# How do centrifuges work (that is, the mechanism)? **URL:** \ **Category:** Factual Questions\ **Created:** [July 16, 2014, 3:23am UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248 "2014-07-16T03:23:56Z")\ **Posts on this page:** 18\ **Page:** 2
**Author:** ![Machine\_Elf](https://avatars.discourse-cdn.com/v4/letter/m/82dd89/32.png) [@Machine\_Elf](https://boards.straightdope.com/u/Machine_Elf)\ **Post date:** [July 17, 2014, 5:27pm UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/21 "2014-07-17T17:27:19Z")
> [@EdwinAmi](#): > > The question from there is, given the g forces you guys describe, what kinds of forces does that come out to mechanically? I’m thinking I’d turn it into a pressure calculation, among other ways of calculating it out A gallon of water weighs about 10 pounds at one g. 3 gallons at 25,000 g would exert a force on the bottom of its container of about 75,000 pounds. The pressure will be determined by the area at the bottom of the container. If your container bottom is 8x8 inches, then the pressure will be 1172 psi. ---
**Author:** ![EdwinAmi](https://avatars.discourse-cdn.com/v4/letter/e/5fc32e/32.png) [@EdwinAmi](https://boards.straightdope.com/u/EdwinAmi)\ **Post date:** [July 17, 2014, 5:54pm UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/22 "2014-07-17T17:54:16Z")
> [@](#): > > A gallon of water weighs about 10 pounds at one g. 3 gallons at 25,000 g would exert a force on the bottom of the container of about 75,000 pounds. You’re forgeting that 25,000 g’s is at the rim. The force closer to the center of rotation is significantly less ---
**Author:** ![Sage\_Rat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/sage_rat/32/399_2.png) [@Sage\_Rat](https://boards.straightdope.com/u/Sage_Rat)\ **Post date:** [July 17, 2014, 8:18pm UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/23 "2014-07-17T20:18:18Z")
> [@Mangetout](#): > > I don’t think that’s possible. I might have got the math wrong, but by my calculation, assuming a 50cm radius (to accommodate an 8 gallon container), your 8 gallon sample will ‘weigh’ about 7500 tons, when spun to 20,000 RPM. I’m not an engineer, so this is an honest question. If you made your centrifuge a double, would that help in any way? I.e. you have a main shaft with some number of arms that spins at X RPM, and the tube in each arm is shaped like [an octopus hot dog](https://www.google.com/search?q=hot+dog+octopus&client=ubuntu&hs=DXE&channel=fs&source=lnms&tbm=isch&sa=X&ei=BC_IU5GbMYbdoASL8oGIBg&ved=0CAgQ_AUoAQ&biw=1547&bih=796#imgdii=_), with its own motor to spin it at a 90 degree angle to the main shaft, at Y RPM. ---
**Author:** ![wevets](https://avatars.discourse-cdn.com/v4/letter/w/e9a140/32.png) [@wevets](https://boards.straightdope.com/u/wevets)\ **Post date:** [July 18, 2014, 12:12am UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/24 "2014-07-18T00:12:58Z")
> [@EdwinAmi](#): > > I haven’t been specifying a g because I can tell you that the process is always done at at least 10,000 rpm, but usually at 20,000 rpm or faster because the separation happens faster. In industrial processes, time is usually the favored variable to shrink, so I’m thinking doing the process commercially you’d opt for the faster speed/time. Unless centrifugation is the limiting rate in the process, it doesn’t make a lot of sense to speed that up before it’s clear it can be done safely/inexpensively. What’s wrong with figuring out how quickly existing industrial centrifuges can spin out your precipitate before custom-making centrifuges which might be prohibitively expensive? The largest centrifuges I know of can handle around 12 liters (~3 gallons) at 7000 g - there might be something bigger out there - I don’t know for sure. ---
**Author:** ![Chronos](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/chronos/32/134_2.png) [@Chronos](https://boards.straightdope.com/u/Chronos)\ **Post date:** [July 18, 2014, 12:22am UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/25 "2014-07-18T00:22:33Z")
**EdwinAmi** , it takes 10,000 RPM _for a centrifuge of the size you’re using_. With longer arms, you could get the same effect at a lower speed. But we can’t tell you what speed you’d need because we don’t know how long the arms are that you’re using now. ---
**Author:** ![Kevbo](https://avatars.discourse-cdn.com/v4/letter/k/e47774/32.png) [@Kevbo](https://boards.straightdope.com/u/Kevbo)\ **Post date:** [July 18, 2014, 12:38am UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/26 "2014-07-18T00:38:02Z")
You might also look into centrifugal separators. These are set up for continuous flow, rather than batch processing. It may require several stages to achieve your aim, but this may work better for volume production. ---
**Author:** ![EdwinAmi](https://avatars.discourse-cdn.com/v4/letter/e/5fc32e/32.png) [@EdwinAmi](https://boards.straightdope.com/u/EdwinAmi)\ **Post date:** [July 18, 2014, 4:08am UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/27 "2014-07-18T04:08:20Z")
> [@](#): > > EdwinAmi, it takes 10,000 RPM for a centrifuge of the size you’re using. With longer arms, you could get the same effect at a lower speed. But we can’t tell you what speed you’d need because we don’t know how long the arms are that you’re using now. uhhhhh… no? The amount of g’s you need to settle out a precipitate is the amount of g’s you need, period. This means the centrifuge would have to deal with the same g forces whether it’s a normal size one going at 10,000 rpm or a really weird one with long armatures going at like 5,000 rpm. Sure handling low rpms would be easier, but the problem we ran into here, as some peope pointed out, is that the forces for such large amounts I’d like to process would be huge. And like I said they’d be the same even on a weird long armature slower spinning centrifuge, since one way or another you have to reach a certain g force ---
**Author:** ![wevets](https://avatars.discourse-cdn.com/v4/letter/w/e9a140/32.png) [@wevets](https://boards.straightdope.com/u/wevets)\ **Post date:** [July 18, 2014, 5:14pm UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/28 "2014-07-18T17:14:58Z")
> [@EdwinAmi](#): > > uhhhhh… no? The amount of g’s you need to settle out a precipitate is the amount of g’s you need, period. This means the centrifuge would have to deal with the same g forces whether it’s a normal size one going at 10,000 rpm or a really weird one with long armatures going at like 5,000 rpm. 1. You haven’t told us how many _g_s you need. You’ve only stated RPM, which is going to vary by centrifuge and rotor choice. See if you can find your centrifuge rotor on [this handy g/rpm/rcf converter](http://www.endmemo.com/bio/grpm.php) to give us better information. 2. Not really - there’s also _time_ as a variable to consider. Most precipitates will come out of solution at a lower g if you centrifuge them for a longer period of time. Just because everyone does this process at 10,000 rpm doesn’t mean the precipitate won’t come out at 9,000 rpm, barring a few unusual circumstances which you haven’t raised in this thread - it could just mean they don’t want to wait to do the process at 9,000 rpm. Is there any special consideration that tells you that 9,000 rpm won’t work? ---
**Author:** ![Chronos](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/chronos/32/134_2.png) [@Chronos](https://boards.straightdope.com/u/Chronos)\ **Post date:** [July 18, 2014, 10:04pm UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/29 "2014-07-18T22:04:05Z")
> [@](#): > > Quoth **EdwinAmi** : > > uhhhhh… no? The amount of g’s you need to settle out a precipitate is the amount of g’s you need, period. How does this differ from what I said? You need a particular number of gs; you don’t need a particular RPM. ---
**Author:** ![Enola\_Straight](https://avatars.discourse-cdn.com/v4/letter/e/dec6dc/32.png) [@Enola\_Straight](https://boards.straightdope.com/u/Enola_Straight)\ **Post date:** [July 19, 2014, 1:29am UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/30 "2014-07-19T01:29:39Z")
Simple Centrifuge seems to cover your needs. [https://www.youtube.com/results?search\_query=centrifuge+simple](https://www.youtube.com/results?search_query=centrifuge+simple) ---
**Author:** ![Mangetout](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/mangetout/32/19_2.png) [@Mangetout](https://boards.straightdope.com/u/Mangetout)\ **Post date:** [July 19, 2014, 10:10pm UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/31 "2014-07-19T22:10:10Z")
> [@Sage\_Rat](#): > > I’m not an engineer, so this is an honest question. If you made your centrifuge a double, would that help in any way? > > I.e. you have a main shaft with some number of arms that spins at X RPM, and the tube in each arm is shaped like [an octopus hot dog](https://www.google.com/search?q=hot+dog+octopus&client=ubuntu&hs=DXE&channel=fs&source=lnms&tbm=isch&sa=X&ei=BC_IU5GbMYbdoASL8oGIBg&ved=0CAgQ_AUoAQ&biw=1547&bih=796#imgdii=_), with its own motor to spin it at a 90 degree angle to the main shaft, at Y RPM. You mean balance the load with something opposite? Yeah, that’s a good idea anyway, just to prevent vibration and uneven load on the bearings. But the point is that a centrifuge that is big enough to accommodate an 8 gallon load, and spins at 20,000 RPM, is (unless I did my sums wrong) generating **hundreds of thousands of gravities** of centrifugal force - that seems like a recipe for a machine that just flies apart and kills everyone. ---
**Author:** ![Mangetout](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/mangetout/32/19_2.png) [@Mangetout](https://boards.straightdope.com/u/Mangetout)\ **Post date:** [July 19, 2014, 10:30pm UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/32 "2014-07-19T22:30:01Z")
> [@Machine\_Elf](#): > > **Mangetout’s** calculation posits a 50-cm (19.7-inch) spin radius. That’s very large compared to most tabletop centrifuges. NB - I think you got it, **Machine Elf** , but for anyone else reading, I guessed at 50cm as a reasonable radius to accommodate an 8 gallon container - I just did the math on this and I’m still happy with that guess - a cylindrical container to hold 8 gallons, and be 50cm tall, would be 30cm in diameter. Even if we had enough unobtainium to build this powerful centrifuge, I’m not sure it would work very well, because of a significant g gradient throughout the container - and maybe problems like coriolis forces too. ---
**Author:** ![Squink](https://avatars.discourse-cdn.com/v4/letter/s/b5e925/32.png) [@Squink](https://boards.straightdope.com/u/Squink)\ **Post date:** [July 19, 2014, 11:15pm UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/33 "2014-07-19T23:15:43Z")
> [@EdwinAmi](#): > > Ultimately, I need to make/have made centrifuges that can spin 20,000 rpm +, handling liquid, but not tiny lab amounts, larger amounts, like 3-8 gallons. [Continuous flow rotor](https://www.google.com/search?safe=off&client=firefox-a&hs=iF0&rls=org.mozilla:en-US:official&channel=sb&q=continuous+flow+rotor&spell=1&sa=X&ei=hfvKU6fZNIPF8AHg3YCoBA&ved=0CBwQvwUoAA&biw=1133&bih=592) ---
**Author:** ![lazybratsche](https://avatars.discourse-cdn.com/v4/letter/l/ba8739/32.png) [@lazybratsche](https://boards.straightdope.com/u/lazybratsche)\ **Post date:** [July 19, 2014, 11:26pm UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/34 "2014-07-19T23:26:01Z")
> [@Mangetout](#): > > You mean balance the load with something opposite? Yeah, that’s a good idea anyway, just to prevent vibration and uneven load on the bearings. > > But the point is that a centrifuge that is big enough to accommodate an 8 gallon load, and spins at 20,000 RPM, is (unless I did my sums wrong) generating **hundreds of thousands of gravities** of centrifugal force - that seems like a recipe for a machine that just flies apart and kills everyone. Seriously. Even with ultracentrifuges designed for 100,000 x g or higher, running an unbalanced or damaged rotor is a [really bad idea](http://web.mit.edu/charliew/www/centrifuge.html). They typically have heavy steel tub to contain any accidents, but sometimes even that is [not enough](http://www.chem.purdue.edu/chemsafety/newsandstories/centrifugedamages.htm). Please, please don’t make a homemade, oversized ultracentrifuge. ---
**Author:** ![Sage\_Rat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/sage_rat/32/399_2.png) [@Sage\_Rat](https://boards.straightdope.com/u/Sage_Rat)\ **Post date:** [July 20, 2014, 1:10am UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/35 "2014-07-20T01:10:48Z")
> [@Mangetout](#): > > You mean balance the load with something opposite? No, I mean you have two things rotating. [![](https://dl.dropboxusercontent.com/s/6pdf6rvpqjagpzq/whirlyfuge.jpg) ](https://dl.dropboxusercontent.com/s/6pdf6rvpqjagpzq/whirlyfuge.jpg) So you have a main shaft, rotating parallel to the Earth. Then each arm that juts out of the main shaft contains a pipe full of solution that we’re trying to separate. The pipe is shaped like an octopus, with arms. Those octoupus shapes spin, perpendicular to the Earth (like a ferris wheel). So the main shaft is sling the solution outwards. The shapes of the pipes is diverting the solution perpendicular to the shaft, and the octopus spinning is adding a second set of slinging, which is (idealy) helping to add extra G to force on the solution in the octopus tentacles. ---
**Author:** ![lazybratsche](https://avatars.discourse-cdn.com/v4/letter/l/ba8739/32.png) [@lazybratsche](https://boards.straightdope.com/u/lazybratsche)\ **Post date:** [July 20, 2014, 4:18am UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/36 "2014-07-20T04:18:06Z")
I take it you’ve never played with a gyroscope? Angular momentum is weird. If you have an object spinning on one axis and apply torque to it on a perpendicular axis, the whole thing will begin to rotate around the third axis. The effect is entirely surprising to our primitive monkey minds, which have done a pretty impressive job evolving an intuition of things that fly through the air, but never had to deal with rapidly spinning objects until very recently. Here, have a [wiki-animation](http://en.wikipedia.org/wiki/Gyroscope#mediaviewer/File:Gyroscope_wheel_animation.gif), a [youtube demonstration](https://www.youtube.com/watch?v=ty9QSiVC2g0), and as always there’s an [XKCD](http://xkcd.com/332/)for that. I think that your centrifuge on a centrifuge idea will add major forces bending the arms up or down, depending on the direction everything is spinning. As a practical engineering matter, you just took a potentially hazardous centrifuge and given it a whole lot of additional forces, and moving parts that will happily vibrate to pieces. ---
**Author:** ![Mangetout](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/mangetout/32/19_2.png) [@Mangetout](https://boards.straightdope.com/u/Mangetout)\ **Post date:** [July 20, 2014, 8:34am UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/37 "2014-07-20T08:34:44Z")
> [@Sage\_Rat](#): > > No, I mean you have two things rotating. > > [Dropbox - Error - Simplify your life](https://www.dropbox.com/s/6pdf6rvpqjagpzq/whirlyfuge.jpg) > > So you have a main shaft, rotating parallel to the Earth. Then each arm that juts out of the main shaft contains a pipe full of solution that we’re trying to separate. The pipe is shaped like an octopus, with arms. Those octoupus shapes spin, perpendicular to the Earth (like a ferris wheel). So the main shaft is sling the solution outwards. The shapes of the pipes is diverting the solution perpendicular to the shaft, and the octopus spinning is adding a second set of slinging, which is (idealy) helping to add extra G to force on the solution in the octopus tentacles. I think that’s goung to be worse. The reason that makes an exciting fairground ride is that the rider is exposed to wildly varying g forces throughout. That’s almost exactly what you don’t want, with a centrifuge designed for settling out liquid, especially with loads big enough for sloshing to be a real issue, and super-especially in light of the apparently absurdly high specification that the OP seems to be pursuing. It would go boom on first use, before even getting to full speed. Centrifuges typically spin perpendicular to natural gravity so that the external forces on them are uniform. ---
**Author:** ![Chronos](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/chronos/32/134_2.png) [@Chronos](https://boards.straightdope.com/u/Chronos)\ **Post date:** [July 20, 2014, 7:53pm UTC](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248/38 "2014-07-20T19:53:58Z")
The octopus-fuge is a bad idea for one reason: It’s more complicated, and doesn’t gain anything from that complication. The goal of any centrifuge is some amount of acceleration. The limiting factor in construction of a centrifuge is also acceleration: You need to make the parts strong enough that they won’t break. Introducing more moving parts just means that it’s harder to make the parts strong enough-- It doesn’t make it any easier to achieve high accelerations. The fact that it produces nonconstant accelerations is also a problem (for most applications), but it pales in comparison to the other problem. [Previous page](https://boards.straightdope.com/t/how-do-centrifuges-work-that-is-the-mechanism/693248.md?page=1)