I found a few interesting boards being sold as scrap a while back, from some kind of marine video system.
I'll get into the circuit details soon, but first, a little background:
The CAGE code on them is 05869 (Raytheon, currently) and looking up the part numbers on the boards resulted in descriptions of "PREAMPLIFIER,DEFLECTION" and "VIDEO AMPLIFIER", and 1989 record creation dates.
This doesn't tell us much alone, but an alternate listed supplier is CAGE code Z4779, for "ASC PTY LTD". ASC PTY is apparently an Australian ship manufacturer, and looking at their
Wikipedia page turned up this:
ASC rose to prominence in 1987 when it was contracted by the Australian Government to design and manufacture a fleet of six Collins-class submarines for the Royal Australian Navy (RAN)[28] in what was the largest defence contract ever signed in Australia. Although the submarine project was marred with difficulties throughout the 1990s, upon completion the Collins class were hailed as the most advanced diesel-electric submarines in the world. The difficulties continued, however, and the very expensive Collins-class submarines have been plagued with troubles and controversy ever since.[29][30][31]
Looking into these
Collins-class submarines some more, as the date range matches these boards...
The combat data system was procured separately to the submarine design; 14 companies were identified as capable of providing what the RAN wanted, from which eight were approached in January 1983 with a separate request for tender.[12][13] Five responded: a consortium led by Rockwell International of the United States...
...Rockwell and Signaal for the combat system.
In 1996, Rockwell sold its military and aerospace division, including responsibility for the Collins combat system, to Boeing.[117] Boeing attempted to produce a workable combat system, but believed that this could only be done if the changes in technology were accounted for in a contract alteration, which the RAN and the Australian Government initially refused to do.[117] Boeing then requested assistance from Raytheon, and after further negotiations with the Government resulted in a reduction of the system capabilities, the companies were able to stabilise the system and deliver Release 2.0 at the end of 1999.[118] Boeing sold its naval systems division to Raytheon in May 2000, making the latter company solely responsible for completion of the combat system.
So, there's a decent chance that these boards came from the combat data system that was managed by Rockwell -> Boeing -> Raytheon. The boards would've been from the early prototyping phase most likely, judging by the 1989 NSN date and the 1986-1987 date codes on the components themselves.
Let's look at the details of the
Video Amplifier Boards:
These two copies are identical, except for the exact parts used for some of the resistors & diodes.
The most visible feature is the hybrid module, which is a Comlinear Corp. CL103 amplifier. This is presumably very similar to the
CLC103, which has a datasheet available. It's a high-current / high-slew-rate amplifier module, with a current-feedback topology, formed from an op-amp and a couple discrete transistors as output drivers.
There's also no shortage of colorful resistors and metal-can transistors, along with some nice relays.
With the very-non-video-amplifier-looking parts, like a bunch of digital logic in the corner, and the relative ease of tracing connections on this board, there was no way I
couldn't figure out the exact circuitry. After a lot of time following traces on overlaid photos of the front and back sides of the board, I ended up with the attached schematic in PDF form. We'll look through the sections one by one - it turns out there's a lot more going on here than just a video amplifier.
Video amplifier
Let's get the obvious part out of the way first. Two video signals enter on the backplane connector. The board selects one of these using a DPDT relay, controlled by a digital input.
The selected video input then goes through two stages of discrete-transistor buffers (Q13 & Q11). The PNP-then-NPN arrangement means that the Vbe voltage drops (roughly) cancel each other, and create relatively little DC offset between input and output. The power supply connections, as I've shown them here, don't entirely make sense - it's difficult to distinguish between different copper planes and I think there's an additional supply voltage in here. But either way, Q12 forms a current sink bias (with R74 as a backup?) for Q11's buffer.
After the buffers, the signal level is set by a trimmer, and a (negative) DC offset is added to it by zener diode D7 & R77/R78. From here, it's amplified by the CL103, and then sent through a 75Ω series resistor, presumably to drive a 75Ω coax cable.
Another reason to suspect that the output drives a long external cable is the protection circuit, which is likely designed to absorb impulses from radar / lighting / ESD / etc. R103 biases C24 at Vcc, so that D12 & D13 will clamp the output a little above Vcc. C24 can absorb short impulses with its capacitance alone, but anything larger is handled by beefy bidirectional TVS diode D11. It's a little strange to me that this only works for positive over-voltages, and does nothing about negative ones...
Speaking of protection, there's also a current-limiting circuit for the CL103's power supply, formed from Q7/Q10/Q6/Q9. The voltage drops across 2x 10Ω series resistors, in-line with positive and negative supplies to the CL103, are compared against the Vbe thresholds of Q6 & Q7. The overall effect is to limit the current draw from each power rail to roughly 60 mA: this protects the amplifier in case of a shorted output cable or too-heavy load (don't want to blow up the expensive board when someone accidentally puts a shorting connector on the output coax).
Mystery analog circuit - video sync pulse generator?
This is an entirely separate circuit which, as far as I can tell, seems to add some kind of synchronization pulses to a video signal. There's a digital input on the left, which gets AND'ed with a "power good" signal (we'll get to that next), so you can think of the AND gate as just a buffer during normal operation.
An analog input signal enters on the backplane connector (pin B35), and passes through another 2-stage NPN-then-PNP buffer (Q3 & Q5). The output of this buffer goes through a diode (D2), another protection circuit, and to an external pin.
However, this isn't the whole story. The mass of resistors on Q5's collector limits the buffer's output current (without increasing its output impedance during normal operation), allowing the output signal to be "overridden" by a lower-impedance signal source.
This ability to override the output signal level without damaging the buffer is important, as we look at what the digital input does.
When the digital input is low, nothing happens - the buffer functions as normal, and the output voltage is a copy of the input voltage.
When the digital input goes high (rising edge), Q4 is briefly turned on by C11, and pulls the output up to a DC level in a short pulse.
As the digital input stays high, Q4 turns off, and Q8 overrides the output with ~ +2V, generated by a 2.7V zener and a series diode.
When the digital input goes low, Q2 is briefly turned on by C1, and pulls the output down to some negative DC level in a short pulse.
This specific sequence of "positive pulse / +2V level / negative pulse" seems like some sort of synchronization indicator. My best guess is that this circuit section is used to add a synchronization pulse to a video signal, from a digital input driven by a timing generator somewhere else.
Power-good indicator
A single comparator here looks at scaled combinations of +5V, a mystery negative rail, and possibly some other supply: I think this is supposed to drive its output high when all power supplies are present and at the correct levels. This "power good" signal both enables the digital input in the previous "mystery analog" circuit, and also turns on a relay (K2) which switches power to some external board.
Video saturation detector
This circuit takes a DC-coupled input from the CL103 amplifier's output, and also an AC-coupled input from the mystery sync pulse generator(?) circuit described earlier. Both these signals are compared against fixed thresholds by a fast (80 ns) LM319 comparator. Two D-flip-flops are used to latch the "threshold exceeded" state, so even a brief excursion of one signal above its threshold will be recorded. The "saturation, latched" digital outputs are sent off-board. A pair of digital inputs allows these outputs to be reset, or manually set (as a test?) under external control.
My first thought was that this is supposed to indicate when either video signal goes beyond its intended limits: showing that something is wrong or there's too much gain in the system, during calibration of levels. There's probably other possibilities too, but I'm not sure what they might be.
Mystery current sink
Finally, there's a stand-alone current sink that drives
something off-board. The zener diode and emitter resistor set the current to a relatively constant ~2.4 mA, while a trim pot allows picking off a specific negative voltage from its collector. There's no way to know what this does in context without seeing the rest of the system.
Inside the CL103 hybrid moduleI can't see a nice module without wanting to look inside, so I removed the CL103's lid on one board:
The biggest die on the ceramic substrate is an op-amp: specifically, the TL072.
There's a collection of other discrete transistors as well:
Finally, there's a mystery device which I can't figure out:
This somewhat suggests a resistor, but I'd expect to see traces,
like the other silicon or thin-film resistors I've looked at under the microscope. It also could be a diode array maybe, or some kind of power transistor.
Anyways, let me know if you have any insights about the mysteries contained here.
Home
Help
Search
Login
Register
