypass valve on the main line, that will circle the water back to the tank without it running threw the pump.
- If obtaining the original power supply failed, build a new one.
- Pump down the column and bake it .
- Start the microscope and see if the emitter works
- Possibly take a quick image of the stage... (using a DSLR on the high resolution screen)

Though, i have a vacuum pump, (well its in the mail right now), its a Leybold S4A (it should be able to pump pump 120L/min) do you reckon it will be enough for the first pump down test? Since the originals where 145L/min im not sure...

edit:
Oh and one question, do the ion pumps need to be on all the time? Or can they be shut off when the microscope is not going to be used for a week or so? (i have no problem either way, im just curious)
John Futter
Posts: 1848
Joined: Wed Apr 21, 2004 10:29 pm
Real name: John Futter
Contact:

Re: Some Diffusion pump Questions

Post by John Futter »

No
You leave the ion pumps on 24/7
When down to the correct vacuum they only consume a few watts.
I have a detector chamber that sits in the low ten to the minus nine and it consumes less than a watt to keep it there @ 10kV ion pump potential
Lukas Bochtler
Posts: 9
Joined: Sun Jun 25, 2017 3:13 pm
Real name: Lukas Bochtler

Re: Some Diffusion pump Questions

Post by Lukas Bochtler »

That wont be a problem.

I was at my microscope today (its not in my house) and found out some interesting things. First off, the microscope still contained a specimen in the sample chamber (i looked in threw the load lock), second the data recorder camera still had a roll of film in it (thanks to me being familiar with medium format cameras, i retrieved without damaging it in any way). So, my guess is that the last time it was used, the person running it left in a hurry, or forgot his sample. Anyway, i will have the film developed and am curious to see what, if anything is on it.

Some other things, i went to the microscope to take some measurements, and check some things. I have taken the broken cathode cable with me, in order to have it potentially restored by that aforementioned company. Though the column connector seems to be potted as well (which is a bit of a bummer). But possibly the company can depot it, alternatively i can make a new one, as its some basic turning in plastic. (not sure which plastic yet though, My gut tells me its PTFE, which explains how i can pull part of the potting out by hand (and some genital force).

Some other things, the ion pump power cables are all severed, they are coax style cables terminated in the plugs seen in the picture. Im not sure if i can pull them or not. Im afraid that ill break the cathode or something if they are vacuum tight or something.
The 3 Ion pump power connectors.
The 3 Ion pump power connectors.
I have measured the cables to be 8.2mm OD, with 6.8mm OD HV insulation, and a 1.6mm diameter conductor (im guessing 14 AWG) made of smaller conductors. Still working on getting the original power supply...
The Ion pump cables
The Ion pump cables
Also good thing i checked the diffusion pump, its not 500W its 700W. the second image is the name plate of the diffusion pump
Diffusion pump
Diffusion pump

I also have this oddity in the column, and im not sure what this is, or what to connect the pipes to:
Unkown part of the SEM
Unkown part of the SEM
Fascinatingly enough, the 2 roughing pumps where right next to the column, something i did not expect. (based on the vacuum hoses, which still had the clamps on both ends, and the hoses being very short.)

Also, i found out that the microscope has a after market vacuum gauge on it, a Leybold PR25. I do recall having a manual for a meter for it, but im not sure anymore. Any information on this gauge would be nice. As well as, with what controllers dose it work with?

Also, with this diffusion pump, how do i check the oil level, and how do i get the oil in, if it needs changing? I have not found any way to do either.

edit:
I have another question, will this pump be enough to pump down the column to the point of me beeing able to test the gun? The original pumps where 145 L/min pumps, where as the one in the image is about 120L/min. (Its a Leybold Trivac S4A)
Leybold Trivac S4B
Leybold Trivac S4B
Though running the pump will be difficult from the microscope. Since the originals where 100V AC, and this is 220V 3Phase. I think i shall use a transformer to drive a AC relay to connect the pump to power.
John Futter
Posts: 1848
Joined: Wed Apr 21, 2004 10:29 pm
Real name: John Futter
Contact:

Re: Some Diffusion pump Questions

Post by John Futter »

The ion pumps look like standard Varian type connectors
You just pull them in the direction of the cable exit to undo
The cable is standard RG8, RG213 COAX any ham would have some this cable we use at work up to 70kV so the 10kV your ion pumps use is well inside it5s spec
Lukas Bochtler
Posts: 9
Joined: Sun Jun 25, 2017 3:13 pm
Real name: Lukas Bochtler

Re: Some Diffusion pump Questions

Post by Lukas Bochtler »

Thanks for the tip that the ables are RG-213. I do wonder, will it be a good idea to try and replace the cables by trying to remove them from the plugs, or just attach a high voltage BNC connector to the cable ends, and connect the cables to the power supply that way?

I fear that if i try and remove the cables from the ion pump power connectors (the one on the top pump is larger btw, and seems to be screw in as well). i will not get the new one in, unless i build (or buy, if thats even possible) new connectors.
Also, i do trust you with your voltage rating, i will call my favorite cable manufacturer (Sommer Kable) to see what they think about it. (they make a RG-213 cable)

Also, the pumps in the S800 appear to run off of 3.6KV. (a 1.8KV 500W transformer (well 500W fuse anyway) connected to a full wave voltage doubler. (with quite frankly an absurd amount of monitoring to boot).

Also, what am i missing with the ion pumps, why would one use a coax cable to power them, when the secondary electron detector (i think) is powered by a standard 20Kv unshielded cable. (well im guessing the accelerator is, not the PMT). Dose it have to do with voltage stability due to interference? Or is it to supply a better ground return path then the ground straps of the column?
ian_krase
Posts: 636
Joined: Mon Nov 28, 2016 2:48 am
Real name: Ian Krase

Re: Some Diffusion pump Questions

Post by ian_krase »

Coax cable provide a safety ground. Intrinsically safe, and for modestly high voltage good coax is often much cheaper and more available than High voltage wire.
Jim Stead
Posts: 94
Joined: Wed Aug 19, 2015 11:44 pm
Real name: Jim Stead

Re: Some Diffusion pump Questions

Post by Jim Stead »

The "oddity" in your column looks like a cold trap to me. Based on the connectors, I would think the left side supports an LN2 feed, and the right side hose is the vent.
John Futter
Posts: 1848
Joined: Wed Apr 21, 2004 10:29 pm
Real name: John Futter
Contact:

Re: Some Diffusion pump Questions

Post by John Futter »

It is a cold trap to limit any contamination of the FE source.
So initial pumpdown would be rough out column fill cold trap with liquid N2 start diffusion pump pump until 10 to minus 6 torr then attempt to start ion pumps keeping diff pump going until vacuum starts getting deeper
I presume you can isolate the diffusion pump when the ion pumps are working
Jerry Biehler
Posts: 975
Joined: Tue Nov 24, 2009 3:08 am
Real name:
Location: Beaverton, OR

Re: Some Diffusion pump Questions

Post by Jerry Biehler »

Like the other have said, cold trap. Mostly to protect the lenses from building up oil vapor which means the column needs to be disassembled and thats no fun on a FESEM.

The pumps were not next to the column. There were two weighted isolators that go inline to the pumps. Basically a piece of pipe though a chunk of cast iron or concrete. It dampens the vibrations from the roughing pumps that are transmitted through the hoses.

There is a reason for a lot of stuff in the ion power supply, the supplies not only run the pump they measure the current on the pumps which is proportional to the pressure and control lock out to make sure HV is not applied to the emitter under insufficient vacuum. And yes, the ion pumps stay on all the time. A lot of FESEMs have battery backup for the ion pumps that keeps them going during a power outage or even moving the machine.

You may have a very hard time getting these pumps started. Honestly, without the proper resources or lots of money this machine is going to be boat anchor.
Lukas Bochtler
Posts: 9
Joined: Sun Jun 25, 2017 3:13 pm
Real name: Lukas Bochtler

Re: Some Diffusion pump Questions

Post by Lukas Bochtler »

A quick update on the pumping down of the microscope. I have managed to make contact with someone who knows this microscope well, and he stated some interesting things that i thought would be good to know.

The microscope needs both roughing pumps present to begin pumping down the column, this is due to the initial pup-down of the sample chamber and column being done via the Load Locks pump, once itd down to the pressure needed to start the diffusion pump, it will start, together with its roughing pump. Once thats had time to warm up, it will begin to pull high vacuum. During the whole process the bypass valves are open allowing pump down of hte ion pumps. Once the diffusion pump reaches a high vacuum i can try to start the ion pumps. If they stay on great, if not, baking time, during which i try and start the pumps again. Once they start i should close the bypass valves.

As to the LN2 cold baffle, it appears i only need it when pumping down the chamber to the vacuum needed to achieve the microscopes maximum resolution, but is not needed when using lower resolution.

Also apparently, if i do need the LN2 baffle, the nitrogen lasts about 8 hours.

Apparently the whole pump down of the column will take roughly a week.
Lukas Bochtler
Posts: 9
Joined: Sun Jun 25, 2017 3:13 pm
Real name: Lukas Bochtler

Re: Some Diffusion pump Questions

Post by Lukas Bochtler »

I have another question about diffusion pumps and general vacuum systems.

I was told by someone experienced with vacuum systems, that i should not discharge the air from the vacuum pumps into the air of my microscope location. But instead discharge it threw some sort of exhaust outside the building.

So, i was wondering, how i should have that made (well what shape diameter etc..)


I was thinking of using a 1" (roughly) pipe (edit: Make that 2") (probably stainless steel or aluminum, in a L shape, pointing down, with a screen on the bottom (as to not have insects enter it). This would go threw the wall of the microscopes building (its not purpose built, there are other things in the room). I would then connect all 3 Vacuum pumps (the 2 of the microscope, and 1 from the sputter coater) as well as the nitrogen exhausts to this pipe, via multi T manifold connected via a KF flange or similar.

Would this be a good idea?

I would be relying on the pressure generated by the vacuum pumps and nitrogen exhausts to shove the air out of this. I don't really want to build a chimney for the microscope. If i need more pressure i can add a fan to push the air out. I would have this exhaust at around a hight of 1 1/2 meters, with a 25cm pipe going pointing down.
User avatar
Dennis P Brown
Posts: 3160
Joined: Sun May 20, 2012 10:46 am
Real name: Dennis Brown

Re: Some Diffusion pump Questions

Post by Dennis P Brown »

Roughing pumps can create an oil mist so many of us vent it outside. Any small pipe will do; yes, it will self exhaust by the roughing pumps - and when the scope is roughed out, the pumps will exhaust very, very little - really just a faint trace of oil mist. Except for roughing down the scope after it is at atm pressure, the vent line will never handle any real volumes of gas. I use an off-the-self flexible garden water hose (using a nice KF to hose adapter on the exhaust port of the roughing pump.)
ian_krase
Posts: 636
Joined: Mon Nov 28, 2016 2:48 am
Real name: Ian Krase

Re: Some Diffusion pump Questions

Post by ian_krase »

You can also get an oil mist eliminator. But ones that completely get rid of the mist are expensive and bulky. There's an faq on this subject.
Jerry Biehler
Posts: 975
Joined: Tue Nov 24, 2009 3:08 am
Real name:
Location: Beaverton, OR

Re: Some Diffusion pump Questions

Post by Jerry Biehler »

Virtually no one does not vent their roughing pumps into the rooms. You can get mist filters or make them. Simplest is getting a truck oil filter and putting a pipe nipple in the center thread where it screws on. Connect this to your exhaust. This works very well, better than some commercial filters I have used.

With a load lock you are not going to be pumping a large volume of air so you will not create much mist.

For basic testing you should be able to just use one pump and tee it off to the load lock and the DP. Save money. There is a very high possibility that this thing will never work and I have seen SEMs working with everything go for only a couple grand around here.
Post Reply

Return to “Vacuum Technology (& FAQs)”