Відео

Something else I've passed in labs that is no longer a operational mystery! Thank you!

Thanks 🙏👍💯😊

In my lab we have two URX Scopes 128GSps, they are not as fast as this one on the front-end side, but still impressive (the fastest is 33GHz). My opinion after 2 years of use is that they are undoubtedly amazing machines with excellent front-end and DAQ hardware. However (and I honestly hope that some Keysight colleagues will read this comment), I would have expected just a little bit more on the software and UI in general. I'm not saying that is bad, don't get me wrong. But we are talking about million dollars machines, I honestly would like to have the absolute best. In particular, I noticed that the stability should be improved, expecially while CDR or heavy measurements/math are used. In several occasions I had to forcefully restart the OS due to hangs. Furthermore, I'm sure that the UX (including responsiveness and even interface style) could be enhanced: controls are slightly less prompt than on the old infiniium scopes (which I loved, especially the 20 years old ones with that DAQ to GPU waveform bus <3 ). For such a beast, the interface should have been based on c++ real-time rendering engines used in games. I want to see an Unreal based UI for the next time ^^, not those old and heavy menus (just for laughs). Finally, the embedded PC side: I notice that it has a CPU / RAM / SSD that are just ok, probably they could have provided something better for that price. Why compromising the PC side performance on a product in this range? I conclude with the most critical point: the LCD display! This unfortunately is a sore point because it's definitely unsuitable for the product and clearly inferior to the flagship products of competitors (see LeCroy). I honestly would have liked to see the best OLED on the market mounted, it had to be better than a freaking iPad. Don't forget that such a beautiful signal path ends on that display (excluding VISA and the like). It had to live up to that level of front-end. A pity. These are maybe small things, but still things which play a role in the daily use and in the overall experience.

Getting rid of the flexible hose and replace it with short thick wall tube, then see what you get.

Glasslinger has a fascinating Video about repairing a turbomolecular vacuum pump in his very own style ...

A common quick method is to wet each join in turn with IPA. or the best way is to ues a helium leak checker but they tend to be extremely expensive

Wouldn't the long hose/pipe make it less probable that the molecules manage to randomly wander down that distance, and hit the later stage's blades?

Very well done presentation. Low vacuums are tricky! I worked for a company for a while that mfg sensors for very low Tor. They used turbo pump systems very similiar to what you demonstrated. They couldn't get reliable employees. They blew up pumps occasionally and it got very costly to fix them!. Sensors were based on a thermocouple design. Great design but the company was way behind the times in their mfg technology (outdated) and they had high sensor failure rates in mfg process due to sloppy hand assembly. Most all units were based on a balanced bridge thermocouple design. Fun job, old school bad management behind the times.

Do you wear an ESD strap on your ankle or trust that there won't be any discharge into the boards when you take them out? 😳 Just wondering, not used to work with so expensive equipment, and would be paranoid if it was just a $1k device. This is a great example of how a very expensive piece of equipment stops working because of a single component failing. And nowadays most consumer electronics are so cheap that instead of fixing it, the whole thing is melted to extract the metals while burning all non-metals polluting the atmosphere. And these metals are then used to produce new electronics.

I have long known about the basic idea of turbo molecular pumps, but my idea was that they were operated in parallel with ordinary roughing pumps. Putting them in series vastly more sensible. Took me a moment to realize that the roughing pump pulls gas *through* the (idle) turbo pumps. What a great idea!

Back in the day (1970's) Where I worked we had a room-sized Mass Spectrophotometer and all of the gaskets were made of gold wire. It provided nearly perfect sealing, didn't outgas, and resisted contamination from samples. Try the soft copper first then move to gold if you get the performance improvement you expect.

I’m intrigued as to why Alcatel - basically a telecoms equipment comps should have been in the vacuum pump business? What’s the connection?

IME, especially for a multi stage system, 10-7 should be achievable within 30min of reaching full speed, even with outgassing so a leak is definitely possible. Did you apply a light coating of grease to them? I would always recommend that regardless - if you have some apiezon or other vacuum grease lying around, great but if not, order some with replacement orings - KF25 o rings are dirt cheap as is. Also try testing the system pressure in stages - put the gauge on the roughing pump inlet only, record the pressure, then add the line to the first stage and blank it off, record the pressure, then the first stage with the inlet blocked off, record the pressure (with only the roughing pump on), and keep going adding one piece at a time. The ultimate pressure should be very close if your system is vacuum tight, but if you see a big difference when you add the hose for example, you know the problem is there.

As most comments are saying, lower pressure is just a matter of "hurry up and wait." In absolute terms, moving to soft copper seals instead of orings will help, and so will baking the manifold, but you must be very careful that your components can take the baking. Many pumps and valves very much do not like getting hot like that, so you have to be careful.

Shahriar: Explains how outgassing is a major issue in high vacuum systems Also Shahriar: Rubs his fingers all over his KF adapters

The maximum breakdown voltage of P-channel transistors does not exceed 600V, so using complementary transistors in high-voltage amplifiers is a bad idea.

Ever since I got back into chemistry I've been wondering how these things work thanks for explaining it all. very well done!

Just looked it up to remind myself. at 10^-12 Torr, you have one molecule per CC. at 10^14 Torr, there's 1 molecule in 10 CC. Viscous flow stops all the way up at 10^-3 Torr or thereabouts.

Many years ago I worked for a company that made ion-beam guns. We usually operated on the -10 to -12 scale. We used turbo pumps to rough down the chambers and then switched over to cryo pumps.

It's challenging to get really low pressure with k f flange seals. Possible but everything has to be perfect. Some people lube the o rings with fombolin but i am not sure of current practice

Super!

I’d def get one of those residual gas analyzers .. it adds more volume to take care of but is a big help when trying to figure leaks … eBay got the older inficon transpector units for cheap .. running one here as a gas analyzer and a single varian tv141 plus an Edward’s rv8 easily go down to the 1e-6 torr range …

I heard you talk about 30k rpm max but as far as I know this speed is only true for pumps with rotors of bigger diameter, all smaller pumps I have seen like you have there and like I do have at home spin at 60k rpm in order to get the speed necessary

I work with these all the time, this was quite interesting. But, uh, why are you not wearing gloves? Finger oils really should be avoided with any part of these mechanisms.

alphaphoenix did a video on finding leaks in high vacuum experiments years ago.

10e-7 is viable with this setup, assuming you don't have significant leaks. Any leaks would be difficult to identify without additional equipment. Outgassing on the other hand can take a while, pressure can improve over several days. You can heat the chamber and pumps to improve this, 60-80C everything should operate fine. Contamination can be an issue, it is worth wiping down exposed surfaces with alcohol is possible, though wearing gloves likely isn't a huge deal at the pressures. I don't know exactly what model of gauge that is (the digital one) but it appears to be a pirani/cold cathode full range gauge where the electronics attached to the gauge head control the switching between the two technologies. Your understanding of the vacuum also depends on how accurate your gauges are. Good explanation of the working principle of cold cathode gauges. In the high vacuum range the cold cathode gauge has at best +/-30% accuracy, an old gauge like this can easily be out by an order of magnitude if it has been exposed to contamination. As you mentioned these can be disassembled and they can be cleaned if you are careful and know what you are doing. I would suggest inspecting all the O-rings that are easily accessible to ensure they are not cracked and also checking the diaphragm on the vent valve you have at the top there, it can be disassembled to inspect. Your flexible hose is another potential place that there could be a leak, these can fail when they flex. That can be difficult to identify but a close inspection, especially close to the welds at either end of the hose is worthwhile. Ideally use a small amount of vacuum grease on these surfaces during assembly if you have this available to you. As a side note for anybody interested, turbomolecular pumps can operate at even higher speeds than noted in the video, at least 90000rpm that I am aware of, the rotational speed has a direct relationship to the size of the pump as the vanes must be traveling at higher than the speed of sound for the pump to be effective, so pumps of a smaller diameter must have a higher rotation speed to be effective. A note on cryopumps would be that they do not achieve much better pressure than a turbopump, though they can have very high pumping speeds. UHV tends to be more about material selection, bakeout, sometimes other pump technologies such as ion pumps. The clamps on those KF fittings look like they could do with being tightened a little more, though this is difficult to judge from a video. You can tighten them as much as is comfortable by hand, that should ensure a good seal if everything else is ok.

Wait, I am thinking at loud here, if Dyson motors are spinnin at about 120.000 rpm-s at "room" pressure can you imagine how fast would they spin at "running turbo molecular pump" pressure (lets presume that motor temperature would not be an issue). :D Another thought, what about sending this pump on a space station and running it widelly open (no top, so you could see spinning blades) and recording it with 100.000 fps cammera? (behind the space station so it would be shielded from sun) :D

Much respect! Thank you for your continued mentorship and priceless knowledge.

Ah yes high vacuum. Pain in the arse.

Can you characterize cheap cables/ bare wire/ baseline substitutes??

Literally so pumped you're getting into the high vacuum game...

I used to work for a company called Lakeshore Cryotronics, they have systems that can go down to 10 millikelvin and power electromagnets to produce 9 Teslas. Think you might find their machines interesting if you're diving into cryo

Never expected to learn this today. Very interesting.

As probably has been said, the vacuum level you're seeing is probably water and oil contaminants that have built up on the interior surfaces. Cleaning and baking (when appropriate, not every O ring holds up to it) will definitely help, but while I do think you should be able to get to 10e-7 Torr levels with effectively no chamber attached, with a chamber of any real size, you'd probably want metallic gaskets for the majority of the seals - even good vacuum seal grease and appropriate O rings will outgas before metal does. While cleaning the interior surfaces is a good starting point, instead of baking in place, you could just hold the vacuum for hours. While baking drives off what's gotten into the surface layer of the metal faster, it does still work its way out slowly without, and if you look around for pumpdown curves you'll see the sort of knee one would expect on a system that is first coming down from air and which wasn't vented with dry nitrogen or similar when brought up to atmospheric pressure. When there density of the gas gets so low, there are a lot of tiny pitfalls, and it seems like excellent system design still often needs day or more long pumpdowns to achieve around the final operating pressure on many systems.

Nice explanation of all this. But that long corrugated hose between the pumps will seriously reduce your pumping speed.

Only way I know how to get to 1e-10 torr (1e-8pa) is with ion getter pumps which function the same as a Cold Cathode Gauge

Could you have a look at those small 2$ RCWL-0516 style 3.2 GHz motion sensors? They have a doppler radar implemented with just a just one transistor and some distributed element magic. I don't have any test equipment that goes up that high, but they seem to be using a super-regenerative style circuit.

does the roughing pump vibration cause damage to the turbopump over time?

Some other comments already suggested wearing nitrile or latex gloves, but you also need to clean your orings and other contaminated parts metal parta with IPA on dust-free tissues to remove fat and dust. Then you add a tiny bit of vacuum grease to orings. This small volume still needs to pump for an hour to fully outgas from water. I would not bake these parts as you must not heat the pump, but larger chambers are baked to speed up outgassing.

You probably don't have an actual leak; you're just seeing outgassing of water molecules from all the internal metal surfaces. One thing you could change would be to make a better connection between your first and second stage pumps; the conductance of that long convoluted hose is terrible. You could speed up the outgassing by warming the high vacuum chamber, but don't warm up the turbopump. That could damage it. The high speed turbopump bearings are extremely fragile.

You don’t really need more than 10-3 mBar to do cryogenic stuff. The vacuum is mostly there as insulation and it’s diminishing returns as you go lower. A better vacuum is nice though. The big worry is that particulates will like to contaminate the coldest part, so having a getter somewhere on the cold end is useful. Also try and avoid anything you really care about being the coldest part.

I'm surprised about this arrangement. You normally never put 2 turbopumps in series unless you have a very good reason for it, you will reach the turbopumps ultimate vacuum anyways. The problem by this is also that if the pressure goes down to much on the outlet of a turbo pump the oil for the bearing could start to migrate upwards and thus run the bearing dry after a while and seize. Keep the pressure at the outlet to what the manual say and that is around some 10 mTorr or so, but that differ quite much since all depends how the pump is built and when it was built since they can have drag stages includes which means that it can accept little higher.

2:04 No, that size of pump run at 70k rpm but the bigger one runs at around 35k rpm.

That corrugated tube has fairly high impedance to ballistic flow. I am surprised that it doesn't go below 10^-8 torr with no load. You can improve the base pressure significantly by baking the moisture off the internal surfaces.

We use helium leak checkers to find vacuum leaks. You basically hook up the leak checker unit to the vacuum system which has its own turbo pump internally and a mass spec head. You then squirt a small amount of helium around each potential leak point. The unit will show the level of helium detected and hence show where the leak is.

A vacuum pump's output is it's input or output? Is the input of a vacuum pump an output or input?😢😢😢

I'm building a vacuum chamber with a Leybold TMP. It sucks.

3:56 I've only seen one catastrophically failed TMP (somehow, through 4 years in a fab) but it was complete. It was quite a large pump, about 100-150Lbs. There were no impellers on the shaft at all, just a pile of titanium shards in the bottom of the housing.

You don't need a cryo pump to get into 10e-9 torr. Turbos with many differential stages can even get down to 10e-10 torr! One thing that may be limiting your ultimate vacuum could be the conductance between the first and second turbo. That bellows between the first and second turbo is likely hurting the pumping speed of the second pump. You may want to think about running the turbos in parallel as opposed to in series and backing both with the mechanical pump that you have. The only hang up you might have is if the mechanical pump that you have cannot back both turbos at once. Turbo manufacturers will list the required pumping speed needed for backing and as a general rule of thumb, I like to have at least 2x the required backing speed. By having both turbos in parallel, their pumping speed will more or less add and should improve your ultimate vacuum substantially.

Hmmm, don't see why you would 'Back' a turbo with another turbo when you have a rotary pump of some sort there. Are you sure that middle pump is a turbo ? cos it looks a bit daft to me unless of course your worried about hydrocarbons and junk from oil based systems, anyhow interesting video with a bit of a physics crossover......cheers. You won't hit the sort of vacuum your asking for in half a minute !! patience and maybe heater tape !!!