Question - the bootstrap startup process: click the button, flick the motor That's all inside the case - you have to dismantle that box every time you get a powercut/cycle the power on the clock? Cheers.
The main workings seem to be on rails inside the box, so I think you undo the two big thumbscrews on each side (labelled 'open'), and slide it out just far enough to flick the switch and spin the motor. Then slide it back in and screw it shut. I guess you're intended to leave it permanently powered along with the atomic clock.
Yes you do have to open it up to restart it. But the clock is supposed to be powered by the crystal source, which has one internal and one external battery backup (that's why the input power is 24V DC and it has no mains plug). We'll demonstrate that when we run it with the HP 106B. The clock is useless anyhow as soon as it stops, you have to realign it to an atomic clock no matter what, so in practice that's not a concern. What would be terrible is that a non knowledgeable person could restart it easily and you would not notice that it had lost atomic traceable time.
@@CuriousMarc Yes it should only be restart-able by men in white jackets (or whatever time boffins wear) with lots of letters after their names, front panel button proders (as in every day domestic clock restarters) need not apply for these jobs.
I remember seeing an Amiga (late 80s/early 90s computer) for sale on Ebay where the vendor advertised that they had replaced ALL the capacitors as a matter of course!
@@zh84 Sadly, this seems to be commonplace. Inevitably, it introduces new problems, as Adrian's Digital Basement found recently. He replaced the caps on a Mac SE/30 motherboard and in doing so accidentally bridged a 5V signal line to the 12V power rail. I've seen all sorts of other problems caused by recapping - reverse polarity, massively wrong value, under voltage spec etc.
@@Zadster I have seen folk replace smoothing capacitors with much higher values "to make it better", only to find that the increased inrush current blows up the supply.
You know, that generation of engineers continues to astound me with what they could do with so few options. How one can create a sort of “minimalist complexity” still looks a bit like magic to me, wizardry I say.
@benjaminhanke79 I takes TH-cam a while to render the higher res. If you just watched all was finished. NeillABliss was a bit hasty commenting on the quality.
@@bennylloyd-willner9667 Some of the video was shot out of focus with a comment over the top apologising for the "potato vision" (e.g. 11:18) - nothing to do with TH-cam's processing.
That tripler circuit (6:40) is at once crude and elegant: driving the thing so hard it distorts wildly then using the fact that a single ended circuit will have dominant odd order harmonics and picking off the third. So clever.
In the late 60s, I worked at a NASA STADAN station at Winkfield UK. We had a visit from the NASA calibration team. They flew around the world calibrating the test equipment. They also carried an atomic clock to check each station's clock.
As a 60 year old mechanical engineer watching this, please, please, please clean the mechanical clock mechanism as per the service manual. Do it before it is too late. The tight tolerances of the bearings and the gearing are being abnormally worn every time you start the motor. The clue that it needs to be cleaned is that you must give the motor a very strong torque to start it spinning. You should be able to start the motor by a simple twist of the knurled handle between your thumb and index. Also HP would never have let such a noisy mechanism out the door. It should be whisper quiet.
I have absolutely no justification but I want one after seeing this video. The engineering is gorgeous, and that display is a work of art. Thanks for the continued awesome videos!
Dr, you have no idea how much I like the content on your channel; The analysis, the theory he uses makes his video super interesting. I thank you infinitely for all your effort and time in preparing each video. A cordial hug from Mexico
25:50 That's wild, the tripler is happily working as a quadrupler instead, thanks to its asymmetric single-transistor design evidently also generating plenty of _even_ harmonics and not just the odd harmonics you'd get from nicely symmetrical clipping.
This is a great example of why I so admire early HP engineering. Just Incredible stuff, from the year I was BORN! Remember, engineers used slide rules in those days! AND their BRAINS! 👍
@Runco990 Of course, but then with these modern design tools folks today do amazing things on top of them. Back in the day, they did the best they could with precision mechanicals. No one is doing that today, because it's hugely inefficient. I'll bet if the folks from those days had today's equipment, they'd use it too. Just as almost no one today, given competence, would choose a typewriter over a word processor.
@@henrivanbemmel Maybe you mis-understood my comment. It's amazing how clever engineers were back when they had such few tools to solve problems that may have seemed impossible. Today, we take assistive tools for granted because anyone can have access to them and use them. What seems easy to us today, must have been an unbelievable challenge when you had to be first ever to do it.
@@Runco990 Indeed, but today's tools remove much of the scut work allowing today's scientists, who are every bit as smart, move things ahead in situations that are far more abstract. When something has not been done before its always hard as it requires imagination both in theory and experimental apparatus choice and augmentation. When doing original discovery often the instrument of the day is being pushed near its maximum and signal must be teased from noise. It is effectively the same issue at most times in history. Finding the parallax of a star in 1838 would be then as challenging as using today's telescopes to measure the properties of an exoplanet's atmosphere. In both cases the equipment of the day is being pushed near its limits and the resulting noisy must be carefully reduced. I feel it's a matter of degree. Millikan's oil drop experimental was a masterful piece of work in 1909, today it is done some high schools, but it is no longer discovery as we no longer need the imagination and the apparatus is pre-designed. I agree though with the history aspect. Seeing such elegant equipment is like other historical devices of influence such as decent padlocks or a steam engine is amazing and I'm a fan.
I'm a software engineer who considered being an electrical engineer when I was in high school in the 1970's. I like your methodical approach to debugging the various problems without making unwarranted assumptions. It is similar to how I debug a misbehaving program or OS kernel crash. It's a shame that so few professionals in our respective fields seem to have good debugging skills.
I once had an old transistor radio where the transistors were mounted in holes in the chassis with little rubber grommets. Another example of tube-era construction techniques applied to transistors. I guess transistors were expensive back then and deserved a bit of pampering.
Amazing bit of kit, transitional technologies are just so fascinating... love the germanium transistor mounts. Also, yours is the only channel which makes me want a noise blanker in the audio path
This device appeared in the recent Nova on PBS. Just for a few frames, but it was so obvious after having seen Part 1. Will try to get a link and offset for y'all.
Just watched it. Great episode on how quantum mechanics affects everyday life: www.pbs.org/wgbh/nova/video/decoding-the-universe-quantum/ . The HP 115BR appears briefly at 17:35, followed by the HP 5061A Caesium clock equipped with the later integrated Patek Phillipe divider and clock. Awesome science documentary, they are right up there with Veritasium. Or is it the other way around?
It's interesting that they reduced an oscillator to a mere ringing circuit. Some of the 80's game consoles included a crystal resonant circuit followed by a buffer amp for that same purpose, but this is the first time I've seen one that used feedback.
Suspect the tantalums! Great job Marc. Love the point-to-point electronics, so easy to work on. That synchro motor is noisy... perhaps some Hamnmond tone generator oil is needed? I just bought some Kroil Microil, looks like the exact same thing... the mechanics sound a lot like a Hammond organ that wants oil!
Syncrho motor is fine, but there is a bad bearing in the clock mechanism. Replacing it is a bear though, you have to take all the gears out, so I haven't done that yet.
Wow, that world tour to synchronize clocks sounds fascinating, it's History, significance and all. Is there a documentary or an interesting resource on the subject?
37:56 Interesting how they fastened the seatbelts! However, it looks like the equipment was travelling economy class. That's a bit of an insult to H.P. equipment. Should be first class all the way! At least the cabin attendants gave them blankets and pillows.
I figure the unit would be on long term/indefinitely hence the exclusion of an auto start for the dividers and synchronous motor but could one be added?
Are the start switches duplicated on the outside? Or do you have to remove the unit from the rack and open it to press buttons and spin the motor every time you want to start it
Now we know what clock was used by the producers of ST:TOS that was next to Sulu when the ship was flung backwards through time on (I think) "The Naked Time".
The "old" HP hardware is just beautifully designed and constructed, much nicer than the plastic stuff that's cranked out these days. It is somewhat odd that one needs to "kickstart" the device (open the rack?) every time?
One question I have - since we know from relativity that both gravity and velocity have an impact on clocks - how did they take it into account when flying clocks around the world? Granted, it's likely not to be a huge effect if you're just going by the mechanical clock with 10us precision, but is that the reason? The effect of relativity is simply ignored because the clocks weren't precise enough?
Effects of relativity did not matter for this round, which was a 1 µs sync. It starts to matter when you want to do better than a few ns sync and fly your clocks real fast.
Ah! The second you said "tantalum" it became clear. Unlike most capacitors, the failure mode of a tantalum cap is a short rather than an open circuit. It would have flooded your mixer with a bias, my guess saturating the transistor or introducing distortion at the very least. On a side note, bad tantalum capacitors is a common cause of the failure of many early PC motherboards. When you measure a short between one of the power rails and ground before powering it up, guess what the culprit is 95% of the time. (FYI... typing this on my HP laptop)
Remove the top and start the oscillators. Plus unscrew the plate over the wheels and readjust the time. Plus hook a scope to WWV and adjust the tick. Make a prayer to the Gods of horology (don't forget that step). It's a whole ceremony to get this clock restarted.
I wonder, how is the startup sequence designed if the unit is installed in a rack? You can't reach the buttons for starting the loops nor reach the synchro?
Also, given the unit has to be manually started - is it just that it has a battery backup to keep it running or you literally have to keep removing the covers, push two switches and spin the motor everytime you use it?
The one thing I don't understand is, do you have to open it up and flick the switches and spin the motor every time you turn it on to use it!? Since it's a clock I assume you would leave it running often, but the idea of having to take it out of the sealed rack and remove the cover plate to spin the motor sounds... cumbersome!
My guess is that interfering with either of those elements would destroy any precise timing that had been established. It would be critical that they were not touched during operation. The internal coverplate doesn't need to be removed to spin the motor.
Done on purpose, so the clock can't be changed easily. Also, it stops dead in its tracks if the 100 kHz disappears for the shortest moment, as the regenerative divider will stop and won't restart. It's a "feature", so it will not restart and indicate a time that's not traceable to the atomic standard without anyone noticing.
Presumably in its new life, the unit won’t be used in any terribly critical applications. I wonder whether the Arduino could be repurposed into a discreet start-up sequencer. “Pushing” buttons would be easy. What’s a good way to spin up the motor? A tiny tiny version of an automotive starter motor would be ridiculous, but fun. 😁
sorry that I did not understand but, how originally this motor starts itself? is it that right now You give it a little twist that it can run on single phase, but to start itself does it needs a two phases ?
The dial can be zeroed afrer the correct time is set on a counter after manual start? Or need another one of these instrument to get the exact time and get the 0.0000 on the dial?
They run about 10 years until the 7 grams of Caesium in the tube get consumed. Otherwise they don't age. They still make the same tubes 60 years later! $20-30k a pop for a new tube, and you have to wait over 6 month, they can't supply the world's demand right now.
It's a whole adjustment ceremony! Open the rack and slide it out. Remove the many screws and remove the metal plate that's over the wheels. Rotate the minute and the hours wheels by hand, there is a little lever to free them up while adjusting. The seconds has a clutch spring lever clutch, so you unclutch it and hold it. Start the clock by spinning the motor. Release the clutch at the right moment. Then adjust the tick by cranking the handle while looking at WWV on a scope. I need to make a video on that!
20 วันที่ผ่านมา
Wow! HP magic again! Did they really have to open up the cover to start it up every time? Understood that the sync would be lost if a dude pressed on these start up buttons. But HP should have found a way to make it a protected front panel feature, or worst case, rear panel accessible.
Anyone else watch the latest Nova about quantum physics. There was a brief glance of the 115BR in the section about atomic clocks. Close up of the second fraction bar.
So often it’s a capacitor! I get grumpy at them - fair enough for a 60’s era germanium transistor to die - but how hard can it be to make a sheet of conductive material and a sheet of insulator?!! Very hard apparently!!
These were MIL-spec tantalums. Much better than electrolytics. But in our experience they fail, mostly due to corrosion, with a leakage fault rather than loosing any capacitance. So they create hard to catch faults. But we catch them eventually.
It is mostly about keeping atmosphere out. The seals are crucial. I shiver when I see people bend the leads without pliers to keep the force off the seals, or when they push them in TTH boards with another size footprint.
1kHz signals have to be single most annoying tone in existence. In the Apollo series, it was somewhat bearable, but the sound of the stationary, synchronous motor is driving me nuts.
It's been running continuously since I have it repaired and coupled to the 106B. No 1000 Hz noise once it's running. The bad bearing though is annoying. We need to replace it.
A question. During your repair us used that manual start buttons and had to give the motor a small spin. Is that not automated? If not, does it mean the operator of that device has to open it, turn it on and then close it again? Or how is that meant to work in day to day operations?
Yes and yes. It's not automated - on purpose. You don't want the clock to restart on its own and you not noticing the time is not traceable to an atomic standard anymore. You need to open it all up to readjust it and restart it. It's really meant to be operated with battery back up, as well as the crystal or atomic source, and never be readjusted save for the microsecond adjustment. None of that daylight saving stuff...
@@CuriousMarc Awesome. Thanks. And thank you for your great work in general. In particular in the analog field of electronics I don't know that much, but your explanations are at a level I can get the concepts.
Do you know something,with a really smart,humble guy like ken,as part of this group of talented engineers,i would bet that he would be able to write a cool piece of machine learning code to go through the whole video post production, and refocus the potato frames,whilst marc sips his coffee.No more potato frames uploaded to TH-cam. This is doable right? I hope ken isnt going to groan at this suggestion!! You might even put out a video as to how he did it. I know,I'm really pushing it! Maybe people will miss the potato.😂😂
Many synchronous motors have a 50/50 chance of turning either direction. Henry Warren invented and patented a self starting synchronous motor in the 1918 which always turned the same direction. Warren’s Telechron company was bought by US’s General Electric. Most other manufacturers’ clocks had a manual start knob to get the motor turning the correct direction. I’ve seen a few old clocks with a ratchet mechanism which would jam if turning backwards and bump the motor to turning the desired direction. Apparently HP didn’t want to pay royalties for a self starting synchronous motor along Warren’s design. Warren had the idea of using the AC power grid to transmit an accurate time reference frequency to home clocks which would count cycles via synchronous motors gear divided to display time on a clock dial. In the 1920s, there weren’t national grids and line frequency was approximate. Warren invented a companion power station master clock to monitor grid frequency. I have one in my clock collection. The master clock consisted of an electric clock section using a Telechron rotor motor. There was also a high accuracy pendulum clock. The two clocks were connected to a differential which was attached to a pointer to indicate need to speed up or slow down the power station alternator. The Telechron motor also kept the mainspring of the pendulum clock wound. The pendulum clock was regulated by listening for a radio time signal and using tweezers to add or remove tiny weights on tray on the pendulum rod.
@@wtmayhew cool, NTP before NTP. I guess, why not sync clock from it now, at least in the US and other 60Hz countries. For anyone on 50Hz, well, sorry?
@@c1ph3rpunk The rotor assembly of a Telechron motor is enclosed in a brass or aluminum housing pressed by hand into the laminated pole assembly with the field coil. The rotor assembly encloses a gear train to drive a 1 RPM (usually) output shaft. The whole assembly with gears was called “the rotor.” Rotors were available for at least 25, 50 and 60 Hz and possibly others. The same field coil assembly could be used for all the rotor types. The master clock I have originally had a 25 Hz rotor but the person who owned the clock before me swapped it for a 60 Hz rotor
I don't know, but where others posed similar questions, people answered to the effect that it wasn't that accurate a comparison where relativity would make a difference.
Hang on a sec, wouldn't flying the atomic clock around the world throw off its time? I remember a story about how they tested Einstein's theory of special relativity by flying around an atomic clock and comparing it to one on the ground, and the two showed a difference, so if they flew one of these around to synchronise atomic clocks around the world, wouldn't they still be different by whatever the difference was from flying it around the world? Or were they somehow able to compensate for the flights?
@@CuriousMarc That explains it. I was really surprised not all of the tantalums were failing short after all those years... (yes, I've been burned by tantalums... now MLCC seem to become a fitting successor to those.
th-cam.com/video/0kdnvCuWXV4/w-d-xo.htmlsi=xnMYzuFPLv7d90qS&t=1496 This is correct. You have series (current) resonance at input and parallel (voltage) resonance at output. So, at 1 kHz input voltage drops to near zero since of filter input impedance drops in resonance. I'm glad you didn't go this path too far away.
@@CuriousMarc Take look at the schematics. There is a tuned transformer that has transistor bias winding between transistor collector and -12 V rail and two other windings, one feeds frequency tripler and second that is actually divider output. Primary side path is not obvious, since bias winding, frequency tripler winding and capacitor forms parallel resonance tank that you mentioned, but frequency tripler winding, the capacitor and transistor collector forms series resonance path. My guess is that tripler winding + cap is about 1 kHz, but tripler winding + bias winding + cap frequency is lower (because there two coils in series). Proof is you can actually see two resonances at input: series at 1 kHz and parallel at about 700 Hz.
Hmm, that explanation does not make any sense. The winding 1-2 is not a bias winding at all. This is simply a transistor shooting into a proper LC parallel tank circuit load, with the collector injection tied to 2 instead of 3 so the low output impedance of the transistor doesn't kill the Q of the tank circuit. That's how all tripler stages are made too, they are also all parallel LC resonant. But in the triplers, this is the only source of resonance. However, on the mixer, the secondary winding (4-7) **is** wired in a series notch configuration! I suppose that's to kill the 19 kHz mixing product, has nothing to do with the 1 kHz resonance. Also there is C319 across pin 2 and 7 connecting the input and output windings, which I would not expect there. I have not yet figured out why it is added. I wonder if you'd get some unplanned resonances when you excite at 3 as I did with the experiment. I should have tried to excite it at 2, but that would have required me to disconnect the collector of Q302. If I ever re-open the thing, I should try to excite at 2 with Q302 disconnected, and see if I get clean 1 kHz resonances. Still not clear to me why I got a resonance at 700 Hz, but it is not due to the parallel tank circuit between 1-3 and C318, that's for sure. Removing C318 caps there did not change that low resonance at all. Now I think it might have been simply due to my test setup and parasitic capacitance from the generator. This would go away, or at least change drastically, if I excited at 2.
@@CuriousMarc The problem is that transformer is far from ideal. One is small in relation to frequency, windings has to be very tiny cross section. This causes problems like weak coupling between coils, hundreds of Ohms coil resistances and leakage inductance of course. There is also dubious to make a primary winding tap on pin 2, since transistor collector is actually current source so impedance is rather high at this point. Maybe this was emitter tap at early stage of design? Besides, Q of LC tank is already killed by internal transformer resistance. I did some simulation in QUCS. I assumed that primary winding sections are 13.2 and 330 mH (pins 1-2 and 2-3), there is also 10% inductance leakage and 20 + 100 Ω resistances. Coupling factor is 0.7. Load capacitor is 100 nF. Result didn't surprised me: coil tap 1-2 is mainly DC current path for transistor, most energy is injected while primary winding is in series resonance. There is also weak parallel resonance at lower frequency too. This exercise confirmed me also why we should use Meissner topology instead of Hartley when designing oscillators in kHz range (I know that this is mixer not oscillator). BTW I guess that C319 advances signal phase or maybe improves weak primary to secondary side coupling. It's also hard to say what is internal transformer parasitic capacitance. Could be 700 Hz resonance is just internal to transformer. I'm wondering why they didn't use RC network or some mechanical resonator instead?
Question - the bootstrap startup process: click the button, flick the motor
That's all inside the case - you have to dismantle that box every time you get a powercut/cycle the power on the clock?
Cheers.
The main workings seem to be on rails inside the box, so I think you undo the two big thumbscrews on each side (labelled 'open'), and slide it out just far enough to flick the switch and spin the motor. Then slide it back in and screw it shut.
I guess you're intended to leave it permanently powered along with the atomic clock.
Yes you do have to open it up to restart it. But the clock is supposed to be powered by the crystal source, which has one internal and one external battery backup (that's why the input power is 24V DC and it has no mains plug). We'll demonstrate that when we run it with the HP 106B. The clock is useless anyhow as soon as it stops, you have to realign it to an atomic clock no matter what, so in practice that's not a concern. What would be terrible is that a non knowledgeable person could restart it easily and you would not notice that it had lost atomic traceable time.
@@CuriousMarc Yes it should only be restart-able by men in white jackets (or whatever time boffins wear) with lots of letters after their names, front panel button proders (as in every day domestic clock restarters) need not apply for these jobs.
@@paulstubbs7678 you forgot the fancy pants! :)
@@MarcoTedaldi they come with the jackets 🧐
Bad ancient tantalums?
Who would have thought!
Tantalizing repair.
I remember seeing an Amiga (late 80s/early 90s computer) for sale on Ebay where the vendor advertised that they had replaced ALL the capacitors as a matter of course!
Good old tantrum capacitors!
@@zh84 Sadly, this seems to be commonplace. Inevitably, it introduces new problems, as Adrian's Digital Basement found recently. He replaced the caps on a Mac SE/30 motherboard and in doing so accidentally bridged a 5V signal line to the 12V power rail. I've seen all sorts of other problems caused by recapping - reverse polarity, massively wrong value, under voltage spec etc.
This old RF jock thought 'bad caps" right away. If my semiconductors are good, every electrolytic/tantalum cap becomes the next suspect.
@@Zadster I have seen folk replace smoothing capacitors with much higher values "to make it better", only to find that the increased inrush current blows up the supply.
You know, that generation of engineers continues to astound me with what they could do with so few options. How one can create a sort of “minimalist complexity” still looks a bit like magic to me, wizardry I say.
We often don't appreciate how good we have it.
If I drink enough, my vision phases out the potato vision.
You were just early. Now it's up to 2160p.
Only with alcohol from potatoes
I watch everything on my phone, I barely noticed the potato vision.
@benjaminhanke79 I takes TH-cam a while to render the higher res. If you just watched all was finished. NeillABliss was a bit hasty commenting on the quality.
@@bennylloyd-willner9667 Some of the video was shot out of focus with a comment over the top apologising for the "potato vision" (e.g. 11:18) - nothing to do with TH-cam's processing.
Marc: “This is going to be interesting.”
Yes. Yes it is. 👍
That tripler circuit (6:40) is at once crude and elegant: driving the thing so hard it distorts wildly then using the fact that a single ended circuit will have dominant odd order harmonics and picking off the third. So clever.
Hardware from this era was beautiful. Really enjoying this series.
"Miracle dividers" - I couldn't agree more. That 4 transistor divider circuit is sheer analogue genius. Thank you for another very watchable episode.
Four boxes in the block diagram, four transistors. Those were the days.
New Curious Marc videos showing up in my feed always make my day!
In the late 60s, I worked at a NASA STADAN station at Winkfield UK. We had a visit from the NASA calibration team. They flew around the world calibrating the test equipment. They also carried an atomic clock to check each station's clock.
As a 60 year old mechanical engineer watching this, please, please, please clean the mechanical clock mechanism as per the service manual. Do it before it is too late. The tight tolerances of the bearings and the gearing are being abnormally worn every time you start the motor. The clue that it needs to be cleaned is that you must give the motor a very strong torque to start it spinning. You should be able to start the motor by a simple twist of the knurled handle between your thumb and index. Also HP would never have let such a noisy mechanism out the door. It should be whisper quiet.
We know. Clock is clean. There is a bad bearing in the mechanism. It's very difficult to change, haven't done it yet.
Great video. I love your explanations, demos and your test equipment. Quite a project!
Thanks! I am feasting on short clips from your old NASA videos too to look for Apollo comms equipment, I hope you don't mind...
I must be weird too : I enjoyed watching the various frequencies on the scope! 😮
… and a blurry schematic too!😂
I have absolutely no justification but I want one after seeing this video. The engineering is gorgeous, and that display is a work of art.
Thanks for the continued awesome videos!
Dr, you have no idea how much I like the content on your channel; The analysis, the theory he uses makes his video super interesting. I thank you infinitely for all your effort and time in preparing each video. A cordial hug from Mexico
25:50 That's wild, the tripler is happily working as a quadrupler instead, thanks to its asymmetric single-transistor design evidently also generating plenty of _even_ harmonics and not just the odd harmonics you'd get from nicely symmetrical clipping.
Oh, good catch! That's exactly what's happening.
This is a great example of why I so admire early HP engineering. Just Incredible stuff, from the year I was BORN! Remember, engineers used slide rules in those days! AND their BRAINS! 👍
They use their brains today as well, just a different age.
@@henrivanbemmel Sure, but today we use so many computer assisted design tools. You have to admit, engineers did some amazing things with very little.
@Runco990 Of course, but then with these modern design tools folks today do amazing things on top of them. Back in the day, they did the best they could with precision mechanicals. No one is doing that today, because it's hugely inefficient. I'll bet if the folks from those days had today's equipment, they'd use it too. Just as almost no one today, given competence, would choose a typewriter over a word processor.
@@henrivanbemmel Maybe you mis-understood my comment. It's amazing how clever engineers were back when they had such few tools to solve problems that may have seemed impossible. Today, we take assistive tools for granted because anyone can have access to them and use them. What seems easy to us today, must have been an unbelievable challenge when you had to be first ever to do it.
@@Runco990 Indeed, but today's tools remove much of the scut work allowing today's scientists, who are every bit as smart, move things ahead in situations that are far more abstract. When something has not been done before its always hard as it requires imagination both in theory and experimental apparatus choice and augmentation. When doing original discovery often the instrument of the day is being pushed near its maximum and signal must be teased from noise. It is effectively the same issue at most times in history. Finding the parallax of a star in 1838 would be then as challenging as using today's telescopes to measure the properties of an exoplanet's atmosphere. In both cases the equipment of the day is being pushed near its limits and the resulting noisy must be carefully reduced.
I feel it's a matter of degree. Millikan's oil drop experimental was a masterful piece of work in 1909, today it is done some high schools, but it is no longer discovery as we no longer need the imagination and the apparatus is pre-designed.
I agree though with the history aspect. Seeing such elegant equipment is like other historical devices of influence such as decent padlocks or a steam engine is amazing and I'm a fan.
37:56 I love that within the picture within the airplane they have a quad outlet box and three of the knockouts are knocked out on the side.
I'm a software engineer who considered being an electrical engineer when I was in high school in the 1970's. I like your methodical approach to debugging the various problems without making unwarranted assumptions. It is similar to how I debug a misbehaving program or OS kernel crash. It's a shame that so few professionals in our respective fields seem to have good debugging skills.
Thanks Marc for doing these videos, like everyone's comments those HP engineers and manufacturing infrastructure were next level.
5:40am & curious Marc. Breakfast and a cup of tea….perfect end to the week
I once had an old transistor radio where the transistors were mounted in holes in the chassis with little rubber grommets. Another example of tube-era construction techniques applied to transistors. I guess transistors were expensive back then and deserved a bit of pampering.
Merci Marc, another great video.
Nice work loop circuits always gave me a fit when I was working in electronics. Thanks for the video Marc.
A few good beers and 40min of HP goodness repair. That was a nice evening. Thank Marc.
Amazing bit of kit, transitional technologies are just so fascinating... love the germanium transistor mounts. Also, yours is the only channel which makes me want a noise blanker in the audio path
"If you follow the channel?" Of course I follow the channel!
Perfect Friday morning content!
Another great video. Please keep making them!
This is super! A lot of creative engineering from 196s
You really do have the best toys.
This device appeared in the recent Nova on PBS. Just for a few frames, but it was so obvious after having seen Part 1. Will try to get a link and offset for y'all.
Just watched it. Great episode on how quantum mechanics affects everyday life: www.pbs.org/wgbh/nova/video/decoding-the-universe-quantum/ . The HP 115BR appears briefly at 17:35, followed by the HP 5061A Caesium clock equipped with the later integrated Patek Phillipe divider and clock. Awesome science documentary, they are right up there with Veritasium. Or is it the other way around?
@@CuriousMarc Video not available for german users :(
Perhaps it's available to international viewers via TH-cam: th-cam.com/video/t06aTX9jM34/w-d-xo.htmlsi=98Ojo9C3zVJAuWLl&t=1061
At Curious Mark's™ "We've Upped Our Quality. Now, Up Yours"...
It's interesting that they reduced an oscillator to a mere ringing circuit. Some of the 80's game consoles included a crystal resonant circuit followed by a buffer amp for that same purpose, but this is the first time I've seen one that used feedback.
when working with transistors, grab'm by the DCs and the ACs will follow. Cool video, thanks.
Well said!
Another great one.
As beautiful as these constructions are, I'm grateful for the invention of PCBs when I see Marc chasing faults in this 3D labyrinth.
Salute Dr Paul Eisler
True, but nothing beats the smell of molten solder from the stone age 😊
PCBs and ICs are widely over-rated. You don't need them at all ;-)
Suspect the tantalums! Great job Marc. Love the point-to-point electronics, so easy to work on. That synchro motor is noisy... perhaps some Hamnmond tone generator oil is needed? I just bought some Kroil Microil, looks like the exact same thing... the mechanics sound a lot like a Hammond organ that wants oil!
Syncrho motor is fine, but there is a bad bearing in the clock mechanism. Replacing it is a bear though, you have to take all the gears out, so I haven't done that yet.
@@CuriousMarc Are you going much further? Bearings, lubrication and cleaning? 😎 I'm waiting for that ...
What's the purpose of centiseconds markings on the wheel? For use as timestamp on short-exposure photos?
Wow, that world tour to synchronize clocks sounds fascinating, it's History, significance and all. Is there a documentary or an interesting resource on the subject?
37:56 Interesting how they fastened the seatbelts! However, it looks like the equipment was travelling economy class. That's a bit of an insult to H.P. equipment. Should be first class all the way!
At least the cabin attendants gave them blankets and pillows.
I figure the unit would be on long term/indefinitely hence the exclusion of an auto start for the dividers and synchronous motor but could one be added?
Testing if the coils are resonant - I half expected you to break out a GDO!
Are the start switches duplicated on the outside? Or do you have to remove the unit from the rack and open it to press buttons and spin the motor every time you want to start it
Now we know what clock was used by the producers of ST:TOS that was next to Sulu when the ship was flung backwards through time on (I think) "The Naked Time".
The "old" HP hardware is just beautifully designed and constructed, much nicer than the plastic stuff that's cranked out these days. It is somewhat odd that one needs to "kickstart" the device (open the rack?) every time?
And they strapped it in a seat on the plane!😂 Well, it would be a more gentle flight than if it were bolted to the floor.
Very cool
One question I have - since we know from relativity that both gravity and velocity have an impact on clocks - how did they take it into account when flying clocks around the world? Granted, it's likely not to be a huge effect if you're just going by the mechanical clock with 10us precision, but is that the reason? The effect of relativity is simply ignored because the clocks weren't precise enough?
I was thinking the same. Didn't they fly atomic clocks around in planes as one of the ways to prove relativity?
Just a guess but Marc may showed it in another video: you phase sync to identical and take off. After touchdown you compare the shift.
Effects of relativity did not matter for this round, which was a 1 µs sync. It starts to matter when you want to do better than a few ns sync and fly your clocks real fast.
I would say the changing inertial reference frames due to gravity and acceleration were not significant for this level of accuracy.
11:31 “Use your imagination to see this clearly” this would make an engineers life so much easier… 😂
3:06 it's like electronic surgery.. Nurse... soldering iron! Stat!
Ah! The second you said "tantalum" it became clear. Unlike most capacitors, the failure mode of a tantalum cap is a short rather than an open circuit. It would have flooded your mixer with a bias, my guess saturating the transistor or introducing distortion at the very least.
On a side note, bad tantalum capacitors is a common cause of the failure of many early PC motherboards. When you measure a short between one of the power rails and ground before powering it up, guess what the culprit is 95% of the time.
(FYI... typing this on my HP laptop)
that tone counteracts your voice some how and downs out your voice at all volume levels lol 😂 😂 love the follow up ...
It is AGC (automatic gain control) on the camera, you can thank it for saving your sense of hearing :p
@@artej11 On this channel you're not allowed to use the acronym AGC for anything other than the Apollo guidance computer.🤭
As Spock would say: Fascinating 😉
I was going to write exactly that comment. Spock is known to have very good taste 🤩
So, you have to remove the top to start the oscillators and motor manually every time, or can it be done without ?
Remove the top and start the oscillators. Plus unscrew the plate over the wheels and readjust the time. Plus hook a scope to WWV and adjust the tick. Make a prayer to the Gods of horology (don't forget that step). It's a whole ceremony to get this clock restarted.
I wonder, how is the startup sequence designed if the unit is installed in a rack? You can't reach the buttons for starting the loops nor reach the synchro?
I was just about to ask the same thing! Take it out of the rack, remove the cover, push the button then spin the motor? There must be another way!
Now I know why my 5071A’s and Cs4000’s have the 100kHz output. I’d always wondered what that could possibly be useful for.
1. Get an old HP clock
2. Open and repair it
3. Find an free Arduino
4. Profit!
Also, given the unit has to be manually started - is it just that it has a battery backup to keep it running or you literally have to keep removing the covers, push two switches and spin the motor everytime you use it?
The one thing I don't understand is, do you have to open it up and flick the switches and spin the motor every time you turn it on to use it!?
Since it's a clock I assume you would leave it running often, but the idea of having to take it out of the sealed rack and remove the cover plate to spin the motor sounds... cumbersome!
My guess is that interfering with either of those elements would destroy any precise timing that had been established. It would be critical that they were not touched during operation. The internal coverplate doesn't need to be removed to spin the motor.
Done on purpose, so the clock can't be changed easily. Also, it stops dead in its tracks if the 100 kHz disappears for the shortest moment, as the regenerative divider will stop and won't restart. It's a "feature", so it will not restart and indicate a time that's not traceable to the atomic standard without anyone noticing.
@@CuriousMarcI would have thought that it should be accessible from the front behind a cover with a tamper proof seal.
Presumably in its new life, the unit won’t be used in any terribly critical applications. I wonder whether the Arduino could be repurposed into a discreet start-up sequencer. “Pushing” buttons would be easy. What’s a good way to spin up the motor? A tiny tiny version of an automotive starter motor would be ridiculous, but fun. 😁
sorry that I did not understand but, how originally this motor starts itself? is it that right now You give it a little twist that it can run on single phase, but to start itself does it needs a two phases ?
It looks as though it was always designed to be started manually.
The dial can be zeroed afrer the correct time is set on a counter after manual start?
Or need another one of these instrument to get the exact time and get the 0.0000 on the dial?
Great Repair Mark! For how many years can one of these HP cesium clocks be used? Is the isotop becoming weaker?
They run about 10 years until the 7 grams of Caesium in the tube get consumed. Otherwise they don't age. They still make the same tubes 60 years later! $20-30k a pop for a new tube, and you have to wait over 6 month, they can't supply the world's demand right now.
Am i right to say that every time you start the clock you have to open is to give the start impulse and start the "engine"?
could you add a elevator music explanation how the tuned circuit with its transformer works?
Bad cap, bad cap, what you gonna do when Marc comes for you ?
I do have one question; how do you set the time?
It's a whole adjustment ceremony! Open the rack and slide it out. Remove the many screws and remove the metal plate that's over the wheels. Rotate the minute and the hours wheels by hand, there is a little lever to free them up while adjusting. The seconds has a clutch spring lever clutch, so you unclutch it and hold it. Start the clock by spinning the motor. Release the clutch at the right moment. Then adjust the tick by cranking the handle while looking at WWV on a scope. I need to make a video on that!
Wow! HP magic again! Did they really have to open up the cover to start it up every time? Understood that the sync would be lost if a dude pressed on these start up buttons. But HP should have found a way to make it a protected front panel feature, or worst case, rear panel accessible.
Anyone else watch the latest Nova about quantum physics. There was a brief glance of the 115BR in the section about atomic clocks. Close up of the second fraction bar.
So often it’s a capacitor! I get grumpy at them - fair enough for a 60’s era germanium transistor to die - but how hard can it be to make a sheet of conductive material and a sheet of insulator?!! Very hard apparently!!
These were MIL-spec tantalums. Much better than electrolytics. But in our experience they fail, mostly due to corrosion, with a leakage fault rather than loosing any capacitance. So they create hard to catch faults. But we catch them eventually.
It is mostly about keeping atmosphere out. The seals are crucial. I shiver when I see people bend the leads without pliers to keep the force off the seals, or when they push them in TTH boards with another size footprint.
1kHz signals have to be single most annoying tone in existence. In the Apollo series, it was somewhat bearable, but the sound of the stationary, synchronous motor is driving me nuts.
Once it is sealed in the heavy outer box it’s pretty quiet.
I had it running in my maker space 24/7 for a few months.
It's been running continuously since I have it repaired and coupled to the 106B. No 1000 Hz noise once it's running. The bad bearing though is annoying. We need to replace it.
What are the odds that you have the actual clock used to synchronize the world? Like how many of these things could they have made?
Now build a GPS disciplined PPS generator on a Raspberry Pi and adjust it to that to get it time synchronized to GPS as a modern clock reference.
With that much precision, would they have noted time dilating while traveling in the air? 🤔
Others posed the question and the answers given amounted to "no".
I'm sticking to my GPS Clock with WWVL Backup.
WWVB maybe? WWVL stopped transmitting in 1972.
TANTALUM ... oh nooooo! :D
I think the 2nd regenerative stage needs a 12AU6 tube instead of a germanium transistor. 😀
I’m with everyone else, do you have to remove the cover every time you want to start it???
A question. During your repair us used that manual start buttons and had to give the motor a small spin.
Is that not automated? If not, does it mean the operator of that device has to open it, turn it on and then close it again?
Or how is that meant to work in day to day operations?
Yes and yes. It's not automated - on purpose. You don't want the clock to restart on its own and you not noticing the time is not traceable to an atomic standard anymore. You need to open it all up to readjust it and restart it. It's really meant to be operated with battery back up, as well as the crystal or atomic source, and never be readjusted save for the microsecond adjustment. None of that daylight saving stuff...
@@CuriousMarc Awesome. Thanks.
And thank you for your great work in general. In particular in the analog field of electronics I don't know that much, but your explanations are at a level I can get the concepts.
Do you know something,with a really smart,humble guy like ken,as part of this group of talented engineers,i would bet that he would be able to write a cool piece of machine learning code to go through the whole video post production, and refocus the potato frames,whilst marc sips his coffee.No more potato frames uploaded to TH-cam.
This is doable right?
I hope ken isnt going to groan at this suggestion!!
You might even put out a video as to how he did it.
I know,I'm really pushing it!
Maybe people will miss the potato.😂😂
As an alternative, I may want to learn how to operate my new Sony FX-30 4k camera properly...
@@CuriousMarc
Yeah..Yeah,that would be easier,,although not as interesting!!
So do you always have to kick-start the resolver, or does it now start on its own?
Agree with the “but how do you start it with the covers on?” Question. Super fun debug sessions!
@@tyrueiImodel T external crank start?
Many synchronous motors have a 50/50 chance of turning either direction. Henry Warren invented and patented a self starting synchronous motor in the 1918 which always turned the same direction. Warren’s Telechron company was bought by US’s General Electric. Most other manufacturers’ clocks had a manual start knob to get the motor turning the correct direction. I’ve seen a few old clocks with a ratchet mechanism which would jam if turning backwards and bump the motor to turning the desired direction. Apparently HP didn’t want to pay royalties for a self starting synchronous motor along Warren’s design.
Warren had the idea of using the AC power grid to transmit an accurate time reference frequency to home clocks which would count cycles via synchronous motors gear divided to display time on a clock dial. In the 1920s, there weren’t national grids and line frequency was approximate. Warren invented a companion power station master clock to monitor grid frequency. I have one in my clock collection. The master clock consisted of an electric clock section using a Telechron rotor motor. There was also a high accuracy pendulum clock. The two clocks were connected to a differential which was attached to a pointer to indicate need to speed up or slow down the power station alternator. The Telechron motor also kept the mainspring of the pendulum clock wound. The pendulum clock was regulated by listening for a radio time signal and using tweezers to add or remove tiny weights on tray on the pendulum rod.
@@wtmayhew cool, NTP before NTP. I guess, why not sync clock from it now, at least in the US and other 60Hz countries. For anyone on 50Hz, well, sorry?
@@c1ph3rpunk The rotor assembly of a Telechron motor is enclosed in a brass or aluminum housing pressed by hand into the laminated pole assembly with the field coil. The rotor assembly encloses a gear train to drive a 1 RPM (usually) output shaft. The whole assembly with gears was called “the rotor.” Rotors were available for at least 25, 50 and 60 Hz and possibly others. The same field coil assembly could be used for all the rotor types.
The master clock I have originally had a 25 Hz rotor but the person who owned the clock before me swapped it for a 60 Hz rotor
I'm a bit confused about why one has to stick fingers into a rack mounted component to press buttons and spin wheels.
Did the 1954 sync process take relativity into account? I'm not sure you have to.
That came also to my mind as he said, they used it to synchronise atomic clocks around the world, and transportated them by plane 🤔
They did not at first but then realised that it actually matters
Didn't matter for that first round, which just sync'ed to 1 µs. It starts to matter when you want to do it to a few ns.
What type of synthesizer do you have hanging in the background 18:20, KORG, Moog?
That's a Roland Juno 6.
With the awesome Juno 66 upgrade!
I'm starting to follow some of this. Soon I'll have to give up my software hacker's license.
How did they get around general relativity while flying a clock at high altitude around the world?
I don't know, but where others posed similar questions, people answered to the effect that it wasn't that accurate a comparison where relativity would make a difference.
Hang on a sec, wouldn't flying the atomic clock around the world throw off its time? I remember a story about how they tested Einstein's theory of special relativity by flying around an atomic clock and comparing it to one on the ground, and the two showed a difference, so if they flew one of these around to synchronise atomic clocks around the world, wouldn't they still be different by whatever the difference was from flying it around the world? Or were they somehow able to compensate for the flights?
Not for this first 1 µs sync. You are 2 orders of magnitude away from having to worry about relativistic effects.
23:00 my red is not strong enough... welp.
It's a tantalum, what are you expecting?
It's a MIL-spec tantalum. I'm expecting it to last a long time.
@@CuriousMarc That explains it. I was really surprised not all of the tantalums were failing short after all those years... (yes, I've been burned by tantalums... now MLCC seem to become a fitting successor to those.
ATOMIC! Wait on! The click bait said "Steam Punk", so where's the boiler?! 🤣
What if you just put in a GPSDO? 😊
RF engineering is black magic. Change my mind
No worry. We consulted the RF sorcerer and he confirms it's magic.
29:58 slurp 😂
Strange design choice to have to open the it to be able to start it.
So no one (accidental) can affect it once running. It then keeps on running for years. Re-syncing is not an hour job.
✌
kool :)
what kind of genius thought it was a good idea to remove the top of the unit to start the dividers? and remove another shielding to start the motor?
th-cam.com/video/0kdnvCuWXV4/w-d-xo.htmlsi=xnMYzuFPLv7d90qS&t=1496 This is correct. You have series (current) resonance at input and parallel (voltage) resonance at output. So, at 1 kHz input voltage drops to near zero since of filter input impedance drops in resonance. I'm glad you didn't go this path too far away.
Are you sure? Certainly looks like a parallel resonance tank circuit to me.
@@CuriousMarc Take look at the schematics. There is a tuned transformer that has transistor bias winding between transistor collector and -12 V rail and two other windings, one feeds frequency tripler and second that is actually divider output. Primary side path is not obvious, since bias winding, frequency tripler winding and capacitor forms parallel resonance tank that you mentioned, but frequency tripler winding, the capacitor and transistor collector forms series resonance path. My guess is that tripler winding + cap is about 1 kHz, but tripler winding + bias winding + cap frequency is lower (because there two coils in series). Proof is you can actually see two resonances at input: series at 1 kHz and parallel at about 700 Hz.
Hmm, that explanation does not make any sense. The winding 1-2 is not a bias winding at all. This is simply a transistor shooting into a proper LC parallel tank circuit load, with the collector injection tied to 2 instead of 3 so the low output impedance of the transistor doesn't kill the Q of the tank circuit. That's how all tripler stages are made too, they are also all parallel LC resonant. But in the triplers, this is the only source of resonance. However, on the mixer, the secondary winding (4-7) **is** wired in a series notch configuration! I suppose that's to kill the 19 kHz mixing product, has nothing to do with the 1 kHz resonance. Also there is C319 across pin 2 and 7 connecting the input and output windings, which I would not expect there. I have not yet figured out why it is added. I wonder if you'd get some unplanned resonances when you excite at 3 as I did with the experiment. I should have tried to excite it at 2, but that would have required me to disconnect the collector of Q302. If I ever re-open the thing, I should try to excite at 2 with Q302 disconnected, and see if I get clean 1 kHz resonances. Still not clear to me why I got a resonance at 700 Hz, but it is not due to the parallel tank circuit between 1-3 and C318, that's for sure. Removing C318 caps there did not change that low resonance at all. Now I think it might have been simply due to my test setup and parasitic capacitance from the generator. This would go away, or at least change drastically, if I excited at 2.
@@CuriousMarc The problem is that transformer is far from ideal. One is small in relation to frequency, windings has to be very tiny cross section. This causes problems like weak coupling between coils, hundreds of Ohms coil resistances and leakage inductance of course. There is also dubious to make a primary winding tap on pin 2, since transistor collector is actually current source so impedance is rather high at this point. Maybe this was emitter tap at early stage of design? Besides, Q of LC tank is already killed by internal transformer resistance.
I did some simulation in QUCS. I assumed that primary winding sections are 13.2 and 330 mH (pins 1-2 and 2-3), there is also 10% inductance leakage and 20 + 100 Ω resistances. Coupling factor is 0.7. Load capacitor is 100 nF. Result didn't surprised me: coil tap 1-2 is mainly DC current path for transistor, most energy is injected while primary winding is in series resonance. There is also weak parallel resonance at lower frequency too. This exercise confirmed me also why we should use Meissner topology instead of Hartley when designing oscillators in kHz range (I know that this is mixer not oscillator).
BTW I guess that C319 advances signal phase or maybe improves weak primary to secondary side coupling. It's also hard to say what is internal transformer parasitic capacitance. Could be 700 Hz resonance is just internal to transformer. I'm wondering why they didn't use RC network or some mechanical resonator instead?
....more headache😅
Tants, the worst! 🎉