I used to work at a potato chip factory. I don't know our power consumption in watts but we had two separate 3 phase 7200v feeds, and our power bill was around $250,000 per month. And the fryers were heated with natural gas! Anyway, every time there was a loss of power for more than around half a second (30 cycles), all of the control equipment would reset. Our process was only certified sanitary under "steady state", so we'd have to dump all food currently in process, and get back to steady state. This cost around $30,000 of raw materials (so much more in lost revenue). During some times of year this could happen several times per day. Given the short time horizon, and the massive power draws, management viewed it as an inherently unsolvable problem. Well. I found a backup power solution that had carbon fiber disks in a vacuum chamber spinning at 500,000 RPM. They could come online within five cycles, and provide power for 15 seconds - which would cover 90% of our outages, but was also enough time to start diesel diesel generators kept in a "hot ready" state (coolant and oil kept at operating temperature with external heaters). Anyway, based on historical power outage data, the flywheel system would pay for itself in 18 months just from raw material losses. I tried for over a year to get my proposal accepted, and the fact that management wouldn't take it seriously was one of the main reasons I left the company. This made me think of that story.
By "this" I mean you computing the energy of your spinning turbine. I don't remember the size of the spinning disks, but given ½mv² and all that (I can't remember the formula for it spinning disk off the top of my head but you get the idea), the company had focused on rotational speed versus size. They found that these insane RPMs let them store a lot of energy for a short period of time.
You may not be an engineer but you're definitely a mechanic. 13:58 the difference between an engineer and mechanic 17:43 never makes the same mistake twice, just slight variations of the same mistake 19:47 The mechanic slap of approval
Ah, thank you very much! That was another guess of mine, but I didn't think that this type of seal works in a vacuum pump. And I had no idea what to call this type of seal. Thanks for the information!
@@AdvancedTinkering I second this. It doesn't look like the type of labyrinth seals that are meant to keep gasses and liquids out but the kind designed to keep small particulates (especially magnetic ones that might get created by what ever process is in the vacuum chamber) that might get into the very close tolerance gaps in the electric motor of the turbopump.
@@AdvancedTinkering I am not an expert, but I did recently read the Pfeiffer "Vacuum Technology Book Vol 2" for a project. It quickly mentions labyrinth seals in their turbopumps. I believe it is to create a flow restriction for sealing gas injected into the bearing/motor area if you use corrosive process gas. But it could also be to keep the oil from moving into the main body, the book is not 100% clear what it is ultimately used for. It certainly isn't a vacuum seal, that's for sure.
@@AdvancedTinkering There is a youtube channel called AgentJayZ that has several years of videos covering jet engine maintenance.. He has mentioned labyrinth seals and probably has a video dedicated to them some place. The labyrinth seals for a jet engine are made of carbon to handle the high temperatures. The labyrinth seals I have seen on his channel would prevent axial flow, while what you have I would guess is a radial seal.
@@AdvancedTinkering I know this problem all to well, I renovate old machines myself and both having to store them and use them come with some space hogging requirements.
@@oliverer3 That, and wire EDM. Some pumps I've seen actually take a cheaper route for the stator vanes, they use stamped sheet metal instead with the vanes twisted to angle them.
@@imajeenyus42 parsons first steam turbine was pretty damn basic... but other than improved machining techniques, not much has changed since then... had the maths down first go, basically. expansion ratios and blade pitches, areas, that sorta thing...
Interesting pump, never seen one with active lubrication. Regarding the nut falling into the pump, it is a good idea to place some paper, lint free cloth or Al foil in any opening the instrument you are working on may have. That will keep small items from falling in, even the most dangerous: those you may not have noticed.
“ . . . even the most dangerous: those you may not have noticed.” Yes yes yes . . . ! I dropped a tiny screw while working on my mom’s engine. Never found it, but I had spares. I was also fortunate that when my moms ignition system shorted out from that screw bouncing around in her distributor, and it burned the primary ignition wire all the way back to the key switch, she was only about 5 miles from home. I hot-wired it so we could drive at home & I re-wired it the next day.
2 turbo pumps for a few cases of beer is one awsome trade. These were no doubt pm pulls but they would be easily good for hobby level deep vacuum experiments ❤
Hi, I'm a 360 degree science enthusiast and "taking it apart to see how it works" is one of my favorite things, as well as obviously designing new ones, the point is this, I'm so passionate about your video and reading the comments that I didn't realize that the tinnitus I suffer from 24 hours a day had disappeared the whole time! Now it's back but, think, would you have ever imagined that your videos were even healing!
I'm very happy to hear that you enjoyed the video and it distracted you from your tinnitus. I wish you all the best and hope it will vanish eventually!
i suffer tinnitus. i think its mental. hum the note back to yourself... it fades. i find theres several, work one at a time... ignore the traditional explanation of damaged hairs. its your brain trying to bias itself, get itself in calibration. damaged hairs maybe, but whatever, its more about having a base reference for "silence". for a teenage metalhead that blew a lot of things up, i still got well above average hearing in my forties...
You bring up an excellent point concerning the potential energy stored in a rotating mass. We had a failure of such a device (a GeneVac rotary evaporator) at work about twenty years ago and it was quite exciting. (The failure was entirely our fault, we overloaded the system with home-made sample cradles machined from solid blocks of metal rather than stamped from sheet.) Probably a similar amount of stored energy to your large pump (slower, but more rotating mass). The net result was a 2" (50 mm) steel shaft snapped, a 4" (100 mm) deep dent formed in the 1" (25 mm) thick aluminum case, the whole 250 pound (113 Kg) device did a 180 degree turn on the bench and jumped about a foot (300 mm) to one side, tearing out the 1" (25 mm) steel vacuum line that connected it to the vacuum pump (a liquid ring pump, quite a curious device in its own right). Also made a hell of a bang.
That is incredible! At first I marveled at your great memory. By the end, I realized that all the details would be emblazoned into my brain for the rest of my life, too.
@@berrigo2 Fortunately not, but I was involved in the initial installation and operation and got to see the corpse while it was still warm. Even more fortunately, I was not involved in designing the modified cradles or approving their use. In the aftermath there was a fair bit of back and fourth with the manufacturer and it was their engineers who determined that our cradles overloaded their hangers which was what actually failed, allowing the cradle to contact the case at very high velocity and bringing the other 11 cradles and (132) samples to an abrupt halt.
What a beautiful piece of machinery! Thanks for showing how it goes together. You had me laughing at the payment in beer; and also at the dropped nut twice!
The conical nut itself I believe to be a centrifugal oil pump. The oil is dragged around between the nut & the outside by viscous friction. Centrifugal force pulls the oil up in the conical gap.
12:54 - Was this turbine rotor CNC milled all in one block of material in one go or was each rotor disc CNC milled one-by-one separately then later assembled into a giant rotor turbine unit? Thanks.
I only know how the smaller models are made. But I suspect they are made the same way. They CNC mill each rotor disc separately and then assemble them onto a shaft using a shrink fit.
That was beautiful, watching it spin effortlessly, I totally enjoyed this video, pointing out the roughness of veins, I don’t know the sounds it made coming apart and back together it was just a beautiful thing thanks!!!
Can you imagine the paperwork of a lab engineer selling used lab equipment with possible health and safety implications to a private individual? From a university or research department that is entrenched in complicated paperwork but not set up for actually selling items? Hours of work for several people including the meetings and paperwork... Alternative B: Nobody wants them, so the alternative is getting rid of them properly. Slightly fewer meetings and hours of paperwork, but now the cost is significant. The professionals receiving them will want to know every material that has ever been used even near the thing. "Take them off your hands for a case of beer"? You bet that is AMAZING beer!
That's such a valuable information! Thank you a lot! I knew that the material plays a role in the torque spec but I always thought it is also highly dependent on the use case.
@@AdvancedTinkering its really about the clamp load. For example an m8x1.25 torqued to its spec (31ft lbs) has nearly 6000lbs of clamping pressure. So the total amount of clamp pressure required for a job is basically how many bolts you need for that specific task, and the torque rating is basically the max that the thread teeth themselves can handle with out being damaged. Thats why in aluminum will alsways be a bit less. Theres dry and wet torque specs too.. if you use a bit of oil, and then use dry torque specs you can and will start to snap bolts. Some bit of extra fun... if youre making something with threads, your thread depth / material thickness only needs to be as thick as the nut for that thread would be to acheive full clamp load. (You dont ever need to tap something m8, 2" deep for "extra strength".. yes the teeth def wont strip, but the bolt will strech and snap instead) Great chanel, we'd prob make good friends 😂 I should get back in the youtube game but try to be a bit less awkward about it lmao. You ask in some videos about your glass work and if you should leave clips in.. I think you should do a few vids on only glass work. How to make a hole, how to weld a joint, how to pre heat post heat, whole 9 yards... Theres lots of other "advance tinkerers" out there who might like to learn from you.. myself included. 👍🍻
14:42 the piece you put in is a seal mechanism, tortious path impedes movement of particles to go into the motor winding area or for anything else to come out of the area.
I had the opportunity to disassemble our KYKY TM pump recently as well. That one didn’t go back together as the internal water cooling channel leaked into the turbo… while it was at speed. Didn’t end well.
You got me beat by one size. I run the ATH1603M on my chambers. I did just put together a similar chamber to the video ending. DN200CF ATH1603M on a 200mm x 400mm chamber. With 5 DN63CF ports around. With an Ebara EV-A10 it roughs in 30 seconds and reaches 5e-7 torr in the 5 minutes it takes to get the pump up to speed.
Great to see such a big pump getting taken apart. Putting the big pump isn't a particularly ridiculous idea to me. It might be useful if you want to do all kinds of depositions etc where you want to be able to say that you only get exactly the thing you want to deposit onto your sample and everything else gets pumped out. But if you put it onto your chamber, you should consider the crash moment (you did work out the energy after all) and bolt your setup to the ground or similar. Nobody wants a vaccum chamber flying through the room.
That is one *honking* big pump! Turbomolecular pumps seem common these days, back when I was doing vacuum stuff back in the mid-70s, we used diffusion pumps for basic coating work, and molecular seive/ion/titanium sublimation pumps for ultra high vacuum. Turbos seem much simpler and higher-performance from a system perspective, I’m guessing turbos hadn’t been as well-developed at that time, so ion/TSP was a more effective combination in that period.
As for your last idea, I support it. Having a big turbo pump is a luxury. If you had limited funds and built the setup with new components, absolutely not, but if you have a big pump available, definitely go for it. Source: We happened to randomly end up with a similar setup at the load/lock of one of our systems. Pumps down in no time, and when necessary we can open the gate valve and pump down the whole system quickly that way. Nobody is in a hurry to get back to a small turbo. 😆 The risk of unfortunate things happening when you mess with the turbo yourself is your own business in this case! Let us hope your energy calculation was not foreshadowing. 😇
I am working on it ;) but I doubt it will be a permanent setup. Since the pump must be mounted upside down, I would have to lift it every time I want to open the chamber 😅 Maybe in the future.
I used to work in helium leak detection. We would evacuate a small roughly 12 cu ft stainless steel bell jar with two large Stokes pumps to a certain vacuum level before we could expose the Turbo's to the vacuum level in the bell jar. If we exposed the turbos to atmosphere while running, the compressor blades would disintegrate from the inrush of air and from a soft startup at atmosphere level they would never ramp up to speed (70 to 75,000 rpm i think. it's been a long time ago.)so they ran all the time and were operated by a valve that opened them up when the bell jar reached the right vacuum level. I too opened one up one time. It never ran again.
14:30 The molecular stage or Holweck stage is working on molecular drag on moving surfaces. It does not have the volume throughput of the turbo stage but it is able to pump against a much higher pressure. If you have only turbo stage you need a primary pump with an end pressure better than 5 Pascal but with the Holweck stage you can work with a base pressure of 100 Pa obtained with a membrane pump.
Please try to get videos with close up shots of individual components. It is usually quite hard to find pictures of functional components of these things. Awesome video, thanks.
Thanks! I'm glad you like the videos! Yes, those pumps are indeed very expensive. Without the generosity of their previous owner, I would never be able to show them to you guys.
Oil pump systems do usually discharge high pressure through a smaller diameter path than the larger low pressure path. For reasons including thermal expansion, aeration, and deceleration while collecting in the reservoir to reduce frothing. 17:26 bwahahaha the xylophone of unemployment. Holy shite you did it twice.
Wow those are the same size of the ones for purging a MOCVD chamber we baked GaALAs laser diodes. When they fail they FAIL ime, thankfully the most common is bearing failure, if the temperature so much as hints at going up, power down the pump or at least reduce the RPMs enough to finish your task.❤
Hallo, hier der Pfeiffer Techniker, hier mal ein bisschen Feedback zu deiner Demontage. Magnetlager: Die Unterlegscheiben sind sehr wichtig für die Funktion der Pumpe. Sie werden zur Einstellung des Magnetlagers benutzt und kommen von unterschiedlichen Dicken (0,8mm - 2,0mm) in 0,05 mm abständen zum Einsatz. Sie werden bei jeder Montage neu gewählt damit das Magentlager die korrekte Position zum Rotor hat. Es kann sein, dadurch dass du die Fläche beschädigt hast (bei deiner demontage) der Abstand nicht mehr korrekt ist und du dadurch höhere Vibrationen in die Pumpe bekommst. Rotor / Statorpaket: Der O-Ring ist ein Toleranzausgleich. Durch die vielen Distanzringe können sich Toleranzen der Ringe aufbauen, die das Gehäuse nicht ausgleichen kann und die Statorscheiben könnten nicht ausreichend gequetscht werden und auch hier Vibrationen erzeugen. Ölkreislauf: hier liegst du komplett richtig mit deiner Annahme. Labyrinthscheibe: Diese Scheibe zwischen Rotor und Gehäuse ist ein Gasschutz für den Motor. In dem Unterteil ist außen eine Schraube in der ein Schutzgas angeschlossen werden kann. Gerade wenn korrosive Gase gepumpt werden muss die Pumpe mit einen Schutzgas (z.B., Luft oder Argon) betrieben werden damit die korrosiven Gase durch den Druckunterschied nicht in den Motor kommen. Schrauben: Ich weiß nicht wie oft mir schon diese schrauben in die Pumpe gefallen ist. Du bist nicht alleine!! Ich habe deinen Schmerz gefühlt, vor allem beim zweiten Mal..... Wenn du weitere Fragen hast kannst du dich gerne melden.
The ridged vanes you call "molecular drag stage" I believe to be a simple serpentine barrier between the motor/bearing & the working volume of the pump. Normally a small flow of purge gas is applied to the motor section to keep "krud" from getting into the motor/bearing.
That's an awesome sound, the sound of the pump ringing with the nut falling through it, twice! Just kidding, that's the expensive sound nobody wants to hear. I love the sound of it running at the end though, I'd love to hear it what it's like, installed in a setup.
Turbo pumps have improved alot for this application. While not indestructable, newer systems are more caoable of handling some process upsets like occasional small qty droplet impact
Not the biggest I've seen by a long shot, the machines I work on have turbos that weigh more than I do! Granted they also are evacuating chambers large enough for several adults to climb inside haha. They work beside cryos running at around 10K. Working in a wafer fab is pretty extreme lol. Edit: oh and the pumps we use are mag lev on ceramic dry bearings, when your process is dirty dry pumps are much nicer, also helps reduce particles.
Those must be big wafers you are baking. Ours were 2.5cm gallium arsenide wafers for laser die making. It used two turbos about this size to pump down the chamber where it needed to be. ❤
That‘s not a big vacuum chamber. For large area PVD glass coating, vacuum chambers generally measure about 80-140m in length and 4m in width. Usually, there are 160-240 of these Pumps attached.
@@christopherleubner6633 eh older style 200mm wafers, my plant makes analog automotive chips like sensor interfaces and power transistors etc. We're a boutique fab so we handle a lot of the smaller batch special order items
@@xxportalxx. Not too low, as the quality requirements are not as high as in semi conductor industry. Base pressure is at around 8x10-7mBar (8x10-5 Pa), process pressure at around 5x10-3 mBar (5x10-1 Pa). On most of these Coaters, you have around 200 Turbo Pumps on the main chamber and 20 on the intro and outpus chambers. There are 40-80 Magnetrons with a length of 4m installed. Per Magnetron, the electric power is 50-150kW. So in total, someting around 4MW of electric power that is required.
The oiler makes me think of one-shot lube systems used on metalworking equipment. In those, the oil is lost rather than recirculating. I suspect the port you identified as for return oil is actually to equalize vacuum in the oiler so it doesn't have to work so hard. It probably injects a tiny amount of oil periodically - a small enough amount that it won't mess with anything downstream.
Oh this nice sound of the Blades and Vanes from a falling screw still makes me think "Oh F..." I've worked in the Aviation industry on Turbofan Engines and once, i've had a screw lost in the compressor. That chimes were beatifull and horrifing at the same time.
14:30 It's a dust trap I would say. The “up and down labyrinth” makes it more difficult for dust to get towards the bearing block 21:00 Same picture as mine yesterday. Only my pump is a little smaller. TMH521 with DN160. 13:10 I would like to see a turbo crash at top speed in slow motion. The kenetic energy must be incredible. We had a TPH2200 that suffered damage at high speed and fell apart into countless pieces
I once knew a research group that managed to drop a nut into a pump at speed. The rotor was basically a solid at that speed, so the nut just bounced back into a remote section of the chamber. THE FIRST TIME! The next time, the pump injested the nut, blades, etc.
Considering that rotational speed and diameter i think it is machined with insane nm tolerance, and such disassemble unfortunately disbalance rotor, even touching it by bare hand it deflect. I hope that magnetic bearing and hydrodynamic oil other side have some trickery to self balance for small deflection of main rotor.
@@AABB-px8lc I witnessed one that was cleaned professionally tear itself apart - $10K shot to hell - it was loud, but brief, but they replaced the unit. The tech probably nicked something while re-assembling. They never let us know what the cause was, but our MassSpec was down for nearly 2 weeks, right at the time we were doing our annual testing for the govt. I think it was an HP, but can't remember (30 years ago). 30,000RPM or so. I've had 14K RPM SCSI drives fail catastrophically as well, but it's more self-contained - the shrapnel didn't get through.
They are not vacuum pumps they are jet engines ha ha !!! On an more serious level !! The flow of oil is usually the largest hole is the inlet and the small hole is the high pressure side or out let !!!! Thay are huge and I bet they will quickly get a deep vacuum !!! Good luck with them !!!😊
14:18 it looks like a rotating seal. The rings slide past each other with such a short distance that a vacuum can be developed. I had thought I had seen these in jet engines, in that case sealing oil inside a particular area of the engine.
As far as I remember from pulling apart a broken TPU50 once, it's pretty simple. That magnetic cartridge @ 5:10 basically pulls the rotor upwards. And a simple ball bearing on the bottom, that gives the rotor a little freedom in tilting. With enough upward pull, the rotor can't tilt much sidewards, i.e rotor is in a less stable condition, if tilted sidewards. Reason for that magnetic bearing is that you can't have oiled bearing in high vacuum. So either hollow rotor with bearings on air, in the inside, magnetic bearing, or symmetric split design, liuke in TPU330, with bearings on either low vacuum side.
The oil fed bearing is to keep that cone tolerance gap filled with pressurized oil that suspends the shaft of the pump on the surface tension of the oil on the tight tolerance of the cone and the two metal surfaces can never touch like the crankshaft of a car rides on a thin film of oil that is pressurized by the cars oil pump.
Wish my pump was this big
It's not the size that matters. It's even very exhausting to have to carry such a large device around.
@@AdvancedTinkeringit's not about the size, its about the sucking power!
:D
Better have more than one, so people can not call you one pump chump 😂
I'm sorry, but dropping the nut into the rotor the second time absolutely killed me... 😂😂😂
I think it was on purpose because he likes the lovely melody
Couldn't stop laughing, felt like I was working on it
@@christiannorf1680 I swear it wasn't on purpose. But it did sound nice!
This is why you put a towel or cardboard on top to cover it off.
This is not just the largest turbomolecular pump I've ever seen, it is also the second one I've seen. My expectations have been thoroughly met. 😊
I'm glad you're not disappointed! ;)
same here, found this channel randomly through suggestions, and it's really interesting
I used to work at a potato chip factory. I don't know our power consumption in watts but we had two separate 3 phase 7200v feeds, and our power bill was around $250,000 per month. And the fryers were heated with natural gas!
Anyway, every time there was a loss of power for more than around half a second (30 cycles), all of the control equipment would reset. Our process was only certified sanitary under "steady state", so we'd have to dump all food currently in process, and get back to steady state. This cost around $30,000 of raw materials (so much more in lost revenue). During some times of year this could happen several times per day.
Given the short time horizon, and the massive power draws, management viewed it as an inherently unsolvable problem.
Well. I found a backup power solution that had carbon fiber disks in a vacuum chamber spinning at 500,000 RPM. They could come online within five cycles, and provide power for 15 seconds - which would cover 90% of our outages, but was also enough time to start diesel diesel generators kept in a "hot ready" state (coolant and oil kept at operating temperature with external heaters).
Anyway, based on historical power outage data, the flywheel system would pay for itself in 18 months just from raw material losses.
I tried for over a year to get my proposal accepted, and the fact that management wouldn't take it seriously was one of the main reasons I left the company.
This made me think of that story.
By "this" I mean you computing the energy of your spinning turbine.
I don't remember the size of the spinning disks, but given ½mv² and all that (I can't remember the formula for it spinning disk off the top of my head but you get the idea), the company had focused on rotational speed versus size. They found that these insane RPMs let them store a lot of energy for a short period of time.
@@PsRohrbaugh I think you mean 1/2Iω² ;)
The torque specifications of HNNG for big bolts and MMH for small bolts made my day
my grandfather always taught me the proper torque spec is "tighten until it snaps off and back off a quarter turn". perfect every time.
looks like 1 gutentite for the larger bolts.
You may not be an engineer but you're definitely a mechanic.
13:58 the difference between an engineer and mechanic
17:43 never makes the same mistake twice, just slight variations of the same mistake
19:47 The mechanic slap of approval
Agreed...
It's excellent to put it on your tiny chamber! If the experiment fails and the system vents, the chamber will walk right over to you for maintenance!
That comment made me laugh quite a bit!
But yes, I would definitely think of some way to secure the chamber.
The thing with the concentric lands and grooves looks like a labyrinth seal. They're used in various axial compressors, like jet engines.
Ah, thank you very much! That was another guess of mine, but I didn't think that this type of seal works in a vacuum pump. And I had no idea what to call this type of seal. Thanks for the information!
@@AdvancedTinkering I second this. It doesn't look like the type of labyrinth seals that are meant to keep gasses and liquids out but the kind designed to keep small particulates (especially magnetic ones that might get created by what ever process is in the vacuum chamber) that might get into the very close tolerance gaps in the electric motor of the turbopump.
@@AdvancedTinkering I am not an expert, but I did recently read the Pfeiffer "Vacuum Technology Book Vol 2" for a project. It quickly mentions labyrinth seals in their turbopumps. I believe it is to create a flow restriction for sealing gas injected into the bearing/motor area if you use corrosive process gas. But it could also be to keep the oil from moving into the main body, the book is not 100% clear what it is ultimately used for. It certainly isn't a vacuum seal, that's for sure.
@@HuskyMachiningI concur.
@@AdvancedTinkering There is a youtube channel called AgentJayZ that has several years of videos covering jet engine maintenance.. He has mentioned labyrinth seals and probably has a video dedicated to them some place. The labyrinth seals for a jet engine are made of carbon to handle the high temperatures.
The labyrinth seals I have seen on his channel would prevent axial flow, while what you have I would guess is a radial seal.
at this rate you can call your channel "the turbomolecular pump channel" love the content. These things are hecka instresting.
Thanks! The only problem is, that I really have to find some space to put those things.
@@AdvancedTinkering I know this problem all to well, I renovate old machines myself and both having to store them and use them come with some space hogging requirements.
"I'm not taking this apart further"
(2 seconds later)
(grabs screwdriver)
"So this is apart now"
You just told the story of my life and personality.
Haha! I had to really hold myself back from also taking apart the base completely to understand how the ball bearing lubrication works.
I feel seen. Love this channel.
electroboom started his channel with a video where he genuinely shocked himself, dropping the nut a second time is a+ material
The engineering and machining levels on that pump are incredible, it's like military aviation level.
The machining involved in making those vanes must be pretty incredible!
Looks pretty simple compared to any jwt engine tbh 😂
I'd assume 5 axis cnc
@@oliverer3 That, and wire EDM. Some pumps I've seen actually take a cheaper route for the stator vanes, they use stamped sheet metal instead with the vanes twisted to angle them.
@@imajeenyus42 parsons first steam turbine was pretty damn basic...
but other than improved machining techniques, not much has changed since then... had the maths down first go, basically. expansion ratios and blade pitches, areas, that sorta thing...
That sound from the intro will make any jet engine technician shiver in absolute horror. lol
It reminded me of the video by AgentJayZ.
Excellent video!!! I just replace them. Never had the time to disassemble one. Thanks!
Thanks! I'm glad you liked the video!
Interesting pump, never seen one with active lubrication.
Regarding the nut falling into the pump, it is a good idea to place some paper, lint free cloth or Al foil in any opening the instrument you are working on may have. That will keep small items from falling in, even the most dangerous: those you may not have noticed.
“ . . . even the most dangerous: those you may not have noticed.”
Yes yes yes . . . !
I dropped a tiny screw while working on my mom’s engine. Never found it, but I had spares. I was also fortunate that when my moms ignition system shorted out from that screw bouncing around in her distributor, and it burned the primary ignition wire all the way back to the key switch, she was only about 5 miles from home. I hot-wired it so we could drive at home & I re-wired it the next day.
i'm 8 years into my job and still can't believe how often beer is seen as a totally legit form of "payment" between two companies
2 turbo pumps for a few cases of beer is one awsome trade. These were no doubt pm pulls but they would be easily good for hobby level deep vacuum experiments ❤
I want to see this pump spin some more! And I bet the sound is just as awesome as the spinning!
There will be a video with the pump reaching full rpm.
19:15 I think you need not worry about alignment because the 3 screws are not 120 degrees apart on that plate. Same with the 4 on the outer one.
Hi, I'm a 360 degree science enthusiast and "taking it apart to see how it works" is one of my favorite things, as well as obviously designing new ones, the point is this, I'm so passionate about your video and reading the comments that I didn't realize that the tinnitus I suffer from 24 hours a day had disappeared the whole time! Now it's back but, think, would you have ever imagined that your videos were even healing!
I'm very happy to hear that you enjoyed the video and it distracted you from your tinnitus. I wish you all the best and hope it will vanish eventually!
i suffer tinnitus. i think its mental.
hum the note back to yourself... it fades. i find theres several, work one at a time...
ignore the traditional explanation of damaged hairs. its your brain trying to bias itself, get itself in calibration. damaged hairs maybe, but whatever, its more about having a base reference for "silence".
for a teenage metalhead that blew a lot of things up, i still got well above average hearing in my forties...
You bring up an excellent point concerning the potential energy stored in a rotating mass. We had a failure of such a device (a GeneVac rotary evaporator) at work about twenty years ago and it was quite exciting. (The failure was entirely our fault, we overloaded the system with home-made sample cradles machined from solid blocks of metal rather than stamped from sheet.) Probably a similar amount of stored energy to your large pump (slower, but more rotating mass). The net result was a 2" (50 mm) steel shaft snapped, a 4" (100 mm) deep dent formed in the 1" (25 mm) thick aluminum case, the whole 250 pound (113 Kg) device did a 180 degree turn on the bench and jumped about a foot (300 mm) to one side, tearing out the 1" (25 mm) steel vacuum line that connected it to the vacuum pump (a liquid ring pump, quite a curious device in its own right). Also made a hell of a bang.
That is incredible! At first I marveled at your great memory. By the end, I realized that all the details would be emblazoned into my brain for the rest of my life, too.
sounds like you had the write the report
@@berrigo2 Fortunately not, but I was involved in the initial installation and operation and got to see the corpse while it was still warm. Even more fortunately, I was not involved in designing the modified cradles or approving their use. In the aftermath there was a fair bit of back and fourth with the manufacturer and it was their engineers who determined that our cradles overloaded their hangers which was what actually failed, allowing the cradle to contact the case at very high velocity and bringing the other 11 cradles and (132) samples to an abrupt halt.
@@robertlapointe4093 i wish i could get a coffee with and listen to you talk about crazy things like this
For us Americans, that torture spec roughly translates to 4-5 Ugga duggas then about 2-3 ugga duggas
You Americans with your weird units. It's so much easier to calculate with nnnggghhhh than with Ugga Duggas.
What a beautiful piece of machinery! Thanks for showing how it goes together. You had me laughing at the payment in beer; and also at the dropped nut twice!
The conical nut itself I believe to be a centrifugal oil pump. The oil is dragged around between the nut & the outside by viscous friction. Centrifugal force pulls the oil up in the conical gap.
12:54 - Was this turbine rotor CNC milled all in one block of material in one go or was each rotor disc CNC milled one-by-one separately then later assembled into a giant rotor turbine unit? Thanks.
I only know how the smaller models are made. But I suspect they are made the same way. They CNC mill each rotor disc separately and then assemble them onto a shaft using a shrink fit.
I love the sounds it makes when you diassemble / reassemble it... it reminds me of Einstürzende Neubauten's "jet turbine".
The copper tubing on the flange isn't to cool the flange, it's to cool the oil. It's used as a oil cooler on the return side of the pump.
I have to admit I laughed out loud when the nut fell in. Very relatable moment
I was so annoyed when it happened :D thankfully it came out without dissasembly.
Yup had the nut drop or screw drop too... and they are usually nonmagntic as well.😂
That was beautiful, watching it spin effortlessly, I totally enjoyed this video, pointing out the roughness of veins, I don’t know the sounds it made coming apart and back together it was just a beautiful thing thanks!!!
You have two, I think that means one on the top of the chamber and one on the bottom.
You have no idea how expensive those large flanges are -.-. Otherwise....
That must be some AMAZING beer you traded... 😂
Can you imagine the paperwork of a lab engineer selling used lab equipment with possible health and safety implications to a private individual? From a university or research department that is entrenched in complicated paperwork but not set up for actually selling items? Hours of work for several people including the meetings and paperwork... Alternative B: Nobody wants them, so the alternative is getting rid of them properly. Slightly fewer meetings and hours of paperwork, but now the cost is significant. The professionals receiving them will want to know every material that has ever been used even near the thing. "Take them off your hands for a case of beer"? You bet that is AMAZING beer!
Fyi. Torque specs are always based on the bolt, and material its going into. So you can look up m10 torque spec in aluminum and be pretty-alright.
That's such a valuable information! Thank you a lot! I knew that the material plays a role in the torque spec but I always thought it is also highly dependent on the use case.
@@AdvancedTinkering its really about the clamp load. For example an m8x1.25 torqued to its spec (31ft lbs) has nearly 6000lbs of clamping pressure.
So the total amount of clamp pressure required for a job is basically how many bolts you need for that specific task, and the torque rating is basically the max that the thread teeth themselves can handle with out being damaged. Thats why in aluminum will alsways be a bit less. Theres dry and wet torque specs too.. if you use a bit of oil, and then use dry torque specs you can and will start to snap bolts.
Some bit of extra fun... if youre making something with threads, your thread depth / material thickness only needs to be as thick as the nut for that thread would be to acheive full clamp load. (You dont ever need to tap something m8, 2" deep for "extra strength".. yes the teeth def wont strip, but the bolt will strech and snap instead)
Great chanel, we'd prob make good friends 😂 I should get back in the youtube game but try to be a bit less awkward about it lmao. You ask in some videos about your glass work and if you should leave clips in.. I think you should do a few vids on only glass work. How to make a hole, how to weld a joint, how to pre heat post heat, whole 9 yards... Theres lots of other "advance tinkerers" out there who might like to learn from you.. myself included. 👍🍻
What an amazing machine! I have to admit I laughed pretty hard when you dropped the nut into it a second time, thank you for including that bit!
This whole video was therapeutic 😁😍
Ahh yes, the best way to sell something for a symbolic price: "Bringst halt ein Kasten mit, dann wird's schon passen."
Good beer choice btw.
Universelles "Zahlungsmittel" für solche Gelegenheiten.
14:42 the piece you put in is a seal mechanism, tortious path impedes movement of particles to go into the motor winding area or for anything else to come out of the area.
That at the bottom is a Labyrinth Seal. You cannot use O-rings at these speeds but they still wanted to seal the motor off a bit.
It was perfectly clean man hahahaha you just wanted take it apart.
Haha, maaaaayyybeeeee
Really enjoyed the breakdown. I would have had to take it apart as well, those designers did a super nice job.
I had the opportunity to disassemble our KYKY TM pump recently as well. That one didn’t go back together as the internal water cooling channel leaked into the turbo… while it was at speed. Didn’t end well.
"I don't think I'm going to disassemble this any further" Haha we all say that but who are we kidding!!. Awesome pump!
You got me beat by one size. I run the ATH1603M on my chambers. I did just put together a similar chamber to the video ending. DN200CF ATH1603M on a 200mm x 400mm chamber. With 5 DN63CF ports around. With an Ebara EV-A10 it roughs in 30 seconds and reaches 5e-7 torr in the 5 minutes it takes to get the pump up to speed.
Would love to see some pictures of that chamber!
Beautifully done!
Thank you!
Great to see such a big pump getting taken apart.
Putting the big pump isn't a particularly ridiculous idea to me. It might be useful if you want to do all kinds of depositions etc where you want to be able to say that you only get exactly the thing you want to deposit onto your sample and everything else gets pumped out.
But if you put it onto your chamber, you should consider the crash moment (you did work out the energy after all) and bolt your setup to the ground or similar. Nobody wants a vaccum chamber flying through the room.
I have been working lately on Pfeiffer ATH 2303M pumps, they are stunningly massive, but I have not yet dropped a nut into them ;)
3/10, cannot recommend.
Sounds nice but it's a pain to get back out.
That is a fine piece of engineering.
You R the "Turbomolecular Pump Whisperer".... and that double-nut drop got me to hit SUBSCRIBE. Very clever of you. Cheers!
That is one *honking* big pump!
Turbomolecular pumps seem common these days, back when I was doing vacuum stuff back in the mid-70s, we used diffusion pumps for basic coating work, and molecular seive/ion/titanium sublimation pumps for ultra high vacuum. Turbos seem much simpler and higher-performance from a system perspective, I’m guessing turbos hadn’t been as well-developed at that time, so ion/TSP was a more effective combination in that period.
As for your last idea, I support it. Having a big turbo pump is a luxury. If you had limited funds and built the setup with new components, absolutely not, but if you have a big pump available, definitely go for it. Source: We happened to randomly end up with a similar setup at the load/lock of one of our systems. Pumps down in no time, and when necessary we can open the gate valve and pump down the whole system quickly that way. Nobody is in a hurry to get back to a small turbo. 😆 The risk of unfortunate things happening when you mess with the turbo yourself is your own business in this case! Let us hope your energy calculation was not foreshadowing. 😇
I am working on it ;) but I doubt it will be a permanent setup. Since the pump must be mounted upside down, I would have to lift it every time I want to open the chamber 😅
Maybe in the future.
Love the machining on these parts these have to be very expensive pieces of kit when they are new
Bayreuther Hell 🍻❤️
These pumps are frighteningly expensive, you're very lucky to get them for a case of beer.
You know in your heart that the giant pump belongs on the top of your chamber; they belong together!
very cool video. truly appreciate your time and effort.
Thank you! I appreciate hearing that!
I used to work in helium leak detection. We would evacuate a small roughly 12 cu ft stainless steel bell jar with two large Stokes pumps to a certain vacuum level before we could expose the Turbo's to the vacuum level in the bell jar. If we exposed the turbos to atmosphere while running, the compressor blades would disintegrate from the inrush of air and from a soft startup at atmosphere level they would never ramp up to speed (70 to 75,000 rpm i think. it's been a long time ago.)so they ran all the time and were operated by a valve that opened them up when the bell jar reached the right vacuum level. I too opened one up one time. It never ran again.
14:30 The molecular stage or Holweck stage is working on molecular drag on moving surfaces. It does not have the volume throughput of the turbo stage but it is able to pump against a much higher pressure. If you have only turbo stage you need a primary pump with an end pressure better than 5 Pascal but with the Holweck stage you can work with a base pressure of 100 Pa obtained with a membrane pump.
I sure am glad I never have problems like that nut falling in for a second time.🙂It must be so frustrating.
Please try to get videos with close up shots of individual components. It is usually quite hard to find pictures of functional components of these things. Awesome video, thanks.
my mans here is gonna get one so large he will control all the air in the world.
Thanks for sharing, you so lucky
Dope content my friend❤ I imagine it must be very expansive piece of labgear
Thanks! I'm glad you like the videos!
Yes, those pumps are indeed very expensive. Without the generosity of their previous owner, I would never be able to show them to you guys.
Oil pump systems do usually discharge high pressure through a smaller diameter path than the larger low pressure path. For reasons including thermal expansion, aeration, and deceleration while collecting in the reservoir to reduce frothing.
17:26 bwahahaha the xylophone of unemployment.
Holy shite you did it twice.
Thanks for that information! "Xylophone of unemployment" is the best term for it I've heard so far 😄
And I though my EXT255H was oversized for my system. You win 🏆
@19:34. good note. Also I am pretty sure they made the 4 screws asymmetric, therefore the back center cover only goes on one way.
Wow those are the same size of the ones for purging a MOCVD chamber we baked GaALAs laser diodes. When they fail they FAIL ime, thankfully the most common is bearing failure, if the temperature so much as hints at going up, power down the pump or at least reduce the RPMs enough to finish your task.❤
I wish I could get a turbo molecular pump for a case of beer.
What's inside? Well turbomolecules of course.
Hallo, hier der Pfeiffer Techniker,
hier mal ein bisschen Feedback zu deiner Demontage.
Magnetlager: Die Unterlegscheiben sind sehr wichtig für die Funktion der Pumpe. Sie werden zur Einstellung des Magnetlagers benutzt und kommen von unterschiedlichen Dicken (0,8mm - 2,0mm) in 0,05 mm abständen zum Einsatz. Sie werden bei jeder Montage neu gewählt damit das Magentlager die korrekte Position zum Rotor hat. Es kann sein, dadurch dass du die Fläche beschädigt hast (bei deiner demontage) der Abstand nicht mehr korrekt ist und du dadurch höhere Vibrationen in die Pumpe bekommst.
Rotor / Statorpaket: Der O-Ring ist ein Toleranzausgleich. Durch die vielen Distanzringe können sich Toleranzen der Ringe aufbauen, die das Gehäuse nicht ausgleichen kann und die Statorscheiben könnten nicht ausreichend gequetscht werden und auch hier Vibrationen erzeugen.
Ölkreislauf: hier liegst du komplett richtig mit deiner Annahme.
Labyrinthscheibe: Diese Scheibe zwischen Rotor und Gehäuse ist ein Gasschutz für den Motor. In dem Unterteil ist außen eine Schraube in der ein Schutzgas angeschlossen werden kann. Gerade wenn korrosive Gase gepumpt werden muss die Pumpe mit einen Schutzgas (z.B., Luft oder Argon) betrieben werden damit die korrosiven Gase durch den Druckunterschied nicht in den Motor kommen.
Schrauben: Ich weiß nicht wie oft mir schon diese schrauben in die Pumpe gefallen ist. Du bist nicht alleine!! Ich habe deinen Schmerz gefühlt, vor allem beim zweiten Mal.....
Wenn du weitere Fragen hast kannst du dich gerne melden.
The ridged vanes you call "molecular drag stage" I believe to be a simple serpentine barrier between the motor/bearing & the working volume of the pump. Normally a small flow of purge gas is applied to the motor section to keep "krud" from getting into the motor/bearing.
Thank you for the information! Other comments agree and called it a labyrinth seal.
That's an awesome sound, the sound of the pump ringing with the nut falling through it, twice! Just kidding, that's the expensive sound nobody wants to hear. I love the sound of it running at the end though, I'd love to hear it what it's like, installed in a setup.
Turbo pumps have improved alot for this application. While not indestructable, newer systems are more caoable of handling some process upsets like occasional small qty droplet impact
Not the biggest I've seen by a long shot, the machines I work on have turbos that weigh more than I do! Granted they also are evacuating chambers large enough for several adults to climb inside haha. They work beside cryos running at around 10K. Working in a wafer fab is pretty extreme lol.
Edit: oh and the pumps we use are mag lev on ceramic dry bearings, when your process is dirty dry pumps are much nicer, also helps reduce particles.
Those must be big wafers you are baking. Ours were 2.5cm gallium arsenide wafers for laser die making. It used two turbos about this size to pump down the chamber where it needed to be. ❤
That‘s not a big vacuum chamber.
For large area PVD glass coating, vacuum chambers generally measure about 80-140m in length and 4m in width. Usually, there are 160-240 of these Pumps attached.
@@christiansprojects-cgmanuf1426 that's certainly some impressive scale! What preasures are we talking?
@@christopherleubner6633 eh older style 200mm wafers, my plant makes analog automotive chips like sensor interfaces and power transistors etc. We're a boutique fab so we handle a lot of the smaller batch special order items
@@xxportalxx. Not too low, as the quality requirements are not as high as in semi conductor industry. Base pressure is at around 8x10-7mBar (8x10-5 Pa), process pressure at around 5x10-3 mBar (5x10-1 Pa). On most of these Coaters, you have around 200 Turbo Pumps on the main chamber and 20 on the intro and outpus chambers.
There are 40-80 Magnetrons with a length of 4m installed. Per Magnetron, the electric power is 50-150kW. So in total, someting around 4MW of electric power that is required.
thats very cool, but whats amazing is you bought it for beer ;-D
Thank you for the torque spec joke, 😊.
Imagine if 'Manuals library' had a manual for one of these. 😊
I want to see how fast that thing can pump down your chamber! You could set it up to compare your other pump against this one.
I will do that!
Makes a nice fan for the summer ... lol
Pfeiffer Vacuum Wetzlar ?
The oiler makes me think of one-shot lube systems used on metalworking equipment. In those, the oil is lost rather than recirculating. I suspect the port you identified as for return oil is actually to equalize vacuum in the oiler so it doesn't have to work so hard. It probably injects a tiny amount of oil periodically - a small enough amount that it won't mess with anything downstream.
Interesting. Of course, that would also be possible.
Anyone else desperately want to turn that into a jet engine?
Well compressor section anyway...
Oh this nice sound of the Blades and Vanes from a falling screw still makes me think "Oh F..."
I've worked in the Aviation industry on Turbofan Engines and once, i've had a screw lost in the compressor. That chimes were beatifull and horrifing at the same time.
I hope they have a video somewhere showing how these are made because that is an insane machine. I'd build a room sized chamber if I had one of those.
14:30 It's a dust trap I would say. The “up and down labyrinth” makes it more difficult for dust to get towards the bearing block
21:00 Same picture as mine yesterday. Only my pump is a little smaller. TMH521 with DN160.
13:10 I would like to see a turbo crash at top speed in slow motion. The kenetic energy must be incredible. We had a TPH2200 that suffered damage at high speed and fell apart into countless pieces
Excellent video, this is why l ordinary subscribed to TTube. 8:08
Thank you for the heads-up.
Thanks, I appreciate it!
I think I may have the perfect concept to power this thing
I'm not even embarrassed to say I would try to make a small turbine motor out of them.
They just BEG to be franensteined into small turbojets!
How do these pumps stop air from flowing back into the chamber when the pressure gets very low
13:57 see also the german standard torque specification of "gütentight"
I once knew a research group that managed to drop a nut into a pump at speed. The rotor was basically a solid at that speed, so the nut just bounced back into a remote section of the chamber. THE FIRST TIME! The next time, the pump injested the nut, blades, etc.
I have a similar pump amd really love these videos of you taking these apart. So cool. You're a smart dude.
What are you studying?
I wonder if the thing at 14:28 is a magnetic shaft coupling. The concentric fins look like magnetic field guides to me.
Those are things of absolute engineering 'beauty' - they are basically jet engines without fuel and ignition!
0:28 herrlich das Bayreuther aus meiner Stadt! :)
Almost as good as ripping into it myself
These pumps are assembled in clean room conditions and very carefully.
Considering that rotational speed and diameter i think it is machined with insane nm tolerance, and such disassemble unfortunately disbalance rotor, even touching it by bare hand it deflect. I hope that magnetic bearing and hydrodynamic oil other side have some trickery to self balance for small deflection of main rotor.
@@AABB-px8lc I witnessed one that was cleaned professionally tear itself apart - $10K shot to hell - it was loud, but brief, but they replaced the unit. The tech probably nicked something while re-assembling. They never let us know what the cause was, but our MassSpec was down for nearly 2 weeks, right at the time we were doing our annual testing for the govt. I think it was an HP, but can't remember (30 years ago).
30,000RPM or so. I've had 14K RPM SCSI drives fail catastrophically as well, but it's more self-contained - the shrapnel didn't get through.
They are not vacuum pumps they are jet engines ha ha !!! On an more serious level !! The flow of oil is usually the largest hole is the inlet and the small hole is the high pressure side or out let !!!! Thay are huge and I bet they will quickly get a deep vacuum !!! Good luck with them !!!😊
14:18 it looks like a rotating seal. The rings slide past each other with such a short distance that a vacuum can be developed. I had thought I had seen these in jet engines, in that case sealing oil inside a particular area of the engine.
You are correct. It's called a labyrinth seal. It's an air seal I think..
Very nice Pump.
Would be interessting how the magnetic bearing is concepted.
As far as I remember from pulling apart a broken TPU50 once, it's pretty simple.
That magnetic cartridge @ 5:10 basically pulls the rotor upwards.
And a simple ball bearing on the bottom, that gives the rotor a little freedom in tilting.
With enough upward pull, the rotor can't tilt much sidewards, i.e rotor is in a less stable condition, if tilted sidewards.
Reason for that magnetic bearing is that you can't have oiled bearing in high vacuum.
So either hollow rotor with bearings on air, in the inside, magnetic bearing, or symmetric split design, liuke in TPU330, with bearings on either low vacuum side.
The oil fed bearing is to keep that cone tolerance gap filled with pressurized oil that suspends the shaft of the pump on the surface tension of the oil on the tight tolerance of the cone and the two metal surfaces can never touch like the crankshaft of a car rides on a thin film of oil that is pressurized by the cars oil pump.
YIkes! This thing was probably $50,000 new.
Love to see pump down speed
There will be a video about that.