Apollo Comms Part 30: Double PLL Trouble!

We build a Phase Locked Loop (PLL) FM demodulator for extracting the Apollo downvoice from its subcarrier, using the classic 7046B IC. But in the meantime, as luck would have it, NASA's own S-band receiver's PLL seems to have gone bonkers, and we can't lock to the spacecraft microwave downlink anymore! So we'll get very hands on with PLLs in this episode. Expect a healthy dose of PLL elevator music.
Apollo Comms Playlist: • Apollo Comms Part 1: O...
Older video on Fractional-N PLLs, used for frequency synthesis: • HP 3325A Bonus Materia...
Links to doc:
www.curiousmarc.com/space/apo...
virtualagc.github.io/virtuala...
www.ibiblio.org/apollo/
www.ibiblio.org/apollo/Docume...
Elevator Music Credit: Crinoline Dreams by Kevin MacLeod
Our sponsors
- PCBWay: fast turn PCBs, www.pcbway.com
- ElectroRent: www.electrorent.com
- Keysight: test instruments: www.keysight.com
- Samtec: connectors: www.samtec.com
- R&D Microwaves: www.rdmicrowaves.com/
Support the team on Patreon: / curiousmarc
Buy shirts on Teespring: teespring.com/stores/curiousm...
Learn more on the companion site: www.curiousmarc.com
Contact info: kzbin.infoa...

[@steve_case]

I still love “Apollo 12 switch SCE to AUX” on the generator!

[@CuriousMarc]

Hmm I wonder who could have programmed that file ;-)… Thanks Steve for making the file and supporting the channel with Electro-Rent test equipment. Video with Steve here: kzbin.info/www/bejne/maWxnox8m918bdk

[@ncot_tech]

I'm adding "datashites" to my vocab. It can go next to "crapacitor" and "krapton tape".

[@ManofCulture]

and dieodes too

[@thesushifiend]

May as well add “Killer hurts”

[@ke9tv]

Don't forget the sloppy disks for vintage computers, and the smoke-emitting diodes.

[@Wizardess]

Dark emitting diodes, DEDs. WOMs. (Write Only Memories) {^_-}

[@KeritechElectronics]

@@Wizardess light emitting resistors!

[@dominiquegobeil5831]

Don't be sorry, I'm here for the elevator music.

[@EdwinSteiner]

24:19 "How many colored rings to mark your oscilloscope probe?" Marc: "YES"

[@Miller7747]

I remember working as an RF electronics technician on an RF receiver that had a PLL. The technician that tried to repaired it before me left excess solder all over the loop filter. At very low frequencies, his solder blobs would not have made a difference. But this thing worked above 1 GHZ. Needless to say, the excess solder had some adverse effects on the "loop response" and the result was that the loop could never LOCK. I removed the excess solder, and than clean the board, after that all I needed to do was to align the circuit, and it worked!

[@fredinit]

Align the Circuit - put it in a vice, grab a good dial indicator, tram it to within a tenth of a thou, should be good to go. 😀

[@Damien.D]

This story is endless. It was "restoring all the Apollo comm equipment" for a long time, and now it's "upgrading repairing and maintaining all the Apollo comm equipment" ;)

[@MichaelEhling]

Thanks for the EM (elevator music) education about PLLs. I had a hand-wavy understanding about them before. What an amazing bit of engineering. And that you can do it digitally, too!

[@keresztesbotond740]

Ah yes, user error The bane of every engineers existence

[@CuriousMarc]

Yes, but we swear, we tried to turn it off and power it on again! You have to give us some slack though, we don’t have the operating manual for this, so we sort of have to reverse engineer how that is all supposed to work every step of the way. This was all the more puzzling because the phase error dial was really broken, and we had been able to lock it before - before we made downvoice modulation work that is. So that led us astray for a while. But the good thing is that we learned something new in the end.

[@keresztesbotond740]

@@CuriousMarc Wasn't meant as a stab at you All of us do similar stuff with or without an operating manual, whether we admit it or not :D

[@c1ph3rpunk]

We needed a period correct method to test all of this, so in our next episode we’re going to start our Saturn V build. First up, F1 engine fun!

[@fredinit]

A few years ago a group of engineers pulled the turbo pump off of an F1, put in new seals and spruced it up, bolted it to a test stand, fired it up, and subsequently blew out the windows in a nearby building. Yea, the *turbo pump* was rated at 55,000 BHP. By comparison, a GE AC6000CW locomotive is 6,000 BPH, worlds largest container ship is 110,000 BHP (two turbo pumps), and an Iowa class battleship 212,000 shaft horsepower (four turbo pumps). I'd love to see a full F1 back on a stand and fired up. Hard part - finding a stand large enough and remote enough not to cause problems from noise or a possible RUD event.

[@lepinjalmao]

Dude, you don't realise how much history you are restoring. I'm sad i wasn't alive at the times of the vintage computers.

[@bobvines00]

Don't give up on the "vintage" computers -- you surely can find one or more if you keep your eyes open. Collectors of minicomputers & the newer "home" computers are getting older and some are starting to sell their systems to ensure that their survivors don't throw them away thinking that they aren't worth anything. I know one guy up in New England (i.e., in the NE U.S.) that is getting closer to offering his PDP-8/I for sale -- I just wish that I could afford an -8/I!

[@624Dudley]

Seriously, Marc, the level of fearlessness is downright astonishing, not to mention the comprehension. This is world-class stuff! 👍

[@Mickwah]

There could not be a more fitting YT channel for the “stay tuned” closing text..!!

[@av8bvma513]

"Phasers Locked-On Captain! Ah, correction, Phase Locked-On Captain, oh my Captain!"

[@siberx4]

One of the coolest aspects of Phase Locked Loops that isn't covered here is that by inserting a combination of multipliers/dividers into the feedback path, you can get a PLL to generate an output frequency at an arbitrary ratio to the input frequency while still maintaining a phase relationship with it. This trick is used extensively when synthesizing various clocks at different frequencies in modern systems, as it produces very stable output references from a single high-quality input reference. That reference usually comes from a crystal, and is usually at much lower frequency than the output because it's hard to make good crystals that resonate at hundreds of megahertz or higher.

[@CuriousMarc]

I covered fractional-N PLLs a while ago in this video here: kzbin.info/www/bejne/sF6YYaOJmMSqnZo . These were the pre elevator music days!

[@fritsified5952]

@@CuriousMarcThanks, didnt see that one yet. It does have elevator music!

[@leonmartinville7628]

Toujours passionnant. Merci pour votre chaîne et pour la beauté de ces montages électronique des années 60-70.

[@vincei4252]

National Semiconductor and AMD (well before they were known for x86) made absolutely excellent data sheets and data books. I look at modern data sheets and it's a darn shame.

[@CuriousMarc]

National's data sheets were out of this world. They were stellar courses in practical electronics. No doubt influenced by Bob Pease.

[@vincei4252]

@@CuriousMarc They were, I learned so much just reading them. I still regret leaving all my books behind when I moved from the UK to the US in the 90's. I have a pristine National Semiconductor NS32000 designers kit with all the data sheets, data books and the chips in a proud spot next to my fireplace. I just stumbled across it on eBay and paid whatever the seller was asking. It's a shame that the DNA Ti absorbed when they acquired National Semiconductor didn't extend to their modern data sheets.

[@mistermac56]

@@vincei4252 Burr-Brown also had great datasheets before the TI acquisition. Fortunately, some of the BB D/A and D/A converter ICs that were manufactured before the acquisition, TI fortunately still provides PDFs of the original BB datasheets with addendums. TI still manufactures some of them. Analog Devices has good datasheets.

[@mistermac56]

Totally agree. Datasheets from some manufacturers are so bare bones.

[@BobHannent]

Part of the decline of data sheets seems to correlate with manufacturers also wanting to stop publishing anything publicly. Data sheets become marketing materials and not a design resource.

[@michaelhawthorne8696]

When I was at GEC Coventry UK in the early 80's (Working on QPSK Modulators for the Telephone network, 'System X' ), we used HP and Fluke and Tektronics equipment. I can't remember using any sort of down market equipment apart from 1 Philips scope which I needed to set up the QPSK modulator in connection with a whole wall of high end equipment. Hell we had Tektronics sampling scopes working in the Pico range. I was involved in the 2GHz but mainly 11GHz range and 19GHz was being played with when I left. The difference in Waveguide size from 2GHz to 19GHz was like being knee high to an adult. I can see here, the whole setup seems to be only the best equipment, HP, Keysight, Agilent, Fluke......nice to see. Awesome setup guys....Would love to be there.👍

[@chrisprocter6107]

OMG this explains why (some) 1930's science fiction assumes space ships wouldn't have radio communications to earth, which always seemed like a very clunky device to make "castaway" type plots work but I now know was actually been the state of the art at the time! A very random crossover between my tech youtube and book youtube watching but very informative elevator music, thanks!

[@vincei4252]

12:58 Keysight crying into their cocoa. * I was going to say beer but pretty sure no one at Keysight drinks beer.

[@BrianJorgensenAbides]

Your circuit analysis “threw me for a loop!” If I had been given this problem in school or industry, I would have succumbed to “there has to be an easy answer I just don’t know about” when I saw the bandwidth requirements and no closed-form specifications on a 4th order filter. This project demonstrates: sometimes we tune things by rejecting the common mode of theory, we learn the transfer function by twiddling.

[@orbitingeyes2540]

Ahh the Costas Loop! I learned PLL magic by playing around with the venerable LM565s. I made many decoders and detectors from those! Fun stuff!!! I really miss my local Radio Shack!

[@TheGregEgg]

Going to Radio Shack was my favorite thing to do in 1980s and 1990s.

[@danielpeck6269]

@@TheGregEgg I was fortunate to frequent Radio Shack in the early 2000's. I made a lot of dumb stuff with cheap parts!

[@TheGunnarRoxen]

Love it! The PLL elevator description was super cool. Those pioneering engineers were incredible

[@CatFish107]

I have a PLL module in my modular synthesizer. I picked it up as an affordable means to add extra depth and character to a voice. Mixing in the square wave synced to the main oscillator helps fatten the sound, and adjusting to have ot not immediately lock when the pitch changes can add some unpredictable excitement to the sound. Thanks for the info regarding demodulating fm signals. I will be patching up a use case for that.

[@souta95]

This was quite a fitting video for me as I had to teach myself about PLLs over the last couple weeks to repair one of my ham radios (a Ten-Tec 585 Paragon).

[@mfree80286]

You started describing PLLs and the 4046 and my mind just went off on a RADAR tangent... doppler shift happens small scale as well, and combined with reflected signal strength and a mechanically swept directional receive antenna you could derive bearing, azimuth, and relative speed. Neat.

[@DavePKW]

This is still one of my all time favorite projects you guys have had. Simply outstanding work Marc. Thank you for keeping it going.

[@hauptuhrdotnetblog6700]

I am listening to Wide World of Sports on 2GB Sydney. They just did an ad for the program and it had Marc saying “We have to perform a fourier analysis”. That was insane and cool. Your accent. Sure it was you. I’ll find the recording. I almost fell off my chair. Curious Marc and rugby league in the same spot.

[@gvii]

"Come see my homemade PLL!" That's not something you hear too often. Or at all. Lol.

[@chefchaudard3580]

In my time as a service engineer, PLL were the most difficult circuits to troubleshoot. Specially if it was distributed through several boards. There was no point where you could start from, as every signal was wrong. An off tolerance resistor or capacitor, a resistive short between two tracks, could generate all sorts of issues all over the place.😊

[@CuriousMarc]

You are darn right, PLLs are touchy things. And things that work in loops are notoriously difficult to debug when things go awry. Switch mode power supplies also fall into that category. Breaking the loop and testing the individual blocks is often the best approach to troubleshoot, as we did here on the bench.

[@KeritechElectronics]

Thanks for the nice explanation of PLL - not getting too far into maths either. SCE to AUX, this never gets old! "You cannot trust the indication when you don't have modulation" makes a nice catchphrase. More of it, please.

[@Wizardess]

Been there many. Done that, once. Once was enough to teach me the testing sequence with such loops. I was half worried at the start that the NASA loop required the voice loop to be on frequency giving you a loop within a loop situation. That is obscenely difficult to work with. So I abandoned it back in the day. Anyway you had me mentally screaming, "Check the bloody VCO voltage!" And when you discovered things behaved better when modulation was turned off on the incoming frequency I gently and slowly shook my head. Incidentally that Motorola phase detector derived circuit has a possible problem feature for you to deal with. Right at zero phase error it has a small dead band. That can give very annoying loop testing problems. Now, if you try to implement that FM demodulator with an SDR - note that almost all SDR PLLs I have seen feature an arithmetic underflow and don't work right if they use simple "float" rather than "double". Fixed point "int32" might work. The two pole loop filter falls back to single pole operation, frequency lock without phase lock. {^_-} Boy howdy that takes me back a long way, to the early 70s when I first played with PLLs.

[@benfreeman9717]

I don't mind hearing the elevator music early. I always look forward to that part.

[@av8bvma513]

Haha! This is better than "Sherlock Holmes!"

[@skfalpink123]

As ever - just amazing!

[@islandhopperstuart]

Fabulous work Marc. Didn't understand much, but fabulous work!

[@cbmsysmobile]

I love your channel Wish I understood what you were talking about when it comes to this RF stuff.

[@danielmadden6831]

Three more years and you'll be ready for launch

[@JoelWilhiteKD6W]

Great job Mark and gang on the whole Apollo series! Miss you guys from Maker Fair days and electronic flea markets since we moved to AZ. Hope to see you around.

[@PumaTwoU]

I learn an awful lot around here. Electronics is almost a form of black magic in my mind, but I do learn enough here to understand what you are doing and why.

[@Renville80]

Nice bit of a refresher for me - this was one of the trickier circuits to get my mind around back in tech school!

[@NH3rrm4nn]

..do I still have that ttl cookbook? late 90's edition.. brb..

[@gertebert]

My favourite phase detector is the 4046.

[@lpbkdotnet]

Those mini marshals as monitoring amps are a great idea! I’ve been looking for something like that recently and was thinking I might have to build my own - but just stuffing a BNC on the side of a mini marshal would do the job so much better!

[@ludmilascoles1195]

Oh goodie more Apollo stuff and stuff on pll something to learn

[@guyh3403]

I could watch this for hours!

[@DiverCTH]

Greetings earthlings! Time for more reverse engineering!

[@clytle374]

Good work

[@petint]

I literally just watched some episodes yesterday and now there's a new one. This actually happens to me a lot.

[@pjladd04]

Did you notice the retrofit kit for the Juno 6/60 at 18:52 ?

[@petint]

@@pjladd04 No

[@user-pc9it3kt2o]

You guys have way too much fun.

[@TRBORADIO]

May be this help, It's a new rev of the LM567. This is a tone detector IC and can be used as Wide FM decoder. A new version of this IC is the LMC567 a CMOS or HCMOS version of the old one. From TI datasheet saids no more than 500KHz but if is HCMOS i'm thinking it will work in few MHz. May be it's a good idea supply it with 9V. Nice work with 4046!

[@AmauryJacquot]

that episode comes at the exact time I'm trying to build a circuit to replace a temic U2860B and ST TSH512C ...

[@bob_mosavo]

Thanks 👍

[@SubTroppo]

Before seeing the little Marshall amp I never thought of Marc as being "Rock and Roll" due to the nature of the elevator music.

[@8o86]

there's at least two different ones!

[@user-cb3qr9dt2k]

I remember those tiny Marshal amps from the 90's

[@markgreco1962]

Marc needs at least a half stack.

[@KeritechElectronics]

I'll make it run on subminiature tubes. Nuvistors even, if I had any.

[@analogdesigner]

Wow, that was a case of beer project!

[@MonsterMovieTV]

Mini Marshall stack for the win.

[@ScienceAppliedForGood]

- I was worried for a while. Yep, me too.

[@fredknox2781]

At 16:49 I wonder why they went with the "dead bug" mounting method for the transistors? Seems kind of odd to me for someting of this production status, although I do see "dead bug" in NASA standards. More a Bob Pease prototype method, I think.

[@RogierYou]

Cool breadboard

[@nb_shopnb_shop]

I tried to fix a y-x Plotter 20 years ago. Today I'm sure it used some kind if PLL to drive and hold the axes morors. I never got it working, because i never could wrap my head around how they could sub millimeter accuracy over am A1 paper from two 8 bit DACs. The motor didn't had an encoder on it, just an detector when a full revolution occurred. The circuit somehow used the four poled motor windings itself to detect the current position and hold it.

[@R.Daneel]

Love this channel. Take note, algo.

[@mike95826]

Another IC that is similar is the NE567.

[@W6EL]

Oddly I am dealing with a broken Honeywell MS-2T panel meter at this very moment. What are the chances?! Mine seems to be missing the spiral spring. It’s just… not here? Now I’m looking at your microscope shot over and over again. Great video as always!

[@jamesburgess9101]

love the artwork during elevator music time, where did they come from?

[@joelalain]

interesting, i missed 6 videos for which i didn't get ANY alerts for... there are some wacky things behind the scene....

[@User720-ik6mv]

*deedlededee in the doodledeedoo*

[@eduardocarvalho1547]

28:26 Talk to me Goose!

[@erictourangeau3651]

All that to say Allo 😂

[@CuriousMarc]

Last episode I tried “13 give a stir to your oxygen tank” but somehow my buddy astronaut on the other end did not appreciate it.

[@emulateiam]

I think you'll find that that is a picture not of Mariner 2 but of Mariner 10 :)

[@graemedavidson499]

My iPhone phase detector unlocks every time I go near it. With others phase error is always an issue.

[@PaulLoveless-Cincinnati]

I remember using "PLL Tones" in order to access 2m repeaters. You could transmit on the repeater input but if you did not include the PLL tone, the repeater would not "accept" your input. I recall that the PLL tones were frequencies that were preset in the radio and from which you could select.

[@johnopalko5223]

Those weren't "PLL tones." They were "PL tones." PL stands for "Private Line," which is a trademark of Motorola. More generally, they are called CTCSS, for continuous tone-coded squelch system.

[@PaulLoveless-Cincinnati]

@@johnopalko5223 Oh! Thank you for the correction!

[@bobvines00]

Marc & Team, this is "partially' related to this video because of hte repair you tried to do in the meter. Could you guys show how to set up for and actually successfully solder *very small* (diameter/gauge) wires by hand? Small solder blobs in this case won't unbalance a meter movement! I've got a minicomputer tape drive 10-track head with broken wires inside and don't have a good idea about how to go about repairing the breaks. I can't just go out and buy a new replacement head for a '60s-'70s vintage tape drive!

[@CuriousMarc]

This was actually one of the most difficult and acrobatic repairs I have done. It was so hard I did not film it. Entirely done under the binocular microscope, with a tungsten wire tool that I custom made and a micro hook tip that I custom ground. I had to fish the broken wire with the tools, then add and form a piece of enamel wire, and use a special very long fine solder tip, which fortunately I had ordered exactly for challenging jobs like this. Took many tries. But it can be done. I imagine a magnetic head might be a bit easier with better access. Give it a try, be very patient!

[@ericpaul4575]

16:25 Warranty void if open?

[@CuriousMarc]

Yup we’re done. Motorola won’t send us any new modules from 1965.

[@bborkzilla]

Data shite indeed - mostly "here are the specs. Please have your lawyer review this NDA if you want real information."

[@danielpeck6269]

How does the 1.25Mhz signal compare to modern signals like 4G/5G cellphones? I understand that higher frequency equates to more data transfer at the cost of distance. Was that why the engineers picked that wavelength?

[@Wizardess]

Often called DC.... 4G and 5G are above 2 GHz. Now, Shannon tells you something about a channel's capacity given it's bandwidth and noise. In the total absence of noise you can put infinite bandwidth through a keyhole at 1.25 MHz. But, the world is noisy, very noisy compared to cellphone data rates at 1.25 MHz carrier frequency. (BW probably less than 5% of that given components and junk like antennas that goes around the electronics.) At 2 GHz with frequency hopping over +/- 1 GHz you could cram ONE rather wide bandwidth signal through the link given clever signal design and coding. But allocations are smaller than that. And allocations at 10 GHz, for example, are wider than they are at 2 GHz. So for the same noise levels you can cram faster data through 10 GHz than 2 GHz. And it scales fairly well in that fashion through light through fibers. Anything much deeper than that and the elevator music would be drowned out by the exploding heads, I fear. I spend 5.5 years paying a school to cram as much as they could into my head in the mid 60s. And it doesn't quite lead to exploding head syndrome today. I'll defer to ANYBODY else who wants to describe going deeper. I'm blissfully retired. {^_-}

[@edgeeffect]

I've recently been playing around with Op-Amps and I was thinking that if you wanted a very over-simplified description of a PLL you could say "It's like an Op-Amp but for frequency rather than amplitude"???????

[@angelorasmijn7306]

Juno 66? 🤔

[@CuriousMarc]

It's an incredible plug-in processor update for a Juno 6 of Juno 60 that a viewer designed and sent to me. It adds not only MIDI to the Juno, but a ton of new capabilities (portamento, monophonic with multi-oscillators, etc...).

[@MarcelHuguenin]

Greetings awesome team! Great episode and thank you Marc for the elevator lesson again!

[@nezbrun872]

10:52 Ain't that the truth: "Datasheet, or datashite now, they're not like they used to be"

[@Slugg-O]

You guys have your hands full. PLL technology from that era was fairly sketchy.

[@TheGregEgg]

No intro music?

[@CuriousMarc]

F1 engine rocket music quintet. That counts ;-)

[@b43xoit]

Particularly including the whine of the turbopump.

[@Hans-gb4mv]

Once again the equipment shows it was built to last. Hell, if you found enough parts, you could probably assemble an entire Saturn V stack with an Apollo module on top and still launch it today.

[@CuriousMarc]

Not to mention our duct tape and super glue technology is much better today, so assembly would be much easier ;-)

[@Hans-gb4mv]

@@CuriousMarc 😆

[@clausvind8010]

PLL has actually been in use before WWII, e.g. in analog TV

[@CuriousMarc]

No, the PLL is a post 1950 invention. You must be thinking about a video sync circuit maybe?

[@clausvind8010]

@@CuriousMarc sscs.ieee.org/images/files/aboutus/history/ISSCC50/communications/communications_7.pdf 1932 si the first mention, AFAIK

[@clausvind8010]

@@CuriousMarc fr.wikipedia.org/wiki/Henri_de_Bellescize

[@CuriousMarc]

But that's not the same thing at all! If you look at Bellescize circuit, it's a coupled oscillator. Which is really a resonant amplifier. Not the Phase Detector / Loop Filter / VCO of a PLL, which is a much later concept.

[@clausvind8010]

@@CuriousMarc Did you delete my comment ? Many people credit Bellscize with inventing the PLL, but I don't have access to his original 1932 article at the moment (the diagram in the PDF is Signetics) . Will see if I can find a copy.

[@electronash]

I often wondered if Marc does the voiceover for the "One moment later..." lol I think one important thing to mention about a PLL, is that once you are locked on to the incoming signal, you can then use the PLL to produce multiples of that frequency, but still keep it phase-locked. Digital PLLs / Fractional-N DPLLs are of course still used to this day in things like FPGAs, and your modern CPUs. Very very handy, especially when especially when you require a weird clock frequency to recreate an ancient arcade board. lol

[@irgski]

My guess is that some scientist in Russia has this all figured out and a working setup!

[@jacotton]

First view /?

[@CuriousMarc]

Google says so. I can award you some good PLL circuits with bad data sheets.

[@jacotton]

@@CuriousMarc Just attend VCF west, I'll look for you.

[@usuthu65]

First! I hope.

[@CuriousMarc]

Almost, Google says 2nd, but close enough, we have plenty of bad modern data sheets to award.

[@usuthu65]

I’ll put them next to my copy of the Signetics write only memory data sheet.