Greeting earthlings. If you follow the channel,
you know that we love all things Apollo, and that during our last visit to Steve
Jurvetson’s amazing space collection, we were given the opportunity to take two
holy boxes of Apollo electronics to our lab. These are the boxes that brought you voice,
data and live TV from the moon, and should be early masterpieces of microwave electronics,
the blackest of black arts in analog electronics. So far we have worked on the two main
microwave boxes we got from Steve, the transponder and the amplifier. But the Unified
S-Band system is quite extensive, and a bunch of other boxes are needed to encode and decode
the signals carried by the microwave front end. And thanks to Marcel, we
actually have many of them, which should keep us entertained for a while, as
we figure them all out and make them work again. Most of them are from the Apollo
Command Module and well documented. But one of them is particularly mysterious. As
Marcel told the story from when he bought it, he first spotted the disassembled chassis at a
recycler and recognized it as probably an Apollo ground service equipment box. It had already been
gutted and had no cards in it unfortunately. But a bunch of modules were strewn about on the floor,
and he realized they might belong to the box. He picked them all up, and here we are: a
mystery box with cut wires and a bunch of modules that we don’t know in which slot they
go, and on which we have zero documentation. However we have an inkling of what it is. It
says updata link on it. The command system to remote control the spacecraft and the guidance
computer was called the updata link or UDL, using Phase Shift Keying modulation
or PSK. All the legends do match. So this is probably the box that tested
sending coded commands to the spacecraft, and the command code would have been displayed
in the front. What a save, Marcel, what a save. Ken, you have been working on a mystery box from
Marcel? Marcel, you're hiding! It's your box. [Marcel] From the scrappers, yes! [Ken] I'm trying
to reverse engineer this box, but it's a mess of white wires that go everywhere. I have figured
out that the display is being driven in octal, and I figured out part of the PSK circuitry. But most of what happens, there's a
paper type reader, it's getting messages, somehow it's comparing them, so we're
going to power it on and see what happens. [Marc] Yeah, but what's the
name of the box? [Ken] Right, it's an Up-data Link Confidence Test. [Marc] So,
we figured out that this was the probably the box that gave commands to the command module. So the
updata link lets the ground switch some relays in the command module or send commands directly to
the Apollo Guidance Computer. [Marc] Right. And then we also know that they didn't send commands
in the blind. They send commands and then, right at the transmitter, they received them
and made sure they had sent the right thing. And there's a whole bunch of buttons that seem
to compare what has been sent to what should have been sent, right? And then we also know
that they didn't send bit by bit, but that each bit is split into sub-bits. [Ken] So yeah, they
used, I think, five bits to transmit one bit, so you could detect corruption. [And those
were called the sub bits. This will spoil them. [Marc] And then we think the bits come out from
the paper tape, which is why we brought our trusty ASR-33. And then we know the
indicators work, because you had a few of them lighting up directly. Yes, it looks like we have 13 glorious
edge illuminated number indicators. Fran Blanche made an episode on them, I’ll put
a link somewhere. They work with regular light bulbs, edge illuminating panes of transparent
plastic with numbers engraved on them that diffuse the light. Sort of a low voltage, though
higher power consumption version of a Nixie tube. So, well, let me make a little
punch tape. Or you want to make one? [Ken] Okay what shall I do?
[Marc] You just punch it. [Ken] Oh wait, I guess it should have
been octal. Oh, well... [Team laughs] [Marc] Okay. All right, so we have the tape and then
we do "Break", and that will spit it out. And it's going to be our trial tape. There we go,
this is going to be our our tape. Can you take one module out? Because you were able
to a few of them reverse engineered. It's quite interesting how this thing
is constructed. So, here you go. [Ken] So each of these modules has a
function name on it: Phase Detector, LP Flip-Flop, this I've reverse engineered as a JK
Flip-Flop. So the different modules have different functions. They're hybrid modules with transistors
and resistors and stuff in an epoxy block. [Marc] Can you remove one? [Ken] Sure [Marc] Because they're the before the IC, right?
You made modules with little transistors in there. [Ken] And they have thirteen pins.
[Marc] Hold on, it's not focused yet. [Ken] The pin spacing is 0.2 inches, you can
plug these into a breadboard, it works just fine. So I've done that for some testing. [Marc]
So you got a flip-flop to work, you got the gate to work. [Ken] I got an emitter follower
to work. There's some block here that I found some of the inputs, but it does weird stuff and
I can't figure out what's going on. [Marc] Right. And a phase modulator, and that seems to be quite
okay, because the box generated phase modulation on the signal. Very low bit rate 200 bits per
second, because it's one kilobit per second but divided in five sub-bits, so it's 200
bit per second on the way up. And the box is missing all kind of stuff, right? It has a
board missing here.[Ken] So this board apparently had some inductors, and a resistor, and a mystery
module that had been cut out of it. [Marc] Right. And also the power supplies were all missing.
There were plenty of cables cut. We kind of figured it out and rewired it was what we think
is the right power supply. That was the 400 - it was powered by 400 [Hertz] three-phase
originally, but no more. So, conveniently it told us what the voltage is where,
because there are test points in the front. [Ken] So this was a big help for figuring out
what the voltages are. [Marc] 28, 6, ground, -6. There's a lot of test points. [Ken] So when
the module is plugged into the box you can't access the the circuitry. But they've wired
up these test points to important signals, which makes it very easy to test this.
If you know what the test points are, you can just stick a probe into the hole, and
find out all the important signals on the board. [Marc] And so at this point, we think it
generates somehow the PSK code and can read the PSK code. Okay. Well, we power it up,
and see if it lights up. I'm going to do +6... [Marcel] +6 dropped to 3V. [Marc] Yeah that's what I thought. All right. Which one the +6V? [Ken] Yeah. OK, we've got digits! Oh,
we got zeros! [Marc] Do you have 6V? [Marcel] 4.5V. [Marc] Oh I have 2A.
I'm at the max of this power supply. Okay, so I need a more powerful power
supply, this is just not good enough. OK, we have the better power supply.
Okay, quite a few amps. This one and this one. This one is very deep
little current. Okay. Did it come up? [Ken] Let me grab my camera. [Marc] Oh it did come up!
[Ken] But we're in 400. [Marcel] Let's clear it. [Marc] Tape stop. So it has a problem
with the tape. [Marcel] Let's push all the buttons. [Marc] Oh, sub-bit
agree! [Ken] You got green. [Marc] Okay, can't read the tape. So that's
probably your first thing to figure out how the tape works. [Eric] Can you
reset it? [Marc] Yeah, I'm going to try. Are you good [Ken]? [Marcel] Do the +6V.
[Marc] Okay, +6V off. Woohoo! [Marcel] It's something else! [Marc] Wow, it doesn't have
the tape error anymore. Maybe it needs the -6V first. [Ken] It looks like we're missing two
digits. [Eric] Are they missing or burned out? [Marc] wWe don't know. [Marcel]
Each digit is a separate light bulb. Okay. So after some disassembly
from the tape drive, right, there was the wire broken. [Marcel]
Suspiciously empty pin. [Marc] Yeah, it's very suspicious. [Marcel] It goes to the yellow wire,
and [there] happens to be a yellow wire down here, which is our main solenoid for
the driver mechanism [Marc] Okay, so that would explain - actually, for
sure it cannot work without that wire. So that's the wire that drives
the whole thing, it's broken. So while Ken was doing the difficult electronic
reverse engineering, I took care of tape reader, which was frozen and bent - it must have taken
a blow. But fortunately I was able to free and readjust it so the transport started to work,
at least when the coil was driven directly. Look, smooth! But even with the repaired wire and the freed
up mechanism, our tape still didn’t budge. It looked like the electro-magnet was energized
all the time. We soon figured out why though. The relay here is driven by by high current pulses
that make it make it click. And those pulses I traced it to this mercury relay here. This is
a mercury wetted relay that can handle high current inductive loads. And if you come from
this angle, you can see the mercury relay has an arrow labeled up. [Marc] Yeah, it only works
with gravity in the right direction. [Ken] The up arrow is currently pointing sideways, so
the mercury is probably shorting it out, and that's why the solenoid was stuck on. So we
just need to put this back in its regular position and hopefully... [Marc] And Eric! In the frame
I have your tape reader. Because we figured out this is a Tally tape reader, and Eric has a Tally
tape reader. Between all of us, Marcel, you, me and Steve Jurvetson, I think we have
everything. [Eric] Yeah. So the mechanism of this tape reader is the same on the
front. [Marc] See that, looks the same to me. [Eric] Basically the same. However, if
we flip it around, the back is very different. And thanks to Eric having recognized the type of tape reader head this was, it allowed me to
get the adjustment manual for this mechanism, which was critical. The drive mechanism was
very different, as his reader is much faster. [Eric] This mechanism uses a 1/12 horsepower
motor, it's very fast. [Marc] But you know why this one is slow speed, right? Because the
link is 200 bits per second, right. [Eric] So you don't need it to be fast. [Marc] Oh
actually you don't want it to be fast, that would be a problem, you would need memory.
Can we make it work you think? [Eric] We can, yes. [Marc] Yes! Oh, that's a super fast one! Here is our experimental setup: we've
put the relay that way up. We elected to do that rather than flip the box,
because we want to be able to probe. And we have the reader with the tape. And we have
a few control points, cleverly chosen by Ken. [Ken] So this should be the three bit octal value
that we read off the tape, and this is a clock, maybe. [Marc] Yeah, we want to find the
clock. I want to find one which I would think is one kilohertz right or two kilohertz,
something that's related to the modulation rate. [Ken] We're getting nothing
on the - oh, there we go! Oh, we've got nice numbers!
All the digits are working. [Marc] This is pretty! [Ken] Let me probe this. Can you tell
me if you see anything interesting? [Marc] This is zero. One. Zero. Ah, yes! [Ken] So what's our clock frequency? [Marc]
100 kilohertz! [Ken] 100 kilohertz. [Marc] That's a lot higher than I expected it to
be. And then we have one digit missing, but we saw them all work. [Ken] Yeah, so it
could be a burned out bulb. [Marc] Yeah. So I want to see one kilohertz. But you must have
the sub bits... [Ken] I don't know anywhere else obvious that would have a different clock. [Marc]
You want to try to read some...? [Ken] Should we flip some switches, see what happens? Okay,
clear did nothing. Wait! [Marc] Something happened. [Ken] The digit, it looks
like our broken digit is now working. Okay, well, that is less than I was expecting. [Marc] I still think
you need to put the one kilohertz somewhere. [Ken] Well why don't we at least try advancing
the paper tape and see if see if it's reading? [Marc] That's zero zero zero, one, two,
three... [Ken] Wait, the blue one moved as well! Probably adjust a bit. Oh no, it did. Four, five, six, seven. Yeah, I have a bad contact on the blue bit. Okay,
so it's reading it, but we can't make it advance. We've made progress. It didn't do anything last
time but we thought we were probably missing a clock somewhere. And Ken, you think you've found,
probably, or [have] a candidate? [Ken] So I've been doing some tracing out of the circuitry,
and I believe this is the board that generates the phase shift keyed signal, which is a 1 kHz
signal that changes phase for a zero or one bit. And tracing out the schematic, I found
that it's driven by two flip flops, which are dividing down an input signal. So
I assume it's going to be: 4 kHz comes in, gets divided to two kilohertz, and gets divided
to one kilohertz to generate the PSK signal. So the question is: where where does that
clock come from? And it comes from this board. And you'll notice there's some components
missing. And there was a component here which was labeled MP1 [Marc] I don't
know, MP1? [Ken] Missing Package 1. So it takes power and ground, and then
has one signal out which is probably the clock. So it's probably some sort of a
quartz oscillator. [Marc] Right. [Ken] So we're going to simulate that oscillator with
the HP function generator here. [Marc] Yes. Ah, no wonder we could not find our clock -
that came from a part of the machine that had been removed. And thanks to Ken’s incredible
reverse engineering skills - remember none of these pseudo integrated circuits are documented
- we now know it needs 4 kHz and where to put it. That's the missing one where do your LP
components and your MP components would be, you think? [Ken] Yeah, those components were
probably too big, and came up off the circuit board, so they couldn't put a card in this slot.
You can see the same effect on this, the green monolith on the end. So we're not missing anything
here, this is the transformer for the op-amp power supply [Marc] Okay [Ken] but we're missing some
power supplies and some amplifier that were mounted on the back. [Marc] Right, but we have
taken partially care of that. Okay, let's try it! And before we go try that, I
should mention that Master Ken also figured out how the tape is
encoded and how it loads the digits. The paper tape reader here, I traced
the wires through to this board, I reverse engineered the modules and circuitry
in this board, and discovered that the format is three bits of data, an octal digit, and the
same three bits inverted for error checking, and then two command bits. So the paper tape
reader reads a row from the paper tape into this module. This module has error checking
to make sure that the three bits and the three inverted bits match. It loads it into a
shift register, and then sends it to one of the display digits. So each time you read a row from
the paper tape, it will read in another digit. And we believe it will shift them into the into
the display so we're hoping to see that work. [Marc] A command message was a 10 digit octal
message, and I think we have 10 digits right here. With the ship address - vehicle address - then
the type of command, then the rest of the data. So it just matches what the command message is.
And that, it says here "message number". [Ken] My theory is that the message number is going to
have to be on the paper tape as well. [Marc] Oh, you think the message number has to be
on the tape? [Ken] Yeah. All these digits are equivalent as far as the circuitry. [Marc]
Okay, so I need to change my tape format. Okay, let's see if we can redo it on video. We
just got live out of the thing, and this is just with our four kilohertz in. We got something
to read. How did you do that? You did clear? Yeah, it went to zero. Program start?
[Ken] Yeah. [Marc] And tape advance. Aha! I think we just read our tape! So we were about
right right, that we are missing the 4 kHz. Cool! [Ken] It's not happy with what's
on the tape! [Marc] I don't know why. But it got something in. [Ken] We have all the
digits active, that's nice! [Marc] Okay, well, now we can go to the next step and figure
out what it's reading and what he wants. Aha! Making progress. We think
we have seen a phase shift. Go for it Ken. [Ken] Okay, let
me load the paper tape back in. Clear, tape start. [Marc] Okay, I got a few.
So it happens repeatedly, but here's one. If you look at the waveform on the green, which is
the one that comes out, you have one that's half a shift smaller than this one. And you can
see a controlling bit here just eats half of a period ,and then the whole thing is has been
shifted. So it's eating data and creating it, but we're still missing some pieces of the
puzzle. But it's getting there, we have data. We were having actually read errors on the tape,
and it's because the tape reader was not properly adjusted. So I made a test tape of alternate
patterns, which you need to adjust the contacts. This is * and U in ASCII by the
way, equivalent of RYRY in Baudot. And I think it corresponds to 5 and 2 on the
device. And by the way, Ken, you got us all of our zeros back, right? [Ken] Yep I replaced a few
bad light bulbs. [Marc] Tada! It all works! And so, if i did it right, the hope is that it
goes 5 2 5 2 5 2. Okay, tape advance, clear. I have my 5 2 5 2 5 2. So this tape
is read properly. [Ken] No errors. [Marc] Try again. Yeah. 1 2 3 4 5 6 7 and the 7 7 7 7 7 is were
the lead-in of the tape. So this read was good. So this is my newest tape, and you can see i have
a whole bunch of commands. One that's all 0's, one that's all 1's, one that's all 2's, etc... So
you have three bits, that's zeros, the anti-bits for checking, and the start of the sequence.
Then 13 numbers. This is 1, and same thing: you have the 1, you have the anti 1,
and then you have the start of sequence. This is 2, same sequence, this is 3,
this is 4, 5, 6, 7. And it worked fine for about two minutes. And then we ran in trouble.
Let me demonstrate. I think, tape advance, clear. So that's... Oh no, it works! That's my
bunch of 0's, command made out of zeros. Command made of 1's. It almost works, except the
last digit. So just wiggling the modules around... So we had a problem: this is a whole shift
register, and we kind of lost it in the middle of the shift register. It works again now. [Ken]
Then we swapped the modules to move the problem to the end, so we could track where the problem
was. [Marc] Okay, let's see if I have 2. Okay, 3, 4, 5, 6 , 7. It works, except for
the first digit. But that might be related to the way I'm doing the
command. We saw that before, right? 6 and 7, yeah, no more tape errors. And that,
there's a tape error, because it's reading the end, right, which has nothing on
it, and it's all 7's. So, progress! This is the achievement of the day. Applause! So we are far from done with this box, which
still has bits missing that we need to re-invent. But we started from a completely
unknown box and got to reading in tape commands thanks to the impressive
reverse engineering work of Master Ken. And kudos to Marcel for having spotted
the box and prevented its destruction. In the meantime, Mike has acquired some new
documentation that shows how our updata box was hooked in the test setup, and Ken has built an
automatic sniffer cards to map out the backplane So we’ll come back to it in future episodes, and hopefully send a successful command to
our spacecraft in a not too distant future.
