Rockwell International produced the 251-9735-010 6 GHz amplifier many years ago. It was a popular amateur radio choice surplus around the turn of the millennium and might still be a decent choice if found at a good price. The one I have looks to be new old stock passed through amateur hands along the way with a paper label handwritten "+37 dBm 5W | 5.76 GHz AD" alongside the factory test tag with matching serial number. That inspection (not really warranty) sticker wasn't about to last very long…
 |
| New-old-stock, as-found Rockwell 251-9735-010 6 GHz Amplifier |
The North Texas Microwave Society's meeting presentations and tech library are quite useful and the latter has Surplus amplifiers for the vhf and up bands - Dave Robinson (pdf) [this device featured on page 16] that, with some sleuthing, lead to Dave's current site for Using the Rockwell 6GHz amplifier. Having that much to go on gave me confidence to buy one when it showed up on ebay at a very reasonable price. Still, I hadn't read the "4.6 lbs" spec and wasn't exactly prepared for a giant, aluminum-covered brass bar to show up trying to tear its way out of the low-cost, thoroughly padded paperboard mailer.
 |
| Arrival in original (remains of) anti-static bag |
 |
| Tested at 5.76 GHz by a helpful ham sometime in the past. Or at least labled for its intended application. |
The unit is marked:
- POWER AMPLIFIER
- 251-9735-010
- 13499 REV H (stamped)
- UM-51D-1
- 622-5834-001
- [][][] (three empty rectangles)
- POWER: 5 WATT
- Frequency:
- 5.925-6.425 GHZ
- SN3754 REV T (stamped, partially double-printed)
- 50585 (stamped)
 |
| Factory production tag has remained with unit 3754 until now |
 |
| Matching inspection tag viewed from the output end of the module |
 |
| Mounting side with retained bolts showing primary brass structure |
 |
| The "RF POWER SAMPLE" port adjacent to the output is covered by an open-circuit SMA |
 |
| The underside of the outer aluminum cover is unadorned except for a part number |
Now that the warranty is well-voided this long after production, what else is under there?
 |
Overview of power, bias, and wiring concealed by the aluminum outer cover |
Now let's dig into the components supporting the contents of the brass block
 |
| On the input power is a single 0.2 ohm power resistor |
 |
| Overview of the bias and alarm board |
 |
| Bonus bypass capacitor for feed-through alongside Motorola 2N6040 PNP |
Since there's not an established reverse engineering of this board, I'm not sure what role the transistor is playing, here. The 60 V, 8 A PNP Darlington pair can certainly switch the rated input for the module.
 |
| RCA CA723T regulator with heatsink |
 |
| Lineup of bias potentiometers after regulator with precision resistors, bypass caps, and wired connections to feed-throughs |
 |
| MC3403L quad op-amp alongside "-4V ADJUST" externally-facing potentiometer |
The classic op-amp (that bragged of performance beyond the 741 series since '79) is part of the signal conditioning that handles the negative-output detector diode used at the final output. Adapting to a positive-output signal requires re-adjusting this potentiometer after the diode replacement on the inside.
Now for getting inside:
 |
| The bonus bypass capacitor ring terminal requires a longer screw in the assembly |
 |
| Releasing the pressure on the brass cover lets it spring back amidst the recently disturbed dust accumulating around the circuits. |
That springiness means there's only so much immediate excitement.
 |
| Brass cover removal revealing beryllium-copper spring shield sheet |
 |
| Uncovered RF compartments with input at left and output/monitor at right |
And there's the first time these transistors have seen the light (not quite of day) in almost 42 years! The assembly is built with three sections of the enclosure encompassing the stages: input pre-amplification, medium power amplification, and paralleled finals with output support. Now for a closer look.
 |
| Input circulator with load and first two gain stages (four ceramic boards) |
The input from the SMA connector starts at a circulator configured as an isolator with a small termination before two cascaded small transistors with three boards for matching, coupling, and bias. Almost every RF node is meticulously tuned with ceramic chips affixed with plenty of adhesive.
 |
| Intermediate stage with circulator and yet larger transistors (four ceramic boards) |
The intermediate stages start with another isolator and two different, larger transistor packages. The latter is even big enough for being marked with "Avantek IM 5964-3B" as a truly power device for the era. Here, the center coupling/bias board required no tuning chips, but all the other boards and the top of the large transistor sport theirs.
 |
| Finals with racetrack couplers and PA transistors (two ceramic boards) |
The finals are a paralleled stage without additional input isolation. Chip-tuned racetrack couplers are used for input division and output combining with similar chip loads for the terminated ports. The PAs are each "Avantek IM 5964-3A" devices. Gate bias is fed through carbon comp resistors and isolating inductors. Drain supply is fed through a multi-turn inductor on ferrite bead, past a snubber, and through similar isolating inductors. Ceramic chip tuning is used again at almost every RF node across the board.
 |
| Output detection, sampler, and isolation circulator (four ceramic boards) |
Finishing up before the output connector is one board that supports the negative detector diode with attenuator, an electrostatic sampling line, and the final board with the bigger circulator and full-sized load resistor. Ceramic chip tuning is still necessary and ferrite blocks are used around the detector diode and along the sampling line. SMA connectors retained with roll pins for both the primary and sampled outputs. The semicircular resistive print for the reflected power dump is definitely a neat part.
All along the way the different boards have been hand-soldered together with gold (plated?) stripline segments to jump between copper traces. Just one small problem that you might have spotted it if you looked closely above.
 |
| Loose stripline piece running around behind upper PA's bias network |
 |
| Removed stripline fragment placed on the backside of the BeCu spring shield |
Maybe it's a good thing this one never made it into service? None of the stripline junctions look damaged, so this might have been an assembler leaving debris while mounting/adjusting a junction. The fragment doesn't look big enough to short-circuit anything critical but could have been a pain (failure) if it got jostled to the wrong place.
After all that, there's still a long way to go until I can implement this thing. Still, understanding what's going on inside the aluminum/brass box helps know what to expect and explains the modification steps on Dave's page about the amplifier. Plus I don't need to use my 6 GHz connectorized isolators on this amplifier because they're already built-in.
And with the lighting I can manage now, look forward to more pictures of the Toshiba 3.4 GHz amplifier seen on Dave's site.
Comments
Post a Comment