Get NordVPN 2Y plan + 4 months extra plus up to 20 GB Saily data voucher ➼ nordvpn.com/networkchuck It’s risk-free with Nord’s 30-day money-back guarantee! In this video, I install a high-precision time server in my studio using the OpenTimeCard Mini from TimeBeat. Instead of relying on external Network Time Protocol (NTP) servers, I’m hosting my own time server to achieve nanosecond-level accuracy using Precision Time Protocol (PTP). The OpenTimeCard Mini combines a GPS module that receives signals from satellites equipped with atomic clocks and a Raspberry Pi Compute Module 4 (CM4) acting as the server. This setup allows me to synchronize all the clocks on my network with unprecedented precision. Additional Information: If you’re intrigued by the fascinating world of time synchronization and want to delve deeper, here are some interesting insights and resources: • The Precision of Quartz Clocks: Modern quartz crystal oscillators are remarkably accurate, typically drifting about 15 seconds per month. While factors like temperature and aging can affect them, advancements have significantly minimized these deviations. • The Birth of Atomic Clocks: The first practical atomic clock was developed in 1955 by Louis Essen at the National Physical Laboratory in the UK. By 1967, the second was redefined based on the vibrations of cesium atoms, marking a significant leap in timekeeping accuracy. • GPS Satellites and Timekeeping: The Global Positioning System (GPS) began with the launch of its first satellite in 1978. These satellites carry atomic clocks that provide precise time signals, revolutionizing navigation and global time synchronization. • Understanding NTP Synchronization: The Network Time Protocol (NTP) dynamically adjusts its synchronization intervals, typically ranging from 64 to 1,024 seconds, to maintain accurate time across devices connected to the internet. • Time Zones and Global Standardization: The concept of standardized time zones was proposed by Sir Sandford Fleming in 1879. The 1884 International Meridian Conference endorsed these time zones, paving the way for the global timekeeping system we use today. • Hardware Timestamping in Networking: Hardware timestamping captures the exact moment a network packet is sent or received, allowing for nanosecond-level precision. This is essential for applications requiring extremely accurate time synchronization, like PTP. • Raspberry Pi and Hardware Timestamping: While the Raspberry Pi Compute Module 4 (CM4) allows for hardware timestamping through additional network interfaces, the standard Raspberry Pi 4 does not support this feature on its built-in Ethernet port. The new Raspberry Pi 5 introduces hardware timestamping support directly on its Ethernet interface. • Why Nanosecond Accuracy Matters: Industries such as financial trading, where transactions happen in fractions of a second, rely heavily on precise time synchronization. Broadcasting and aerospace also require this level of accuracy to ensure seamless operations. Further Reading and Resources: • TimeBeat’s OpenTimeCard Mini: store.timebeat.app/products/open-timecard-mini • TimeBeat Software: www.timebeat.app/solutions/ • Understanding Precision Time Protocol (PTP): www.timebeat.app/blog/sync-showdown-ntp-vs-ptp-vs-tsn-vs-ethercat • History of Timekeeping: www.timebeat.app/blog/timekeeping-history-and-technology • Quartz Crystal Oscillators Explained: www.timebeat.app/blog/quartz-crystal-oscillators • Atomic Clocks and GPS Technology: www.timebeat.app/blog/atomic-clocks-gps-timing • Raspberry Pi Hardware Capabilities: www.timebeat.app/raspberry-pi-hardware-capabilities • Network Time Protocol (NTP) Details: www.timebeat.app/blog/sync-showdown-ntp-vs-ptp • Global Standard Time Zones: www.timebeat.app/blog/global-time-zones 🔥🔥Join the NetworkChuck Academy!: ntck.co/NCAcademy **Sponsored by NordVPN
I've seen people use water clocks for real. Burma, now Myanmar, they were using a water clock as part of the process for making gold leaf. You have to hammer the bundle for as long as the clock which was floating (with a hole) until it sank. Then someone else got to hammer.
Not good, instead of those network signal things.. isolating and perfectioning that vibrating atom wouldnt be better idea? so there will be no shifting
Cool, I went down the rabbit hole on being fascinated with precision time when I was doing signal level protocol work in the late 1990s. By the mid-2000s I started collecting some random lab sources, precision time gear, etc that I could find on eBay while in my professional life I was managing large scale telecom systems.
Unless you're needing precise scientific timing, or are in FinTech, NTP is more than enough for most people. If your network hierarchy is pretty small and flat, you don't even need your own NTP server - it's just added complication. Just setup DHCP to have everything sync to a specific NTP server, like NIST's. I run NTP servers for the enterprises I administer, but that's because there's thousands of servers and tens of thousands of client. It's a multi-level NTP setup for Win and Linux systems. When you have this many hosts, you'll want internal NTP. If you run a handful of hosts on a small network, you don't need your own NTP.
I'm in the broadcast industry from TV productions and live events as well as large mass distribution of streams and we use PTP to keep everything in sync with everything from the source of the video (camera) to all the way down to the output which can be in a total difference country through a web browser or a TV...
I run a stratum 1 server. Just because, someone needs to be running them I guess. But, it's easy these days. Raspberry pi with GPS (PPS) board gives microsecond accuracy locally and +/- 2ms within Europe and +/- 5ms to west coast USA. So yes, for almost everyone in the world, NTP is good enough for the accuracy to be better than you could really measure yourself as a human.
I agree. NTP is not even close enough for audio much less video. NTP accurate to a few milliseconds. PTP is accurate to a few nanoseconds. PTP is 6 orders of magnitude more accurate than NTP
Small correction, crystals don't inherently vibrate at that frequency, crystals are just "electromechanical" clocks, their cut and size and mass determines the frequency, which is very stable, and they can be tuned to extreme accuracy and with electronics, that accuracy and stability can be turned into very good clocks and other frequency references. :) 32768hz was chosen because you can divide it in many ways by 2^n, so binary counting systems can keep track of a second by measuring 2^n ticks of that clock. :)
Yes you can get much more precise quartz oscillator chips. It’s just that 32mhz was fine for average consumer need. It also needs almost nothing in terms of power to keep this running (a compact watch battery can power them for up to a decade). Now once you go beyond about twice that temperature variation plays a major part. So you have Temperature Compensating Oscillator, and then above that you have Oven Temperature Control (the quartz crystal is maintained inside a small electric oven that keeps it at a stable temperature over ambient). Of course now you can get single chip cesium and rubidium oscillators too.
@@LogicalNiko yes, and temperature variations happen partly because temperature causes the size of the crystal to change, like anything else -> altering the tuning frequency. :)
@ yep and when you get really precise gravity effects become a factor. Just change the angle and you could watch the frequency swing. In theory you could use this to track small variations in local gravity and altitude, but that’s much easier with other instruments.
The history is interesting. It started with railways. That was a huge influence on Einstein, and you get a flavour as well from these sorts of set ups and syncing. How do you sync when the signals take time?
The USAF spent a week explaining Rubidium and Cesium clocks to me but this was way more fun. Watching you realize that your antenna needed to be outside or in clear view of the sky was pretty funny. Good stuff Chuck.
@@pazsion GPS signals are transmitted at only 50W (i.e. they are as "bright" as a commodity lightbulb) from more than 3000km away. A piece of sheet-metal reliably blocks the signal. And no, a GPS signal does not use cell towers in any way. Your device _might_ use them as _additional_ information, but not the GPS chip itself.
GPS antennae are active in that they have an amplifier built-in. They're powered by a dc bias voltage from the GPS board. Also, the GPS frequencies (GHz region) have a very high attenuation per meter of cable. Antennae height is irrelevant, they must have a clear view of the sky, the greater the better. Trees are bad. Clouds are bad. I used an OEM GPS uBlox module some 20 years ago for timing purposes. Cool stuff.
I love the breakdown of the video and how much depth was put into the details. I'm no GPS expert, but I have been in situations where time was off and systems did not connect. I have also been in situations where an antenna was not positioned properly to get timing. The funny thing is that most people think GPS is easily gained from inside a building because our phones seemingly work inside with GPS. Additionally, Hollywood often gets it wrong, putting people having Iridium (satellite) calls inside ships, underground, etc and the individuals having successful conversations, but then other times showing an Iridium phone NOT having signal when they are clearly in plain view of the sky. However, outside is critical, because the signal comes from space. The building can even stop the signal from being received if too close. A long cable is normal for a setup where you have an external GPS antenna. Overall this video is great and would be very useful for people to get an understanding of how critical time is.
Assuming you guys are here looking for the video that was taken down. It was just him bashing LTT for taking down their video about CompTIA A+ exam and nitpicking every statement Linus made as if he was personally offended while trying to sell his course on training material for the same A+ exam.😂
20 years ago I was running Stratum 1 servers in my data centers throughout the US for database cluster synchronization. Also, for data replication to my offshore data centers. They were very expensive. Its good to see tech has brought the size and cost down to the homelab.
Slight correction on the quartz crystals. They resonate slightly differently based on the size of the crystal. We get them down to the size that is very useful to us. 32,768 is a very useful number for computing, as it is 2¹⁵. So we get the crystals measured precisely to a size that will resonate at a value useful to us
This can't get better than this. I'm a broadcast final year student figuring out how to setup PTP with pi for AES67 streams and Viola!! here my fav youtube friend with the video
Nice to see you discovered time protocols. My son and I started our NTP project right after we learned about using GPS hats on Raspberry PI's for robotics. It was so easy to build out a basic PI install and NTPsec and GPSd. My antenna is in the attic and the cable runs down into the basement to a splitter that feeds two Pi's, they are using PPS as primary source and then validating with two known good Stratum 1 servers. This provides the sanity needed for a known good time server. I have made on of the two servers public for IPv6 only though the NTP Pool projects with IPv6 only because with IPv4 you can get swamped with ISP that use CGNAT and send all their clients to one address the lock into. Have fun, the hole you dig will continue to go deep as you tune your solution.
The one thing that is more important than having accurate time is having synchronized time...that is, if you had to choose between having all your servers in perfect sync with each other or all but one on the correct time, choose in sync. This is because timestamps from various systems become important for log correlation. If all logs are out by a few seconds but all systems generating logs are in sync, you can correctly sequence your logs.
I worked with a lot of stuff with precision time to the nanosecond level in the past. Everything you said is exactly right. We used it for air traffic control. Not just for timekeeping of all the flight positions data and comms but also to measure time against transponder replies to determine distance of aircraft from the receiver called MLAT or multilateration. It’s really interesting stuff. Awesome video!
One of my mentors was at MIT working on tweaking the Kerberos Authentication Protocol as part of Project Athena - one of the issues was that authentication packets were time stamps and "drift" was an issue. As such there was a large window allowed - he suggested NTP, which was happening at U. Delaware at the same time. Syncing all nodes meant they could wind down the window, reducing the risk of decrypting the authentication packets. On the point of understanding a NanoSecond - Admiral Grace Hopper would hand out 11" bits of wire - this is how far an electrical impulse (or light) can travel in a nanosecond. LAter she would hand out the little restaurant paper packets of pepper - these were the same for a "picosecond". Early supercomputers (like the Crays) would have a spiral of etch on someboards, and you would tweak their timing by soldering a wire to some point on the spiral.
I remember the famous lecture by Admiral (then Captain) Grace Murray Hopper where she handed out nanoseconds. It was a wire 11.8 inches long. That's the distance electricity travels in that amount of time.
He is a UA-cam content creator, therefore has to string out his videos so they are long enough to contain multiple ads including the COMPULSORY NordVPN ad.....
Chuck's background is rooted in IT. If he's not dealt with System Time & Synchronization Issues, or RF/Microwave Signal Propagation, I suspect the GPS Signal Issues may simply be outside his 'wheelhouse.' Working in the Telecom Industry for the past +25-years, BITS Clocks & Synchronization is a daily concern for both voice & data traffic, across the network. If we loose our primary timing source (Stratum 1), which is typically our local GPS Receiver (with an outdoor antenna), we'll fallback to our secondary timing source (Stratum 2), which is a fiber-fed Signal from a remotely located Stratum 1 Timing Source. Suffice to say, these days Timing & Synchronization is crucial, for network elements to work in concert with one another. Free-running equipment will eventually lead to chaos in the Telecom & Data World.
I've had 2 NTP servers on my home network for years. One is a little dedicated unit whose sole purpose is NTP, and the other is a RPi 3B+ with a GPS HAT that I put together to feed ADS-B data to the 3 major flight tracking websites. If those both fail, my NAS is built on an RPi 4 and has a high-accuracy real-time clock attached that can take over.
My favorite subject. Time synchronization is immensely important in connected parallel computing - especially clouds and systolic arrays. Intel's latest chips finally support high precision time as do most high end network products. Think about it for a second (pun intended) - you have a bunch of nodes running in their own little time domain world. When they interact with eachother they start to have a sense of fluidity they "wish" to achieve. You want the independent requests and responses to be low latency and instantaneous. And yet every transaction is non-deterministic. For every node you add each interaction between the nodes exacerbates this meshing. Without an agreed time unification the software will spend (more like waste) time polling and responding with very low efficiency. You are not computing when you are asking for something to compute and waiting for a node to "get around" to it. Take that chaos and think of a time partitioned comb (for your hair) with teeth. At tooth #1 we all sync. At Slot #2 all look for global commands like reset. Slot #3 - Node #1 sends requests. Slot #4 - all respond to Node #1. etc. This is simple to prove a point. If you are not involved with Node #1 - you can ignore the slot and keep busy. The array of nodes because become synchro-meshed and message with minimal disruption. It is seriously game changing.
I’d love to learn more about this (systolic arrays and parallel computing) - have you got any good sources to start from? Thanks for the inspiring comment 🙂
A bit mad I wasn't able to post this in time in your hot take video about LTT taking Comptia A+ so I'm posting it here. Not an attack but I actually felt where that video came from. -------- Putting Linus' dramatic and exaggerated antics aside, I still think his points are valid. I'm from the Philippines and got my degree in IT here. What ticked me was Comptia "partnered" with some instructors and gaslight (mild exaggeration) you to taking it. They were suggesting that this is the way to get into a great employment. I mean I get that it definitely helps you become more marketable, but its not like THATS THE ONLY WAY to get in. I definitely felt their ways were predatory especially for students who is just about to graduate. We eventually didn't take it because it was too expensive. What usually happens then is most graduates went into freelancing and/or getting into more specific certifications in other areas of IT expertise which I think is a lot better.
He puts a lot of effort in picture it like it's near impossible or very hard to enter into IT without a certification, and i see a lot of that mindset in mostly all forums or videos about certs I think the first scene in LTT nailed that sentiment, and not everyone (obviously not NetworkChuck) noticed it
I work with electrical protections for substations and power generation plants and time synchronization is veeeery important. Thank you for this video 👏🏼👏🏼👏🏼
Been using your classes for a couple weeks but just saw you have an app now! 🎉🎉🎉🎉🎉! So awesome! Thanks for being you and helping me get through all this stuff without getting bored and falling asleep! 😂
We've seen the first part of the history of time measurement. And my 9-year-old daughter was interested in it. We also thought about why time measurement should be accurate. thx alot!
I mean, there are people who make youtube videos of making their own ISP with old tech, so it's not fully out of the question this guy would do it. Although, I doubt this guy would though, since he only pushes garbage he was given for free.
Working with PTP in SMPTE 2110 and AES67 systems, a lot of systems use ptp4l. There is so much about the setup for PTP that wasn’t covered in this that’s very important. Things like the PTP domain is pretty cool. I would encourage anyone who find this interesting to look into PTP and its uses. Running PTP for your homelab or personal computer is wicked overkill.
Agree. There's a lot more to PTP. I work in the broadcast world with 2110 and PTP. Oh, the joy of getting everything sync'ed. Fortunately, I come from the world of SONET/DWDM so timing isn't anything new, but the distribution of time with PTP is definitely some crazy stuff. I thought the same thing after I watched this video. PTP in homelab is overkill.
Me always watching all your videos and don’t even have a computer 🥲🥲 some day I’ll be working in cyber security I don’t know how but here is where I’m starting! Thanks for everything 🙌🏼🙌🏼
From network engineer perspective, we have these cases when routers are loosing connection to hubs, and often it is simply because CMOS dies and server is using time as part of checking licences/credentials. So different time means server will refuse to recognise device and connection is down. Precise time is keeping entire internet up.
RPi, GPS USB Dongle ($10), Chrony, instant time server. Have a couple for my HamNet portable, have a VM on home server (also fed by a $10 USB GPS). Easy-peasy.
@@timballam3675 How many homelab users need 'nanosecond' compatibility? Even PPS sync is 'extravagant' for most purposes. Nanosecond sync to a GPS source can vary across the reliability of the distance measurements alone, let alone thermal issues. If you really need to be that precise, you need a lot more equipment with a lot higher precision than a Pi (even a CM4) at a higher cost than $1000...
@@timballam3675 Then dont buy a Pi.. they are mostly overpriced anyways with loads of compromises which can be solved by putting proper hardware onto the chip.. or just having a proper System. Dont need a Pi anyways for such a task, can just use a small RISC-V based board or ESP
I have had a GPS HAT on a Pi for several years now, running as a stratum 1 NTP time server, with a stratum 2 on my router. I’ve looked at PTP several times but I always end when I look at other requirements. Most places say it requires network switches that support for PTP. Part of the PTP process is taking into consideration latency within the network and making connections based on latency.
Oh cool. I described a tech like this ten years ago to my friend. The goal is that all kinds of devices and recordings and things can be matched up for perfect multi angle recordings or reproduction.
UTC is an acronym that combines both the way coordinated universal time in english and french, which is why it's UTC. One of the few times that a wiki article gave me some good info.
I was first exposed to this 12-13 years ago in a manufacturing plant. The site manufactured intravenous medications that were encapsulated in polymer designed to break down at a constant rate in human blood, guaranteeing a consistent therapeutic dose. Insane amounts of automation and monitoring. Losing 1 second of data could result in a loss of 7-8 figures. We used another product not mentioned here that used PTP under the hood to have microsecond latency but not nanosecond at that time.
i been doing gps stuff and it can take up too 20 - 30 mins for a lock, now the cool thing is you have a warm start after getting that first signal. Next time you boot gps up it should be able to just get a lock.
Yes. Though the data gets stale after a while. My Garmin handheld GPSs will ask me to confirm if they haven't been used in a while so they can go looking for the sats where they are rather than where it thinks they should be.
6 днів тому
I used to have 3*6 antennas up per clusters... All on a ceiling. I won't dive into the details but all I am going to tell you is, there's quite a bunch of hardware work needed. Cable loss to account for, amplifiers and things to put in place in case lighting hits your antenna. IF you decide to run a long cable and deploy that antenna on your roof I highly suggest you to get help from the company who's selling you the product and ask them to be HONEST with the full cost / install aspect. Companies deploying these aren't average joe's businesses and I constantly had to go on the data center roof to do inspection and maintenance despite redundancy. If you set this up without redundancy and a secondary solution you will very likely jinx a lot of business critical services...
PTP introduces the idea of absolute time. With PTP you can measure if two events happens at the EXACT same time on opposite sides of the planet. Accurate "absolute" time makes most time syncing problems easy to solve. 1728925877 seconds since the epoch is the exact same time everywhere on the planet and in space. Accurate within a few nano seconds (0.000000001 seconds). So now we can answer the question "Did two events happen at the EXACT same time."
So this is super cool. Consider this: time is based on planetary movement, various rotations really. So accuracy may not necessarily come down to smaller and smaller fractions of time, but rather hitting a target where the timepiec's accuracy is perfectly nsync, like Justin Timberlake and uh.... Those other dudes, with the particular rotation for which you are currently interested in, ie a year being the time from the particular phase of one harvest to another or a second being a fraction of a fraction (the second fraction, which is why it called a second) of the 24 pieces of time that make up one rotation of the planet. Unless of course the earth is flat, then there is no time, only ice walls and leap frogging plane expeditions into a neverending frozen tundra!
When you get into digital communications on amateur radio time becomes very important, otherwise the delay in transmissions across the world will sometimes cause you to completely miss a transmission. Most of us end up downloading more accurate time sync software like dimension 4 and disableing the standard time sync in order to get accurate enough time to not have issues.
People are complaining about the price and I can understand why to be honest, but compared to other solutions brands like MasterClock, the price can reach 10k and above, so yes it’s cheap. Having a PTP for home use it’s over overkill , but in Video IP Based systems it’s crucial or I should say vital!
@@CodesExplorer-hb1wr to summarize he was just pretending Linus made no valid criticisms, essentially saying Linus just wanted to cheat and that Comptia is perfect and absolutely necessary to get into IT.
Oh this is handy. I remember back when ntp was hacked. There are these old protocols and services where everyone was like, "it's just a matter of time." It happened. Freejacked. That historical attack was over 20 years ago, along with true lan hopping. Good ole days. Down the road, I bought a Vishera. AMD put their own clock on board which basically doubled precision. It took a special system config in Windows. That's when I got interested in timing performance. Ultimately, clocks are least path of resistance design, follow the bottlenecks. It's pretty cool to study. That Vishera FX-6300 still runs today at under 70F, thanks to Arctic MX under every heat sink and a Harley sized CoolerMaster.
Thanks Chuck! As always another great video!! OMG, I’m so onto this “Time Server” board! I actually collect various antique time pieces, and certainly “time” is my favorite mystical subjects. I’m building a “pretend” Time Machine, and certainly will need this device! ;)
Damn, I was about to witness a beef between NetworkChuck and LinusTechTips. Luckily he withdrew like J Cole. 2024 already had so many beef i lost count. The Tech world was about to have one.
"You might be wondering why UTC is the abbreviation for Coordinated Universal Time. The acronym came about as a compromise between English and French speakers: Coordinated Universal Time would normally be abbreviated as CUT, and the French name, Temps Universel Coordonné, would be TUC. "
I believe France wanted it through Paris too so there was definitely some wrangling. The funny thing is, the line for UTC isn't even through where it's marked at Greenwich these days. Tom Scott and Stand up Maths have good videos about it.
I am really glad I watched this. Not only with computers but as a watch collector, now I am just that much more fascinated with the craftsmanship, the science it took to be created. Now wear and tear on the crystal itself is another thing. I wonder if the quartz crystal can be replaced? N
2:07 needs to be a little more specific. No quartz does not perfectly line up with a power of 2 frequency. It's close, but nowhere near close enough to be used as is. What is DOES do, is vibrate with a very high Q favtor reliably, and we can then cut the physical crystal to "tune" it exactly where we want it.
Hey NetworkChuck! Big fan of you and your videos are absolutely amazing! I just want to ask one thing, with the help of which software did you create an absolutely banger website of yours? Please tell
remember back in highschool before everyone had mobile internet and all our cellphones did have a different time. like 5 minutes apart sometimes.. now all our phones just sync automatically
My alarm clock, my microwave, my stove, and the radio in my truck have no digital way to sync. So they are all typically within 2-3 minutes when I set them. Also several of the clocks end up slowing down as time goes on, but usually the power surges reset the clocks every few months so they get reset/updated that way.
We definitely do live in a matrix, I'm currently in the middle of a hobby project making my own timeserver as well. Thing is, how do you trust the time is correct? Well, you need a minimum of 3 devices, as 1 might be wrong, if you have 2 how do you know which to trust, but with 3 and more you can say it's likely that the majority is correct. So I have 1 RPi taking the time from GPS Another asking a local precision measurement server A third that keeps track with a precise crystal. Infact 4. 3 of which again, provide a mean value for 1 main server. All those combined will feed my 1 main NTP server. I was this close of buying a large satellite dish and try to listen to pulsars and parsing data from a noise filter but a smarter friend of mine talked me back into my senses. And no, nobody needs to do this, public NTP servers already do it all for you, but fun projects are fun!
Most people dont realize how accurate timestamps of network traffic are extremely important to many tasks. Packets must be reconstructed in the same order to make the data within it valid, so jitter in the clock that is doing timestamping can cause a bad chunk of data, requiring the whole chunk to be re-transmitted. In audio these bad packets result in a cracks and pops or a "zzzzttt" sound, in video it can result in a buffering issue if the issue is really bad, or a single frame having "static looking" data in part of the frame or even all of it. In HPC program that are crunching data it means lower throughput while you wait for re-transmission, or an error in data output (unlikely unless bad coding), and higher network utilization. This is why high end server NICs have femtoclocks on them, so packets can be timestamped down to sub-nanosecond accuracy As another note, in professional audio settings with modern day equipment that uses mixing consoles and stages boxes and other rack equipment that all runs over a network, setting up which device is the master clock for your audio and making sure it is the most accurate and that every device can receive the messages accurately is extremely important. You run into a lot of clicks and pops in your audio if the network and time stamping is not set up properly or you have too low end of equipment. So I always laugh when I see people commenting on a product like an "audio NIC" meant entirely for streaming audio and saying how "it doesnt matter" and "why would anyone buy this?"
Please research your topic before posting myths like this. TCP for example doesn’t need a clock for ordering, it’s using a variant of the vector clock algorithm. Many distributed systems do this because clocks can’t be trusted. There’s drift and lag involved with clocks, but not with vector clocks.
The people who make this product used to work for Facebook. They saw the project on an open source forum and bought out all the pieces it takes to make this product and started selling it themselves.
Amateur Radio operators (especially ones who use digital modes like FT8/4 need to have as accurate of a time as possible. The connections are all started and controlled by their computers, so being synced up with each other is extremely important.
I was actually for something like this on a USB stick, a little battery backed up time keeper. People don't realize how helpful it is to have all your logs in sync.
@@foxale08 you only need 1 moment to synchronize, not sure why would need to constantly stream timer frames from a USB device, seems like the hard way to do it. an RTC on a USB stick, it can't be that hard.
Yeah, but nanosecond precision logs? come on..Most logs dont even log Milliseconds, and if they do +-1ms isnt super important as you have network delay too of +-0.5ms at best, more at worst
Network chuck is my favourite UA-camr, I am not even joking like every single day I wait for chuck to upload so I can watch it. thank you for making my day better
Get NordVPN 2Y plan + 4 months extra plus up to 20 GB Saily data voucher ➼ nordvpn.com/networkchuck
It’s risk-free with Nord’s 30-day money-back guarantee!
In this video, I install a high-precision time server in my studio using the OpenTimeCard Mini from TimeBeat. Instead of relying on external Network Time Protocol (NTP) servers, I’m hosting my own time server to achieve nanosecond-level accuracy using Precision Time Protocol (PTP).
The OpenTimeCard Mini combines a GPS module that receives signals from satellites equipped with atomic clocks and a Raspberry Pi Compute Module 4 (CM4) acting as the server. This setup allows me to synchronize all the clocks on my network with unprecedented precision.
Additional Information:
If you’re intrigued by the fascinating world of time synchronization and want to delve deeper, here are some interesting insights and resources:
• The Precision of Quartz Clocks: Modern quartz crystal oscillators are remarkably accurate, typically drifting about 15 seconds per month. While factors like temperature and aging can affect them, advancements have significantly minimized these deviations.
• The Birth of Atomic Clocks: The first practical atomic clock was developed in 1955 by Louis Essen at the National Physical Laboratory in the UK. By 1967, the second was redefined based on the vibrations of cesium atoms, marking a significant leap in timekeeping accuracy.
• GPS Satellites and Timekeeping: The Global Positioning System (GPS) began with the launch of its first satellite in 1978. These satellites carry atomic clocks that provide precise time signals, revolutionizing navigation and global time synchronization.
• Understanding NTP Synchronization: The Network Time Protocol (NTP) dynamically adjusts its synchronization intervals, typically ranging from 64 to 1,024 seconds, to maintain accurate time across devices connected to the internet.
• Time Zones and Global Standardization: The concept of standardized time zones was proposed by Sir Sandford Fleming in 1879. The 1884 International Meridian Conference endorsed these time zones, paving the way for the global timekeeping system we use today.
• Hardware Timestamping in Networking: Hardware timestamping captures the exact moment a network packet is sent or received, allowing for nanosecond-level precision. This is essential for applications requiring extremely accurate time synchronization, like PTP.
• Raspberry Pi and Hardware Timestamping: While the Raspberry Pi Compute Module 4 (CM4) allows for hardware timestamping through additional network interfaces, the standard Raspberry Pi 4 does not support this feature on its built-in Ethernet port. The new Raspberry Pi 5 introduces hardware timestamping support directly on its Ethernet interface.
• Why Nanosecond Accuracy Matters: Industries such as financial trading, where transactions happen in fractions of a second, rely heavily on precise time synchronization. Broadcasting and aerospace also require this level of accuracy to ensure seamless operations.
Further Reading and Resources:
• TimeBeat’s OpenTimeCard Mini: store.timebeat.app/products/open-timecard-mini
• TimeBeat Software: www.timebeat.app/solutions/
• Understanding Precision Time Protocol (PTP): www.timebeat.app/blog/sync-showdown-ntp-vs-ptp-vs-tsn-vs-ethercat
• History of Timekeeping: www.timebeat.app/blog/timekeeping-history-and-technology
• Quartz Crystal Oscillators Explained: www.timebeat.app/blog/quartz-crystal-oscillators
• Atomic Clocks and GPS Technology: www.timebeat.app/blog/atomic-clocks-gps-timing
• Raspberry Pi Hardware Capabilities: www.timebeat.app/raspberry-pi-hardware-capabilities
• Network Time Protocol (NTP) Details: www.timebeat.app/blog/sync-showdown-ntp-vs-ptp
• Global Standard Time Zones: www.timebeat.app/blog/global-time-zones
🔥🔥Join the NetworkChuck Academy!: ntck.co/NCAcademy
**Sponsored by NordVPN
first clock invented by muslim
I've seen people use water clocks for real. Burma, now Myanmar, they were using a water clock as part of the process for making gold leaf. You have to hammer the bundle for as long as the clock which was floating (with a hole) until it sank. Then someone else got to hammer.
The episode where we find out that Chuck isn't a ham
Not good, instead of those network signal things.. isolating and perfectioning that vibrating atom wouldnt be better idea? so there will be no shifting
Here time clock
Real time clock
My time clock
This time clock
How word are more here same meaning
My second role beside IT role is Time center techinican,im so happy that someone talk about how much time measuring is mindblowing.
and the fact that 1ft is about 1ns so you need cm accurate GPS for these time servers.
My second role beside IT role is Time center techinican
so your second job is more IT
Cool, I went down the rabbit hole on being fascinated with precision time when I was doing signal level protocol work in the late 1990s. By the mid-2000s I started collecting some random lab sources, precision time gear, etc that I could find on eBay while in my professional life I was managing large scale telecom systems.
Unless you're needing precise scientific timing, or are in FinTech, NTP is more than enough for most people. If your network hierarchy is pretty small and flat, you don't even need your own NTP server - it's just added complication. Just setup DHCP to have everything sync to a specific NTP server, like NIST's. I run NTP servers for the enterprises I administer, but that's because there's thousands of servers and tens of thousands of client. It's a multi-level NTP setup for Win and Linux systems. When you have this many hosts, you'll want internal NTP. If you run a handful of hosts on a small network, you don't need your own NTP.
And if you want a bit more accuracy connect to a stratum 1 server.
I'm in the broadcast industry from TV productions and live events as well as large mass distribution of streams and we use PTP to keep everything in sync with everything from the source of the video (camera) to all the way down to the output which can be in a total difference country through a web browser or a TV...
I run a stratum 1 server. Just because, someone needs to be running them I guess. But, it's easy these days. Raspberry pi with GPS (PPS) board gives microsecond accuracy locally and +/- 2ms within Europe and +/- 5ms to west coast USA.
So yes, for almost everyone in the world, NTP is good enough for the accuracy to be better than you could really measure yourself as a human.
@@attackhelicopter-up3dh I believe the TimeBeat server is a stratum 1 server.
I agree. NTP is not even close enough for audio much less video. NTP accurate to a few milliseconds. PTP is accurate to a few nanoseconds. PTP is 6 orders of magnitude more accurate than NTP
Just wait until he starts synchronizing his coffee breaks with atomic clocks
Now I have a guide on how to synchronize my coffee poops
Don't you mean "coffee breaks"
@@dromerdev procedure for buying atomic clock require government perrmision
Or he turns into Sheldon.
it's only a matter of TIME.
Geezus... That's a frikken $1200 board + $70 for the CM4 adapter + $120 for the CM4 + $80 for the antenna. Cool info, but that's a hard pass from me.
that seems cheep i just looked at the website and a "Open Timecard" was £2495. Time is expensive
Exactly what I thought. At $200 I'd buy one, but 1200 is a quick pass.
@@McGregorMX He got paid to peddle this crap to you.
@@kyle98927 I did it way way cheaper - Hackaday - ESP32 NTP Time Server (Stratum 1)
@@kyle98927 not crap, normal people have no need for this accurate of time.
Small correction, crystals don't inherently vibrate at that frequency, crystals are just "electromechanical" clocks, their cut and size and mass determines the frequency, which is very stable, and they can be tuned to extreme accuracy and with electronics, that accuracy and stability can be turned into very good clocks and other frequency references. :)
32768hz was chosen because you can divide it in many ways by 2^n, so binary counting systems can keep track of a second by measuring 2^n ticks of that clock. :)
History is cool.
Yes you can get much more precise quartz oscillator chips. It’s just that 32mhz was fine for average consumer need. It also needs almost nothing in terms of power to keep this running (a compact watch battery can power them for up to a decade).
Now once you go beyond about twice that temperature variation plays a major part. So you have Temperature Compensating Oscillator, and then above that you have Oven Temperature Control (the quartz crystal is maintained inside a small electric oven that keeps it at a stable temperature over ambient).
Of course now you can get single chip cesium and rubidium oscillators too.
@@LogicalNiko yes, and temperature variations happen partly because temperature causes the size of the crystal to change, like anything else -> altering the tuning frequency. :)
@ yep and when you get really precise gravity effects become a factor. Just change the angle and you could watch the frequency swing.
In theory you could use this to track small variations in local gravity and altitude, but that’s much easier with other instruments.
That history lesson was amazing
The history is interesting. It started with railways. That was a huge influence on Einstein, and you get a flavour as well from these sorts of set ups and syncing. How do you sync when the signals take time?
The USAF spent a week explaining Rubidium and Cesium clocks to me but this was way more fun. Watching you realize that your antenna needed to be outside or in clear view of the sky was pretty funny. Good stuff Chuck.
LOL. 3C2 Tech Control! I remember them FCC-100's and Promina's
he shouldnt need to since the signal penetrates, and triangulates between 3 or more satalites. and cell towers.
??
Just wait until the quantum chicanery clocks come out, even better than modern atomic clocks and they use nanokelvin atoms to do it.
@@pazsion GPS signals are transmitted at only 50W (i.e. they are as "bright" as a commodity lightbulb) from more than 3000km away. A piece of sheet-metal reliably blocks the signal.
And no, a GPS signal does not use cell towers in any way. Your device _might_ use them as _additional_ information, but not the GPS chip itself.
GPS antennae are active in that they have an amplifier built-in. They're powered by a dc bias voltage from the GPS board. Also, the GPS frequencies (GHz region) have a very high attenuation per meter of cable. Antennae height is irrelevant, they must have a clear view of the sky, the greater the better. Trees are bad. Clouds are bad. I used an OEM GPS uBlox module some 20 years ago for timing purposes. Cool stuff.
I love the breakdown of the video and how much depth was put into the details. I'm no GPS expert, but I have been in situations where time was off and systems did not connect. I have also been in situations where an antenna was not positioned properly to get timing.
The funny thing is that most people think GPS is easily gained from inside a building because our phones seemingly work inside with GPS. Additionally, Hollywood often gets it wrong, putting people having Iridium (satellite) calls inside ships, underground, etc and the individuals having successful conversations, but then other times showing an Iridium phone NOT having signal when they are clearly in plain view of the sky.
However, outside is critical, because the signal comes from space. The building can even stop the signal from being received if too close. A long cable is normal for a setup where you have an external GPS antenna.
Overall this video is great and would be very useful for people to get an understanding of how critical time is.
Assuming you guys are here looking for the video that was taken down. It was just him bashing LTT for taking down their video about CompTIA A+ exam and nitpicking every statement Linus made as if he was personally offended while trying to sell his course on training material for the same A+ exam.😂
I like NetworkChuck, but yeah this pretty much sums it up 😂
@@SirDonald Same but this was out of pocket for no reason lol
Been on this channel for years and still enjoy the videos!
20 years ago I was running Stratum 1 servers in my data centers throughout the US for database cluster synchronization. Also, for data replication to my offshore data centers. They were very expensive. Its good to see tech has brought the size and cost down to the homelab.
Slight correction on the quartz crystals. They resonate slightly differently based on the size of the crystal. We get them down to the size that is very useful to us. 32,768 is a very useful number for computing, as it is 2¹⁵. So we get the crystals measured precisely to a size that will resonate at a value useful to us
I was looking for this comment. Thanks!
this is blowing my mind rn
This can't get better than this. I'm a broadcast final year student figuring out how to setup PTP with pi for AES67 streams and Viola!! here my fav youtube friend with the video
Nice to see you discovered time protocols. My son and I started our NTP project right after we learned about using GPS hats on Raspberry PI's for robotics. It was so easy to build out a basic PI install and NTPsec and GPSd. My antenna is in the attic and the cable runs down into the basement to a splitter that feeds two Pi's, they are using PPS as primary source and then validating with two known good Stratum 1 servers. This provides the sanity needed for a known good time server. I have made on of the two servers public for IPv6 only though the NTP Pool projects with IPv6 only because with IPv4 you can get swamped with ISP that use CGNAT and send all their clients to one address the lock into. Have fun, the hole you dig will continue to go deep as you tune your solution.
The one thing that is more important than having accurate time is having synchronized time...that is, if you had to choose between having all your servers in perfect sync with each other or all but one on the correct time, choose in sync. This is because timestamps from various systems become important for log correlation. If all logs are out by a few seconds but all systems generating logs are in sync, you can correctly sequence your logs.
Very interesting stuff! Thanks for sharing! The NTP Project folks appreciate you bringing attention to our work!
You didnt know about PTP???? wow... a little BEHIND THE TIMES!!!!
...thank you, i'm here all evening.
I worked with a lot of stuff with precision time to the nanosecond level in the past. Everything you said is exactly right. We used it for air traffic control. Not just for timekeeping of all the flight positions data and comms but also to measure time against transponder replies to determine distance of aircraft from the receiver called MLAT or multilateration. It’s really interesting stuff. Awesome video!
One of my mentors was at MIT working on tweaking the Kerberos Authentication Protocol as part of Project Athena - one of the issues was that authentication packets were time stamps and "drift" was an issue. As such there was a large window allowed - he suggested NTP, which was happening at U. Delaware at the same time. Syncing all nodes meant they could wind down the window, reducing the risk of decrypting the authentication packets.
On the point of understanding a NanoSecond - Admiral Grace Hopper would hand out 11" bits of wire - this is how far an electrical impulse (or light) can travel in a nanosecond. LAter she would hand out the little restaurant paper packets of pepper - these were the same for a "picosecond".
Early supercomputers (like the Crays) would have a spiral of etch on someboards, and you would tweak their timing by soldering a wire to some point on the spiral.
Came for a networking lesson, stayed for the history on time. what a journey. Great content once again.
I remember the famous lecture by Admiral (then Captain) Grace Murray Hopper where she handed out nanoseconds. It was a wire 11.8 inches long. That's the distance electricity travels in that amount of time.
By the way, the Navy put a version of it on UA-cam.
Not quite at the speed of light, it depends on the cable velocity factor.
After 15 minutes and 3 commercials, he realizes he has to put the GPS antenna outside ????
Occasionally people get lucky and have a roof that doesn't totally destroy the signal, but that isn't always the case.
He is a UA-cam content creator, therefore has to string out his videos so they are long enough to contain multiple ads including the COMPULSORY NordVPN ad.....
Because he's an i_di*t.
Chuck's background is rooted in IT. If he's not dealt with System Time & Synchronization Issues, or RF/Microwave Signal Propagation, I suspect the GPS Signal Issues may simply be outside his 'wheelhouse.'
Working in the Telecom Industry for the past +25-years, BITS Clocks & Synchronization is a daily concern for both voice & data traffic, across the network. If we loose our primary timing source (Stratum 1), which is typically our local GPS Receiver (with an outdoor antenna), we'll fallback to our secondary timing source (Stratum 2), which is a fiber-fed Signal from a remotely located Stratum 1 Timing Source.
Suffice to say, these days Timing & Synchronization is crucial, for network elements to work in concert with one another. Free-running equipment will eventually lead to chaos in the Telecom & Data World.
block them all
To the editor of the video: THANK YOU for taking out the coffee slurps!
Damn, that's cool, I need one of the...
£1,187.85
I don't need one of these
Time sync is also super important for live TV broadcast as well.
This guy is so ahead of the game, time travel in the rack
Can't wait to see the video about the disappearance of the video about the disappearance of the Linus video.
The only reason I clicked on this video. Linus had a much better video.
I've had 2 NTP servers on my home network for years. One is a little dedicated unit whose sole purpose is NTP, and the other is a RPi 3B+ with a GPS HAT that I put together to feed ADS-B data to the 3 major flight tracking websites. If those both fail, my NAS is built on an RPi 4 and has a high-accuracy real-time clock attached that can take over.
My favorite subject. Time synchronization is immensely important in connected parallel computing - especially clouds and systolic arrays. Intel's latest chips finally support high precision time as do most high end network products. Think about it for a second (pun intended) - you have a bunch of nodes running in their own little time domain world. When they interact with eachother they start to have a sense of fluidity they "wish" to achieve. You want the independent requests and responses to be low latency and instantaneous. And yet every transaction is non-deterministic. For every node you add each interaction between the nodes exacerbates this meshing. Without an agreed time unification the software will spend (more like waste) time polling and responding with very low efficiency. You are not computing when you are asking for something to compute and waiting for a node to "get around" to it. Take that chaos and think of a time partitioned comb (for your hair) with teeth. At tooth #1 we all sync. At Slot #2 all look for global commands like reset. Slot #3 - Node #1 sends requests. Slot #4 - all respond to Node #1. etc. This is simple to prove a point. If you are not involved with Node #1 - you can ignore the slot and keep busy. The array of nodes because become synchro-meshed and message with minimal disruption. It is seriously game changing.
I’d love to learn more about this (systolic arrays and parallel computing) - have you got any good sources to start from? Thanks for the inspiring comment 🙂
A bit mad I wasn't able to post this in time in your hot take video about LTT taking Comptia A+ so I'm posting it here. Not an attack but I actually felt where that video came from.
--------
Putting Linus' dramatic and exaggerated antics aside, I still think his points are valid. I'm from the Philippines and got my degree in IT here. What ticked me was Comptia "partnered" with some instructors and gaslight (mild exaggeration) you to taking it. They were suggesting that this is the way to get into a great employment. I mean I get that it definitely helps you become more marketable, but its not like THATS THE ONLY WAY to get in. I definitely felt their ways were predatory especially for students who is just about to graduate. We eventually didn't take it because it was too expensive. What usually happens then is most graduates went into freelancing and/or getting into more specific certifications in other areas of IT expertise which I think is a lot better.
NetworkChuck is a shill, thats what.
He puts a lot of effort in picture it like it's near impossible or very hard to enter into IT without a certification, and i see a lot of that mindset in mostly all forums or videos about certs
I think the first scene in LTT nailed that sentiment, and not everyone (obviously not NetworkChuck) noticed it
I work with electrical protections for substations and power generation plants and time synchronization is veeeery important. Thank you for this video 👏🏼👏🏼👏🏼
Its so much important that somone finally explained the importance of time server
Been using your classes for a couple weeks but just saw you have an app now! 🎉🎉🎉🎉🎉! So awesome! Thanks for being you and helping me get through all this stuff without getting bored and falling asleep! 😂
I used PTP when we were deploying 5G in our network. Base stations need nanosecond level precision for 5G to work. Nice video!
Also, it can take 10 minutes for the GPS to find the sats. The doppler shift means searching the frequency band can take a while.
We've seen the first part of the history of time measurement. And my 9-year-old daughter was interested in it. We also thought about why time measurement should be accurate. thx alot!
the 60 second history was so crazy good! do more of that frfr!
can't wait for the day when he says i want to host the whole internet myself
just a joke 😂😂
or is it??
I mean, there are people who make youtube videos of making their own ISP with old tech, so it's not fully out of the question this guy would do it. Although, I doubt this guy would though, since he only pushes garbage he was given for free.
@@NathanChambers You are kind of right
I was in the middle of commenting on your LTT video before it became private 💀
NetworkChuck is a shill, thats what.
@@wpgspecbpls dude stop commenting this under every comment u don't know whether he is a shill or not
Ah man you’re videos are killing me, good thing I have more time now. Love it!!
Working with PTP in SMPTE 2110 and AES67 systems, a lot of systems use ptp4l.
There is so much about the setup for PTP that wasn’t covered in this that’s very important. Things like the PTP domain is pretty cool.
I would encourage anyone who find this interesting to look into PTP and its uses. Running PTP for your homelab or personal computer is wicked overkill.
Agree. There's a lot more to PTP. I work in the broadcast world with 2110 and PTP. Oh, the joy of getting everything sync'ed. Fortunately, I come from the world of SONET/DWDM so timing isn't anything new, but the distribution of time with PTP is definitely some crazy stuff. I thought the same thing after I watched this video. PTP in homelab is overkill.
Me always watching all your videos and don’t even have a computer 🥲🥲 some day I’ll be working in cyber security I don’t know how but here is where I’m starting! Thanks for everything 🙌🏼🙌🏼
This is geeky to the nanosecond! Love it! Great video, keep doing amazing content
Dang I learn a lot from this guy. My dad told what he knows about time.
From network engineer perspective, we have these cases when routers are loosing connection to hubs, and often it is simply because CMOS dies and server is using time as part of checking licences/credentials. So different time means server will refuse to recognise device and connection is down. Precise time is keeping entire internet up.
Wow you came to Japan!
Please come back, you are always welcome :)
I'm getting my R.Pi tomorrow, cant wait to test it.
RPi, GPS USB Dongle ($10), Chrony, instant time server. Have a couple for my HamNet portable, have a VM on home server (also fed by a $10 USB GPS). Easy-peasy.
The standard pi doesn't have the PPS input unlike the CM4 (this is broken out on the CM4 dev board) you also need a better GNSS board.
@@timballam3675 How many homelab users need 'nanosecond' compatibility? Even PPS sync is 'extravagant' for most purposes. Nanosecond sync to a GPS source can vary across the reliability of the distance measurements alone, let alone thermal issues. If you really need to be that precise, you need a lot more equipment with a lot higher precision than a Pi (even a CM4) at a higher cost than $1000...
@@timballam3675 Then dont buy a Pi.. they are mostly overpriced anyways with loads of compromises which can be solved by putting proper hardware onto the chip.. or just having a proper System.
Dont need a Pi anyways for such a task, can just use a small RISC-V based board or ESP
I have had a GPS HAT on a Pi for several years now, running as a stratum 1 NTP time server, with a stratum 2 on my router.
I’ve looked at PTP several times but I always end when I look at other requirements. Most places say it requires network switches that support for PTP. Part of the PTP process is taking into consideration latency within the network and making connections based on latency.
I came for the networking and coffee, i stayed for the puns and dad jokes
My dad loved pendulum style clocks. When asked why he was obsessed with them, he simply said, "It's about time." :)
Oh cool. I described a tech like this ten years ago to my friend. The goal is that all kinds of devices and recordings and things can be matched up for perfect multi angle recordings or reproduction.
UTC is an acronym that combines both the way coordinated universal time in english and french, which is why it's UTC. One of the few times that a wiki article gave me some good info.
This solution is less expensive compared similar products on the market. Worth a look!
That one in the box would be useful to put on a wall close to outside and the short antenna. Very educational video. Thanks.
I was first exposed to this 12-13 years ago in a manufacturing plant. The site manufactured intravenous medications that were encapsulated in polymer designed to break down at a constant rate in human blood, guaranteeing a consistent therapeutic dose. Insane amounts of automation and monitoring. Losing 1 second of data could result in a loss of 7-8 figures. We used another product not mentioned here that used PTP under the hood to have microsecond latency but not nanosecond at that time.
i been doing gps stuff and it can take up too 20 - 30 mins for a lock, now the cool thing is you have a warm start after getting that first signal. Next time you boot gps up it should be able to just get a lock.
Yes. Though the data gets stale after a while. My Garmin handheld GPSs will ask me to confirm if they haven't been used in a while so they can go looking for the sats where they are rather than where it thinks they should be.
I used to have 3*6 antennas up per clusters... All on a ceiling. I won't dive into the details but all I am going to tell you is, there's quite a bunch of hardware work needed. Cable loss to account for, amplifiers and things to put in place in case lighting hits your antenna. IF you decide to run a long cable and deploy that antenna on your roof I highly suggest you to get help from the company who's selling you the product and ask them to be HONEST with the full cost / install aspect. Companies deploying these aren't average joe's businesses and I constantly had to go on the data center roof to do inspection and maintenance despite redundancy. If you set this up without redundancy and a secondary solution you will very likely jinx a lot of business critical services...
PTP introduces the idea of absolute time. With PTP you can measure if two events happens at the EXACT same time on opposite sides of the planet. Accurate "absolute" time makes most time syncing problems easy to solve. 1728925877 seconds since the epoch is the exact same time everywhere on the planet and in space. Accurate within a few nano seconds (0.000000001 seconds).
So now we can answer the question "Did two events happen at the EXACT same time."
That time history montage is crazy!
So this is super cool. Consider this: time is based on planetary movement, various rotations really. So accuracy may not necessarily come down to smaller and smaller fractions of time, but rather hitting a target where the timepiec's accuracy is perfectly nsync, like Justin Timberlake and uh.... Those other dudes, with the particular rotation for which you are currently interested in, ie a year being the time from the particular phase of one harvest to another or a second being a fraction of a fraction (the second fraction, which is why it called a second) of the 24 pieces of time that make up one rotation of the planet. Unless of course the earth is flat, then there is no time, only ice walls and leap frogging plane expeditions into a neverending frozen tundra!
When you get into digital communications on amateur radio time becomes very important, otherwise the delay in transmissions across the world will sometimes cause you to completely miss a transmission. Most of us end up downloading more accurate time sync software like dimension 4 and disableing the standard time sync in order to get accurate enough time to not have issues.
People are complaining about the price and I can understand why to be honest, but compared to other solutions brands like MasterClock, the price can reach 10k and above, so yes it’s cheap. Having a PTP for home use it’s over overkill , but in Video IP Based systems it’s crucial or I should say vital!
i don’t need a time server, i let my mom tell me the time
I'm glad I was able to watch the video about Linus and CompTIA+ before it went down. But I wonder what happened
He was being criticized for shilling and not addressing Linus in good faith.
Can you tell me what he said because i couldn't watch it
@@CodesExplorer-hb1wr to summarize he was just pretending Linus made no valid criticisms, essentially saying Linus just wanted to cheat and that Comptia is perfect and absolutely necessary to get into IT.
@@LouisGalarza i can't believe Chuck did this, he's always been soo cool
@@CodesExplorer-hb1wr Im subbed to him (after not having any clue who he was) only to post "Shill" on every video he makes.
Nice to see something different discussed for Homelab.
your videographer is insane!
fun video to watch but the price tag on that thing dayum son =))) £900 ouch
Yea, you can better connect to a stratum 1 time server.
@@attackhelicopter-up3dh i’ve recently built one for £100 so a bit of a difference
Chuck sensed that I had my junk together this morning so he decided to scramble my neurons with this.
I like the documentary like explanations for things, you should do it more with other protocols, networking things
Really nice video like always Chuck 😊
Oh this is handy. I remember back when ntp was hacked. There are these old protocols and services where everyone was like, "it's just a matter of time." It happened. Freejacked. That historical attack was over 20 years ago, along with true lan hopping. Good ole days. Down the road, I bought a Vishera. AMD put their own clock on board which basically doubled precision. It took a special system config in Windows. That's when I got interested in timing performance. Ultimately, clocks are least path of resistance design, follow the bottlenecks. It's pretty cool to study. That Vishera FX-6300 still runs today at under 70F, thanks to Arctic MX under every heat sink and a Harley sized CoolerMaster.
Thanks Chuck! As always another great video!! OMG, I’m so onto this “Time Server” board! I actually collect various antique time pieces, and certainly “time” is my favorite mystical subjects. I’m building a “pretend” Time Machine, and certainly will need this device! ;)
Good catch on the edge firewall :) it's always the SOURCE.
Damn, I was about to witness a beef between NetworkChuck and LinusTechTips. Luckily he withdrew like J Cole. 2024 already had so many beef i lost count. The Tech world was about to have one.
Shill. Shill. Shill.
Time for another NetworkChuck video, where he'll go over yet another product I don't need in my life.
"You might be wondering why UTC is the abbreviation for Coordinated Universal Time. The acronym came about as a compromise between English and French speakers: Coordinated Universal Time would normally be abbreviated as CUT, and the French name, Temps Universel Coordonné, would be TUC. "
I believe France wanted it through Paris too so there was definitely some wrangling. The funny thing is, the line for UTC isn't even through where it's marked at Greenwich these days. Tom Scott and Stand up Maths have good videos about it.
I am really glad I watched this. Not only with computers but as a watch collector, now I am just that much more fascinated with the craftsmanship, the science it took to be created. Now wear and tear on the crystal itself is another thing. I wonder if the quartz crystal can be replaced? N
2:07 needs to be a little more specific. No quartz does not perfectly line up with a power of 2 frequency. It's close, but nowhere near close enough to be used as is. What is DOES do, is vibrate with a very high Q favtor reliably, and we can then cut the physical crystal to "tune" it exactly where we want it.
true. he makes it sound like it is just a coincidence that quarz in general vibrates at this multiple of 2
“I don’t need other satellites 📡, i need mine so let’s setup 😌”
As I was watching your Linus Reaction video it also got taken down lol
Wtf i was about to watch that
What a weird thing
@@tjk_prince Be glad you didn't - it was painful, you would think Linus made an entire video bashing Chuck with how personal he took it.
Great video chuck👌
Dang it chuck! Why you gotta be so entertaining, inspiring and persuasive. I’m too poor for this (in all honesty your channel’s so fun! Subscribed!)
sweet learned something new today ... Thanks chuck!
Networkchuck hasn't been getting enough views. That's the reason for all the sponsorships bro got to feed his family
Hey NetworkChuck! Big fan of you and your videos are absolutely amazing! I just want to ask one thing, with the help of which software did you create an absolutely banger website of yours?
Please tell
That's great. I was looking for someone to explain it to me erratically.
remember back in highschool before everyone had mobile internet and all our cellphones did have a different time. like 5 minutes apart sometimes.. now all our phones just sync automatically
amazing video, love an acturate close. BTW, are you still using Lastpass??
When I saw Lastpass I paused and jumped to the comments. Wth is still using Lastpass?
My alarm clock, my microwave, my stove, and the radio in my truck have no digital way to sync. So they are all typically within 2-3 minutes when I set them. Also several of the clocks end up slowing down as time goes on, but usually the power surges reset the clocks every few months so they get reset/updated that way.
Radio set alarm clocks are great. You don't have to do math first thing when you wake up.
oh god that 'tokyo drift' killed me lol
overall, love the video, fascinating!
We definitely do live in a matrix, I'm currently in the middle of a hobby project making my own timeserver as well.
Thing is, how do you trust the time is correct?
Well, you need a minimum of 3 devices, as 1 might be wrong, if you have 2 how do you know which to trust, but with 3 and more you can say it's likely that the majority is correct.
So I have 1 RPi taking the time from GPS
Another asking a local precision measurement server
A third that keeps track with a precise crystal. Infact 4. 3 of which again, provide a mean value for 1 main server.
All those combined will feed my 1 main NTP server.
I was this close of buying a large satellite dish and try to listen to pulsars and parsing data from a noise filter but a smarter friend of mine talked me back into my senses.
And no, nobody needs to do this, public NTP servers already do it all for you, but fun projects are fun!
Most people dont realize how accurate timestamps of network traffic are extremely important to many tasks. Packets must be reconstructed in the same order to make the data within it valid, so jitter in the clock that is doing timestamping can cause a bad chunk of data, requiring the whole chunk to be re-transmitted. In audio these bad packets result in a cracks and pops or a "zzzzttt" sound, in video it can result in a buffering issue if the issue is really bad, or a single frame having "static looking" data in part of the frame or even all of it. In HPC program that are crunching data it means lower throughput while you wait for re-transmission, or an error in data output (unlikely unless bad coding), and higher network utilization. This is why high end server NICs have femtoclocks on them, so packets can be timestamped down to sub-nanosecond accuracy
As another note, in professional audio settings with modern day equipment that uses mixing consoles and stages boxes and other rack equipment that all runs over a network, setting up which device is the master clock for your audio and making sure it is the most accurate and that every device can receive the messages accurately is extremely important. You run into a lot of clicks and pops in your audio if the network and time stamping is not set up properly or you have too low end of equipment. So I always laugh when I see people commenting on a product like an "audio NIC" meant entirely for streaming audio and saying how "it doesnt matter" and "why would anyone buy this?"
Please research your topic before posting myths like this. TCP for example doesn’t need a clock for ordering, it’s using a variant of the vector clock algorithm. Many distributed systems do this because clocks can’t be trusted. There’s drift and lag involved with clocks, but not with vector clocks.
This is crazy! So in-detail. Amazing; Near future U must build a time machine!
The people who make this product used to work for Facebook. They saw the project on an open source forum and bought out all the pieces it takes to make this product and started selling it themselves.
Amateur Radio operators (especially ones who use digital modes like FT8/4 need to have as accurate of a time as possible. The connections are all started and controlled by their computers, so being synced up with each other is extremely important.
I was actually for something like this on a USB stick, a little battery backed up time keeper. People don't realize how helpful it is to have all your logs in sync.
USB has too much jitter for high precision timing.
@@foxale08 you only need 1 moment to synchronize, not sure why would need to constantly stream timer frames from a USB device, seems like the hard way to do it. an RTC on a USB stick, it can't be that hard.
@@BobHenderson-dr2wyIf you don't constantly sync you defeat the purpose of having ultra high precision timing as the host will drift.
Yeah, but nanosecond precision logs? come on..Most logs dont even log Milliseconds, and if they do +-1ms isnt super important as you have network delay too of +-0.5ms at best, more at worst
@@tarakivu8861 If you only need millisecond precision you don't need GPS or PTP, regular NTP is fine.
Network chuck is my favourite UA-camr, I am not even joking like every single day I wait for chuck to upload so I can watch it. thank you for making my day better
video next year "I built my own satellites to synchronize time better"