CuriousMarc - 2024-11-14
It's all working now! We take you through the complicated alignment procedure and atomic sync of our steampunk HP 115BR divider and clock. HP 115BR Playlist: https://www.youtube.com/playlist?list=PL-_93BVApb5-_7sWRxQ0on8vNy7_H-zxS Stuff that supports the channel: - Support us on Patreon: https://www.patreon.com/curiousmarc - Amazon links for the tools I use in the lab: https://www.curiousmarc.com/amazon-links - Channel merch on Fourthwall: https://curiousmarc-shop.fourthwall.com - Legacy channel merch on Teespring (I don't have everything transferred to Fourthwall yet): https://teespring.com/stores/curiousmarcs-store "Elevator Music" Credit: Crinoline Dreams by Kevin MacLeod Our lovely sponsors - PCBWay: fast turn PCBs, https://www.pcbway.com - Electro-Rent: https://www.electrorent.com - Keysight: test instruments: https://www.keysight.com - Samtec: connectors: https://www.samtec.com - R&D Microwaves: https://rdmicrowaves.com Get more technical info on the companion site: https://www.curiousmarc.com Contact info: https://www.youtube.com/curiousmarc/about 00:00 Reuniting the HP 115BR clock with the HP 106B oscillator 02:08 Complicated clock startup procedure 04:37 Synchronization to the HP 5061A atomic clock 08:21 Synchronization to the WWV Fort Collins signal
Marc, as a student at Stanford back in the late 1980s and early 1990s, it was my job to maintain a clock at the Big Dish. We had a Spectracom 8130 ovenized crystal standard and a WWV receiver, and I would follow the procedure you show in the video. But rather than just synchronizing our clock, I had to do some math to calculate how fast or slow our frequency standard was running and adjust its vernier knob accordingly, because the crystal sped up as it aged. I also had to adjust for the radio propagation delay between Fort Collins and Palo Alto, and even allow for the seasonal variation in the height of the ionosphere. It was a point of pride for me to keep that clock as accurate as possible!
Yes, you are 100% correct, the 10 us alignment procedure shown here is for tracking and adjusting the crystal oscillator frequency drift. The delay is orders of magnitude bigger!
@@CuriousMarc Yes, now that you mention it, we had some "slew" buttons on our custom clock to bring the delay back to zero. That's also where I learned that a linear frequency drift results in a parabolic time error, i.e. if you don't keep the frequency offset small then the time error gets big really fast. I would fit multiple time-offset measurements over many days to a parabolic curve, and then calculate the current slope of the curve to know how much to adjust the frequency. Kids these days with their GPS clocks have no idea of what fun we had.
Look! He put the connector cap over the cable before soldering it! That is an experienced connector solderer!
Of course! He is a real connector professional! After all, he works for a major company producing connectors. No way would he do the same stupid mistakes when soldering connector cables, like all the rest of us... And even if he did: being the connector company's CTO, he just couldn't show such a beginner mistake in a video... 😅😂
If you DO forget to put the shell on the wire before soldering, the number of time you repeat this is inversely proportional to the number of wires you have already soldered before discovering the error of your ways.
@@Herby-1620 It's OK if it's the first end of the cable. Then you can weigh up the complexity of getting the cap over from the far end of the cable vs. undoing it all and doing it the right way. If you then forget the cap on the second end of the cable then you should examine your choice of career or hobby :)
@@gshinglesbeen there done that…
It gets more fun with RF connectors, where you have like a bazillion different shims, sleeves, isolators, spacers, gaskets, clamp nuts you all have to stack over the cable in the correct order before you crimp the center contact, which can't be undone, and when you have to cut it off, you just wasted an expensive connector (Well .. not really, you can at least keep the parts as spares for the ones of others that inevitably well teleport away from your workbench into a parallel universe) .. of course you also have to repeat the stripping of the cable to exact lenghts.
I have heard the WWV Boulder Colorado signal many times on the HF band and now i get to see how it gets used. Well done. Enjoyed this series.
Forget about the microseconds, I just love the clock rollover to 16:00 minutes at the end of the video synchronized to the WWV voice👌 Perfection!
It would have been about 1969 that I glommed onto a very nice 5 MHz double oven oscillator designed for portable use in communications equipment. That inspired essentially generating that entire setup with poor-person's parts. I build a RadioShack clock and a prediviider from 5 MHz down to 1Hz the clock chip used. To one second it was easy to set time. The oscillator had a nice vernier frequency adjustment. I used that to advantage for calibration. I found a relatively small 24V lead acid battery. I built a UPS around it. I had my one Hz pulse. I had an oscilloscope. I triggered it with the 1 Hz. I walked the oscillator onto proper time by tweaking frequency as needed. The first day I simply adjusted until I had the rising edge of the first cycle of the 1kHz tick tone burst lined up with my 1 Hz tick and not drifting. On subsequent days I made very small adjustments of the frequency first to realign the tick and then a small tweak in the direction to reduce frequency error. By the end of a couple months that kludge appeared to be staying within about 1 part in 10^9 with no further adjustments needed. I made occasional use of it over the next most of a decade to keep it precise.
So what did I use this for? Well, I was fiddling around on two maters waiting for a nerdnet that usually appeared ad hoc in the evenings. I noticed somebody calling CQ on USB with the carrier about 144.985 MHz. Seems at the time that was the bottom end of the Technician license band less 1.5 kHz, I proved my frequency readings with a tap on the 100 kHz divider chain. I then spend about 20 minutes telling him how the FCC figures that frequency allocation and why measuring frequency of an SSB signal with a counter emphatically does not work. I never reinstalled it when we moved to the Tustin Ca area. I may be a hopeless nerd; but, I am not THAT hopeless.
{^_-} Joanne
When you do the adjustment, don't forget to add the propagation delay between Ft Collins, CO and you. It can be of several milliseconds! And it is not a straight line of the distance; you must consider the signal path bouncing off the ionosphere. Very interesting series of videos! Thank you.
I wonder what issues there are with the fact they are picking up both Ft. Collins (WWV) and Hawaii (WWVH) at the same time. You can tell that from the two announcer voices.
Also, using the average of WWV and WWVH (in Hawaii) can help ameliorate the propagation delay slightly. Unfortunately, it has only minimal effectiveness in the continental U.S. But, it is ever so slightly better than relying on just one transmitter location. The results will be best during daylight hours as ionospheric effects are then at a minimum. Averaging additional transmitters like CHU in Canada and MSF in the U.K. can help further. Again, it is important to do the averaging at a time when it is closest to solar noon in all locations. It's a fairly complex task requiring accurate equipment and is essentially an academic exercise in the age of GPS. Though it would be interesting to see the actual level of accuracy achievable using this method.
Well that delay would be huge! Yes you’d need to subtract 5ms or so. Speed of light is sooooo slow. I did not explain it well, but the goal of such a 10 us sync is to track and correct the drift of the oscillator. You let it sit for a day, a week or a month, do a couple similar sync, plot the resulting readings of the sync knob, and voila, you should know how far off your oscillator is. Correct it with the 10^-11 adjustment, and you’re good to go.
@@CuriousMarcgod, ntp got just so much more worthy.
I am glad I took the "time" to "watch" this episode before I "clocked" out
Appreciating my GPS synchronised Garmin wristwatch after watching this series!
And if you lose that appreciation, just imagine having all that kit in a backpack to do the same job.
This brings to mind what we used as a station clock for a radio site. We didn't have a clock in the traditional sense, but we had a pair of double oven crystal oscillators, and an omega navigational receiver. We didn't worry about what time it was, but we needed to have a very accurate 100 KHz clock to provide to all the radios. We had HF radios, so we needed to make sure that 29,999,999.9 Hz was spot-on. That is where the ten to the ninth precision came in for us, even though the best time keeping device on sight was probably only within a few minutes of true time :-)
I really enjoyed this series, thanks for posting the videos!
I watched all three parts back to back. HP engineering in those years continues to amaze me. And your expertise and persistence in troubleshooting is equally impressive. Thank you for another fantastic project.
It’s amazing how easily we forget that at one point, distant parts of the world were, in fact, distant, and navigating an airliner meant some guy sitting in the back of the cockpit squinting at charts and maps. Prior to super precise atomic clocks and GPS finding your way around actually required a certain level of intelligence since it meant looking at maps and having a compass handy when driving on vacation. The old fashioned gps used to be “Turn here, honey. No wait. Oh we’re going the wrong way! …no you’ve been holding the map upside down the entire time. We were supposed to turn back there.” Fun times.
squinting at maps, notes and bubble level sextant
I can smell that gear running from here!
Absolutely fascinating the amount of work it took to do back then what we can now accomplish with a single command on our PC.
With an internet link ;)
I find the consistent pulse from that radio station relaxing. As a collector of (mostly mechanical) clocks, this would be the ultimate Fallout-esque clock to have.
Absolutely Fascinating! Love this post and love your lab. 😮
Thanks so much!
Oooo! We finally get to see it all in action!
2:24 Another Wizard casually enters the sacred halls… "Hello!"
Or in other fitting words: A wizard is never late, nor is he early, he arrives precisely when he means to.
Ah well, Magic Mike flies in and out at his own convenience 🦉
@@the_jcboneAnd that “precisely” is of course calibrated to a portable cesium atomic clock
Marc's basement seems to be like a place in some open world adventure, you go there to meet people who have special superpowers. And it's always open, no key card required.
@@benjaminhanke79 No keys, but a special force field. A Nerd Field. In the End, probably Marc's wife let Mike in… "Oh hi, Mike. He's downstairs, playing with his toys. You know the way…"
Brings back memories. I had a Collins R-388. When I got it, the previous person had replaced the pear shaped 5U4:with a solid state rectifier module, but didn’t install a ballast resistor. I put it back to the correct 5U4. I believe the too high B+:over stressed the PTO causing a very slight permanent nonlinearity in the 1 MHz segment. The R-388 is really fascinating piece of gear in the band switching mechanism with various ferrite slugs moving in and out of coils as you go through the 30 1-MHz clicks of the band switch. If you’ve ever worked on a 1960s car radio, the R-388 band switch is reminiscent of the car radio slugs, but much more complex. The audio had a very nice warm sound. Compared to modern radios, the R-388 is massive, but sometimes I wish I had kept it. The problem is there’s only so much room in the basement.
Another great video! I love them all, especially anything that that has HP instrumentation. They always give me a short recess from my day to day worries.
Of course you also need to subtract the delay between the time signal transmitter site and your receiver site.
And the delay in the receiver as well.
Today, it is easier to synchronize it to the PPS output of a suitable GPS receiver 😀
Yes, and what path did the radio signal take -- it's bounced off the ionosphere if not groundwave.
Unbelievable, amazing work Marc. Great job diving into history!
Good propagation! You were receiving both WWV and WWVH.
In California it is not uncommon to receive both. Now if my WWVB clock would just receive that signal (it appears to be weaker than it was just a few (2) years ago.
@@Herby-1620 I'm up in Oregon (Portland area) and I never hear WWVH. Maybe if I got off my butt and installed an outdoor antenna...
Ah, that explains it, thanks! I was wondering while I had both voices, and why the man’s voice was so much louder than the woman’s… This was at 5 MHz by the way.
WWVH is voiced by the great Jane Barbe :)
And here I’ve just started messing around building a GPS disciplined NTP server this month. I’d much rather twist the knobs and read the dials to synchronize a clock.
Nothing an arduino with a pot could not fix. Clock skew adjustment manually 😊
I vaguely remember setting a few rubidium oscillators in my youth. I used a plotter that received signals from a national standard (France Inter? I don’t remember). It took few days to adjust one, by turning some precision potentiometer one way or the other until you obtained a straight line.😊
As far as I remember, the clocks precision was in the order of 10^-11, while the standard was 10^-12.
It was a loooong time ago…
Please wire this up to trigger a cuckoo clock to make an overly complicated way to know the top of the hour, or find a Goblin Teasmaid to start a cup of tea.
I suppose you could program a Raspberry Pi to do a graphical version of this.
Lol. That is an excellent idea. Every precision clock should have a cuckoo bird inside. We’ll sell it as the new H29 option!
Thanks Marc! That was a fun romp. 😀
Amazing to see where we started. I did a terrestrial TV station during the forced sell off and repack by the FCC and in the process demoed an old analog TV station that had occupied the space. I have alot of the older racks of equipment from there I clouding the exciter and the ovenized quartz reference. They were monitored and periodically bumped to keep the station on frequency. There were 5 racks in total for all of this. The new transmitter that replaced it was a simple affair with the exciter taking in program material via ethernet and a coax to a GPS antenna I installed on the outside of the building. Amazing how simple and cheap atomic accuracy can now be had! My reference at home is an older cellular time card with an 8 channel GPS receiver to periodically bump an ovenized and very stable quartz oscillator. Dirt cheap off of fleebay. It has a 10MHz output and an odd 2Hz output too. Connect to the serial port and you can get all sorts of info including precise location. Loved this series and we'll as most every other! Thanks so much Marc for not only making me feel stupid but also making me a little bit less so too.😁
Seems like a lot of work to go through to figure out when the coffee break starts!🤣
The only thing I was missing was to hear: "War, war never changes!"
What a beautiful setup!
CuriousMarc, I highly recommend Eric Dollard's books... He's undoubtedly the most authoritative electrical engineer in the public domain and you can't find it anywhere else...
Clock sync is fascinating. Thank you for demonstrating! Are we getting a collaboration with Mr Carlson's Lab to restore the Collins R-388 ?? 🤣🤣🤣🤣
I have that stack of HP. When I fine-tuned mine, I used the color burst signal from a television station. Turns out they use an atomic clock as their reference.
That lil tenths of a second meter/bar is so satisfying to watch, especially when it's all synchronized. Love how ingenious it is to create that w/what they had. Kinda wanna have a desk-side clock that looks like that now >v< (and is just synced w/NTP, but would be neat to play a stream of the time signal too.). Also wouldn't you'd think they'd make more exposed / easier controls for setting the rough time? Crazy how it has to be pulled all out.
You could add a switch that let's you switch between the two precision settings on your clock.
Would be a curiously awesome modification.
Lovely video as usual!
Fascinating (C) Spock.
I know many people who find setting the time on their microwave to be more complicated than syncing these clocks.
I was waiting for that. 25 years ago we would have joked about VCR clocks… but not on YouTube of course! Gosh how time files.
love hp!!
Picking up WWV is sometimes challenging up here in New England. Thankfully, CHU Canada comes in reliably, in both French and English :-)
Now, how about exploring the ionospheric sounders you can hear on the shortwave bands? "Whoop!"
good thing you have the 1950s receiver! Can't do this with a modern SDR receiver as that adds some potentially unknown 10s of milliseconds delay through the DSP demodulation and filtering before you even hear the WWV audio.
Ohhhhh, Marc, you're killing me! No backshell clamp and you didn't safety wire the threaded locking ring (with 500 turns per inch engagement) - Kidding!
I’m all messed up.
Not only microseconds, if you built your own rig, you can zero-beat it to the WWV carrier since it's also regulated to some ridiculous standard like 10^12 or some such.
Please note: You are receiving BOTH WWV and WWVH (Hawaii) (both male (WWV) and Female (WWVH) voice). The propagation time must also be taken into account (5 ms, I guess) from Colorado to California. The propagation time is longer to Hawaii. Nice Collins receiver, the civilian model was one of the 75-A series (I have a 75A2). In general time is a fickle thing. Now we have GPS to get nice accurate time all over the world. You can get GPS receivers to have 1 PPS outputs as well, but that isn't 1960's technology.
Yes, sane people should use the flying atomic clock over our heads, aka the GPS. I’m slightly deranged ;-)
Yep... that's how we sync'd the UTRAO local clock.... once a day, to WWV... NBS at the time. My next RAO job, the VLBA, we had a hydrogen maser-tempered GPS clock... we could really tell when Selected Availability was turned on! I wonder how close our cellphone network is... seems spot-on with GPS time.
atomic ticking is amazing.......
The DCF-77 signal actually can be decoded to single digit microseconds, but you need to correlate it to a well known pseudorandom sequence for that.
Lovely gear!!
thanks Marc always enjoy your videos. I would have been interesting to sync to the GPS time and see what the propagation delay is. What is our Plan B if we lose our GPS timing? Might have to resort to your timekeeping..
i think the radio "click" transmission is not supposed to be used to sync upon like that. It is likely much more suitable only to drift-compensate your local source over long term, as you could likely frequency lock on that second pulse, but locking the phase on it would not be reliable, also due to variable path of the short wave and propagation time changes there.
@CuriousMarc I enjoyed that lovely series very much! While we're talking about clocks... You're sometimes wearing a G-Shock watch, but what's the other watch?
@AlanKilian - 2024-11-14
When I got this clock the crank on the synchro didn’t turn at all.
It took several hours of careful cleaning to get it to turn and then more hours over about a week to get it turning pretty smoothly.
It looks like you’ve got it turning even better!
Thanks for this series. It’s great to see this clock in its perfect home.
@fredinit - 2024-11-14
Thanks for the contribution!
@CuriousMarc - 2024-11-15
Alan, we can’t thank you enough for the clock donation! It cranks really good now. And Eric is the Master Cranker.