EEVblog
HTML-код
- Опубликовано: 2 ноя 2024
- A look at the Caesium beam time and frequency standards at the Keysight standards lab in Melbourne Australia, with Peter Daly.
Forum: www.eevblog.com...
EEVblog Main Web Site: www.eevblog.com
The 2nd EEVblog Channel: / eevblog2
Support the EEVblog through Patreon!
/ eevblog
Donate With Bitcoin & Other Crypto Currencies!
www.eevblog.co...
EEVblog Amazon Store (Dave gets a cut):
astore.amazon.c...
T-Shirts: teespring.com/s...
💗 Likecoin - Coins for Likes: likecoin.pro/@...
I love these kind of super nerdy videos. In my last job I assembled a high speed signal integrity lab on a budget of about $1,000,000US for developing Thunderbolt III and USB 3.1 based storage devices which have to handle 40Gbps data rates. I went the Tek route rather than Keysight. Bias I guess. We had a 32GHz storage scope and TDR capabilities that still stand my hair on end. It's surreal to see timing events measured in femtoseconds, even big chunks of femtoseconds. Thunderbolt III PCBs use data lanes based on 1 mil trace and space where anything less than perfection equals failure.... embedded micros are a hell of a lot more fun.
Peter S Consumer electronics is getting nasty these days, even below 40Gbps. A colleague is trying to get his 4K projector running by trying out which HDMI cable might work this weekend.
Best wishes to your colleague! 4K HDMI involves 18 Gbit/s maximum bandwidth so it's more or less the exact same territory as Thunderbolt II. As an aside, all Thunderbolt devices need to include 1/2 of an HDMI converter for backwards compatibility with DisplayPort where it just becomes a pipe for HDMI. The new Intel chips include that block but I had to add it via a 3rd party part used in Intel's reference design. I don't want another HDMI signal issue as long as I live, lol. Once you're above 4GHz or so the transmission line effects seem to dominate and you switch from being an electrician to being a surface wave plumber.
Could connector wear due to normal use affect 4K?
Knuckles the Echidna I suspect that connectors are the main problem, even in new condition. Every change in impedance has effects.
Wearing the gold off connectors can certainly cause a bit of mismatch that you can easily see in time domain reflectometry but I've personally never encountered enough of a mismatch to affect overall signal integrity. Most of my experience is limited to the lab so take it with a grain of salt. HDMI uses three differential data lanes and a differential clock so usually if the connection is a little loose on one line, the other line in the pair will still swing enough to make the receiver register. I did notice on the scope that if you have a bad connector on the transmitter end the signal is much worse off than if the bad contact is at the far end of the cable and the worse the connection, the less effective the line termination becomes.
I've worked as a radio communication engineer. Our system used GPS time as sync for air traffic and land communication between base stations. Central sync source was a Meinberg-500 (high time stability GPS receiver). It can support good enough sync for around a week, even if there is not signal from antenna, but.... Once, those yankees corrected GPS for a second at the end of a year. The system gone crazy - no air traffic, despite of carriers were in the air... I've had to reboot and reclock every base station transmitter, in order to restore it working. GPS is a good in general source, but it's not very reliable, since you don't control it.
I work for a major telco in the UK. We use very similar Caesium clock sources to time our transmission and PSTN equipment. Ours are manufactured by Oscilloquartz. We also use GPS as a fallback.
Our Cs clocks are currently being replaced after approx 30 years of service. We are able to do this without massive impact s on our network due to the hierarchy of slaved sync equipment, which will free run then slowly track into the new reference sources.
because the PSTN works on network switching / packetized voice. You want your packets synchronized just like any network needs to sync their packets.
I can't imagine how the UK worked when PDH time based networks were in place
Steve Cooper The PDH network is capable of conveying timing across long distances. While the short term accuracy is compromised due to multiplexing and demultiplexing buffering, the long term stability is maintained. A slave timing shelf at the remote end is able to smooth out the jitter with a phase lock loop mechanism. Basically the slave timing shelf has an on board oscillator which will slowly track to the recovered incoming PDH signal by comparing timing phase. Over a long period of time the overall accuracy is maintained.
Fair enough :)
Every lab should have a GPSDO (either Rubidium or OCXO) and 10MHz/1pps distribution box. Its just such awesome technology and relatively cheap to build.
And vacuum lines!
dave, you now need to do a video on the importance/need for such standards and calibrations etc etc , please! thanks for the awesome vids
fradd was thinking the same thing. I really appreciate these videos but an average newbie wouldn't necessarily get it (I don't even remember hearing about it in uni!). As a DMM reseller/manufacturer it would be a great marketing tool to do a traceability for dummies video.
SO cool to see behind the curtain like this!
Really loving this series with Peter Daly
Glad you sorted the mic issue and it didn't ruin the video.
"Physikalisch-Technische Bundesanstalt" is the right spelling :p
Regardless of the fact they want to call themselves Agilent or Keysight nothing can beat the good, old Hewlett Packard branding
7:10 R.I.P subwoofer users
@Argon: I was listening the video with my studio monitors without subwoofer, and I thought it was neighbors kids jumping on the floor. So I wouldn't state that so certainly.
It's not a matter of how to use it if there is a defect in the audio
My laptop was totally unaffected and I didn't notice was it the laugh?
No, some sort of constant scratching noise from 7:10 to 8:37 and intermittent on other parts. On laptop speakers it is very unlikely to hear those low frequencies.
Psh. My frequency standard is much better. It's literally made out of subatomic particles arranged into the shape of atoms with electron fields surrounding them which have all been carefully arranged into the shape of a Casio watch.
Fascinating video, by the way! Pardon the aspie humour.
Same here, can't beat a good Casio watch, Aspie humour is fun too!
How do the companies maintaining the frequency standards handle issues relaying to global inaccuracies, for example if everyone's standards equipment has some kind of drift, wont there be some kind of global drift that happens over time? Hypothetically speaking, if you take a frequency standard from today, and went back in time and compared it to the same systems from 15 years ago, would there be a measurable drift. If so, would that impact the performance of modern equipment?
You are right and Its a big issue for the standards agencies so they are constantly trying to find 'new' reference standards. For example the 'kilogram' is defined by reference to an actual artefact - the mass of a piece of machined platinum iridium alloy (stored in France) but despite all efforts to protect it the natural aging effects are changing it at the atomic level so its drifting over time. The plan is to define all measurements by reference to things we believe are constant (at least to current technology) such as the speed of light, plancks constant and so on which won't drift unless our technology undergoes some paradigm shift.
The second is the duration of 9,192,631,770.20 Hz periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the cesium 133 atom. Since this behavior is dictated by fundamental constants, there will be no drift in the definition of a second - and therefore no drift in Hz.
(Except before this standard has been introduced)
Distelzombie Centzon Totochtin But the way this is actually measured has imprecisions that can cause slight drifts. To deal with that, the major national laboratories constantly compare their clocks through an office in France, which decides which average of the big laboratories was right then reports how much each was off. Only after receiving that report do you know precisely what time it was during the reported period. The delta is very small, but NIST is just one of the references in the official mix.
John Francis Doe
Interesting. If the method of measurement is improved (I remember there are new kinds of clocks developed now) this would at some point be no issue anymore. I thought he was thinking it was something like the kilogram issue. (Which has been solved using watt balance)
the level of inaccuracy being talked about for time is much smaller than that for mass, and is of a different nature. With ceasium clocks, the average of all "reliable" clocks is a better representation than that of any single clock, it's just the nature of how random errors work. You could add more clocks to make it more accurate. The problem with 'kg' is the definition itself is changing, so re-measuring an infinite amount of times will produce a slowly varying answer for the same question.
Dave,
I worked in the USAF doing AFMETCAL tester CAL-VER. The tractability to the standard will drive you CRA-CRA.
Dave, I'd have a small remark for that kind of video: I find that it's really difficult to follow when you interrupt the person you're talking to. That's, I think, mostly because we cannot hear you as loud and clear as the person facing the camera. The consequence is that we often cannot hear clearly what your question or remark is, as you are not as loud and the other person is still talking. To solve the issue, you may want to bring a second mike in those cases, or else try to avoid interrupting.
Anyway, very cool video :)
Yet another great video! I didn't even know this stuff existed before subscribing to your channel! Since then however I know the importance of calibration, even silly little things like HF radios it's noticeable as signals can be a few Hz even KHz out on Chinese portable SSB world band receivers with no option of calibration but my SDRPlay RSP2 software defined radio has a frequency reference in and out!
I got to maintain those exact model primary standards in the military. Nothing past putting in the occasional leap second or swapping the tube to keep the spares from going bad.
Knowing which second it is(time of day), is entirely different from knowing how long a second is.
Nice to see where some of our calibrations go....
So what is the story with speed of light measurements being out of expected tolerance around the world some time ago? I heard they all deviated more than expected.
I'm sure at the entry they had to take all your concealed dismantling tools XD
great stuff, more tours !
The GPS satellites are owned by the US Military and they have made it accessible to the civilian so they don't get lost.
Cool Video! Dave, I recommend HiPassing your audio at about 100hz on these location videos. Crazy low end at times when using capable speakers.
Dave
Have you considered having a look at a National Measurement Institute. NMI Australia has their head office and Laboratories at West Lindfield winch is about 10 km from you. If you need some contacts to do that please contact me as I was a former employee.
Dave
I am not taking on behalf of any NMI only my experience as a former employee.
Peter talked about measurement capability of Keysight to NMIA or other NMI's in their RF measurements. Sadly NMIA has very limited capability in RF measurements as it closed its RF microwave lab a few years ago because of lack of customer demand and them wanting to put more effort in areas of higher customer demand. As Peter said Keysight traces themselves to their primary lab and not the local NMI. This is a consequence of Global measurement trading where countries recognise each other's measurement capability.
Keysight is not an NMI for them traceability is different than the NMI particularly in the USA where Santa Rosa was able gain accreditation that was better in some areas than NIST as to why that is purely conjecture and I will not speculate publicly. Simply to say NMI's have to prove their capability by published key comparisons with other NMI's whereas accredited laboratories only have to convince their auditors.
After a week NIST in Colorado, USA will print out how much GPS is off... (time and/or location). Thank you for the video ! tjl P.S. Have you hooked up an Antenna to your GPS Time Source from a past video ?
Hello from Boulder, Colorado. 😎👍
NIST has the only roadside clock I would set my watch by.
Dave, does the GPS time need to be adjusted for Special Relativity? The old wives tale says that GPS didn't work until dilation was accounted for. Cheers!
I think it does, but not only that. In fact, as of general relativity, the smaller gravitational potential up there even overcompensates time dilation due to relative speed, st satellite time is *faster* than earth time.
Yupp I wanted to say this too.
Yes it does. Here's a great article showing the math behind it: www.colorado.edu/physics/phys2170/phys2170_sp15/Library_files/GPS%20(Physics%20Today%20May%202002).pdf
+Wyvrn Thanks, really great article!
Can confirm the need for SR and GR corrections. We did this in a relativity class
gorgeous outro animation
It's funny to think of that GPS does not relay upon its own frequency standards, simply because they are too imprecise.
When the astronomer did calculate the acceleration of our solar system through the Milky Way to be one atom per second, did they use quasars, that is the most precise frequency standard in existence! :-)
ForMERly Agilent, i trust :P
Two requests: 1) Please stop talking over the guest you're interviewing. 2) Would you be able to provide a link to the GPSCaesium drift data which is published publicly?
Paul
www.measurement.gov.au/Services/Pages/TimeandFrequencyDisseminationService.aspx
Is the page you are looking for at NMI Australia.
Cheers
Steve
There are cesium beam clocks inside of some the US Army communication vans.
Love the new outro.
8:16 - 8:36 Gotta get them fresh eyes on board! :P
@5:30 their chart has "scalar" misspelled as "scaler".
Bill shows the world's smallest atomic clock and then describes how the first one made in the 1950s worked. He describes in detail the use of cesium vapor to create a feedback or control loop to control a quartz oscillator. He highlights the importance of atomic team by describing briefly how a GPS receiver uses four satellites to find its position.
ruclips.net/video/p2BxAu6WZI8/видео.html
Wow the new ending is very electric and scary. I'm afraid I will have to sue for 2,000,000 in emotional anguish and lost productivity.
NPL is actully 'National Physical Laboratory' and not 'National Physics Laboratory'.
In Slow-mo the bottom time display was running faster than the top...
a frige full of time
Do you call your diesel generators UPS’s over there?
Globally there are several manufacturers of integrated rotary UPS/diesel gensets that are widely used globally.
Pillar is one manufacturer, below is a link to additional info.
www.piller.com/en-GB/262/uniblock-ubtd-rotary-diesel-ups-from-500-kw-up-to-40-mw
patw52pb1 - thanks for the info. I have only worked at sites that have separate generator/ battery ups. But I guess packaging them would be only one service contract...
You are welcome.
They are also found in installations where space for a separate UPS/battery plant and genset is unavailable or either very expensive.
This is especially true for data center installations where support infrastructure space competes with client income generating space.
With the integrated rotary flywheel UPS/diesel genset units there are also additional benefits of lower TCO, no battery maintenance/replacement, no fire marshal hydrogen gas flammability/explosion concerns, no battery acid abatement/concerns and no hazardous/toxic material disposal concerns since these units do not use batteries.
patw52pb1 - I keep my battery bank in a 70f cooled room for longer life, generators are pre heated to near 200f for instant startup.
patw52pb1 I have only once worked with a fly wheel ups - several tons of concrete spinning on air bearings. I stayed a step or two away from it ha
God, let the man speak.
Don't turn it on, take it apart!
Cool shit.
SATCOM Hell YEA!!
*formerly
Dave, I (and many others too) already said this, please stop talking over your interview partner. It is not just annoying, it makes the whole converation less understandable. Please take into consideration that not just native english speakers watching your videos. I think I understand english speaking good enough, even in some weird accent, but for me it is hard to understand if you both talk simultaneously.
Can't stop an engineer being excited about learning something from an expert :P
All it takes is lots and lots of money. Please pass the RS Micronta VOM.
OK you lost me at "airlines."
Huh?
did he say millihertz?