Synthesized Sweeping Signal Generators
My test and measurement instrument experiences here at Bill and Dave's have been limited to working mainly with RF/Microwave Signal Generators. With my move to Area 51 I'm still firmly mired in the world of Signal Generators, but the product line I'm working on now caters to CW synthesized sweeping applications. The boxes I was building and testing previously operated at much lower frequencies, I think the top end output was a little over 1Ghz, and were legendary for having extremely low phase noise. That made them very desireable to a specific customer base. This product is useful for signal simulations and for system testing up to 40-50Ghz. For example, testing and tuning systems on fighter aircraft is one job these boxes are adept at.
Okie Carol started me off learning the instrument chassis assembly. She's done an excellent job showing me the ropes around here. Thanks to some interesting mechanical engineering, an operator can slap together a complete chassis in practically no time at all. The whole thing consists of a single large motherboard, two prefabbed dark grey pot metal side rails, an aluminum bucket (for the power supply boards), a few bulkheads, rear panel assembly, and a front frame. To build one all you have to do is grab your parts and place them into a fixture. Then you lock the fixture down and hand load a bunch of Torx-20 screws down both side rails. I leave all the screws completely loose with only two turns just to keep them temporarily in place. Then you flip the fixture upsidedown and drop in the motherboard. This also takes a pile of T-20 screws. Again I hand load all the hardware with only a couple of turns on the screws. It's important to start the hardware by hand so nothing ends up cross-threaded when you torque it all down with an electric driver.
After everything is torqued down and solid, I unlock the fixture and pop the instrument sub-frame out with a big, ugly rubber mallot. Rear panels are already built up for us by a sub-assembly group which saves a ton of time. Depending on the option mix a customer has ordered you simply grab the corresponding rear panel assembly and load it into the frame. The rear panels don't consist of much besides multiple BNC connectors and a mess of cables. Those get hooked up later at Final Assembly. Once that is done a front frame piece is installed and then you can add a set of mounting arms. The arms allow for the box to be loaded onto a rolling cart so we can push the units all over the shop floor and spin them end over end and lock them into any position for ease of access.
Building a complete chassis maybe takes me twenty to thirty minutes from start to finish. In comparison on my old line building a whole chassis would take hours. The older series boxes from the 1970s took a tremendous amount of hardware to put together and everything was in separate pieces. Before you could even start a chassis on that line you would already have had to build a reference casting and a power supply because they made up the bulk of the rear panel. These newer generation instruments are so simple and fast to slam together, which makes my life here at work much easier. In an eight hour shift I can stay super busy only building chassis if I want to and I can probably knock out fifteen of them. I like staying busy. Makes the shift go by in what seems like ten minutes.
Next step in the process is power supply test. The power supply consists of three PC boards. Two of them have to be pressed into the motherboard and then a jumper connection is installed linking the two boards together. At the power supply test station they have a goofy overhead boom that looks like a construction crane. It holds a weird fixture made out of clear acrylic. Inside it has one extender PC board. I have to place this fixture over the chassis I just built and then lower it into the instrument motherboard where it makes a connection. Basically the whole fixture becomes a safety shield when it's in a unit under test. If a power supply board shorts to ground or if a capacitor explodes, anybody operating the setup will be safe from injury. It's kind of a hassle to use though because it's awkward and bulky. A third power supply board is placed in the fixture on it's extender board and then you close the lid. You're ready to test the damn thing. From there the process is software controlled by the test rack. I can either take a little break while it runs or I can jump back to build another chassis. Sometimes when I'm in kick-ass mode I like to jump back and forth all night long from building to testing. Keeps things from getting boring.
Not much can go wrong at power supply test. It either works and passes, or something screws up and you fail it. Troubleshooting a problem is easy. The third power supply PC board has a bunch of error LEDs for the various line voltages so if one comes up red I know there are probably a couple of pins on the motherboard that are making contact and shorting out. It's usually the +5v or +8v supply that goes out when that's the case. I fix it by carefully scanning over the board to locate the culprit pins and straighten them out with needlenose pliers. Occassionally though I do end up with a bad circuitboard which will kill the power, shit happens. I take notes and get a printout of failed data and dump those defective boards into a box and forget about 'em. They're not my problem at that point.
If it passes power supply test the box then goes over to Final Assembly. That's the job around here I like the most. Compared to my old line the final assembly here is easy as hell. The whole box really comes together and it's all about dropping in a dozen circuitboards, adding the front panel display, and dropping the RF deck. Then you cable it all up and shove the unit over to the technicians. It's a breeze. The front panel keyboard and display are built by our sub-assembly group so all we have to do is bolt it to the front frame and we're done. We build the RF decks here on the line, but when I get them at Final someone else has put it all together. So again all I have to do is drop it down the side of the unit and bolt it in place. Quick work.
For some reason I really dig the front panel displays on these boxes. I've always liked display technology, I don't know why. Guess I'm an uber-nerd for that. I have some favorites. Nixie tubes from the 1950s are super cool. In the 1970s the first 7-segment LEDs were a rich ruby red color and I've always liked those alot. These units use a pretty light blue-green VFD screen that looks especially cool in low light. The Japanese manufacture our VFDs for us. Besides a few instrument lines here that use VFDs the only other place I've seen that kind of display technology is in some stereo equipment.
RF Decks are the one place I am going to stay away from. I've tried it a few times and I can positively say that job sucks. It sucks real bad. The RF Deck is a collection of microcircuits that are linked together by rigid metal cables. It's a pain in the ass to build because you have to use a rotating fixture and the instructions didn't take into account the fixture position an assembler would have to place everything to put it together. So I'm sitting there looking at the instructions and everything is upsidedown or backwards from the way it's laid out on the workbench. It's like I suddenly became dyslexic and the shit gives me a splitting headache in no time. It's a mess. Thankfully, Okie Carol is so good at building RF Decks that she's stuck putting them together every night.
Okie Carol started me off learning the instrument chassis assembly. She's done an excellent job showing me the ropes around here. Thanks to some interesting mechanical engineering, an operator can slap together a complete chassis in practically no time at all. The whole thing consists of a single large motherboard, two prefabbed dark grey pot metal side rails, an aluminum bucket (for the power supply boards), a few bulkheads, rear panel assembly, and a front frame. To build one all you have to do is grab your parts and place them into a fixture. Then you lock the fixture down and hand load a bunch of Torx-20 screws down both side rails. I leave all the screws completely loose with only two turns just to keep them temporarily in place. Then you flip the fixture upsidedown and drop in the motherboard. This also takes a pile of T-20 screws. Again I hand load all the hardware with only a couple of turns on the screws. It's important to start the hardware by hand so nothing ends up cross-threaded when you torque it all down with an electric driver.
After everything is torqued down and solid, I unlock the fixture and pop the instrument sub-frame out with a big, ugly rubber mallot. Rear panels are already built up for us by a sub-assembly group which saves a ton of time. Depending on the option mix a customer has ordered you simply grab the corresponding rear panel assembly and load it into the frame. The rear panels don't consist of much besides multiple BNC connectors and a mess of cables. Those get hooked up later at Final Assembly. Once that is done a front frame piece is installed and then you can add a set of mounting arms. The arms allow for the box to be loaded onto a rolling cart so we can push the units all over the shop floor and spin them end over end and lock them into any position for ease of access.
Building a complete chassis maybe takes me twenty to thirty minutes from start to finish. In comparison on my old line building a whole chassis would take hours. The older series boxes from the 1970s took a tremendous amount of hardware to put together and everything was in separate pieces. Before you could even start a chassis on that line you would already have had to build a reference casting and a power supply because they made up the bulk of the rear panel. These newer generation instruments are so simple and fast to slam together, which makes my life here at work much easier. In an eight hour shift I can stay super busy only building chassis if I want to and I can probably knock out fifteen of them. I like staying busy. Makes the shift go by in what seems like ten minutes.
Next step in the process is power supply test. The power supply consists of three PC boards. Two of them have to be pressed into the motherboard and then a jumper connection is installed linking the two boards together. At the power supply test station they have a goofy overhead boom that looks like a construction crane. It holds a weird fixture made out of clear acrylic. Inside it has one extender PC board. I have to place this fixture over the chassis I just built and then lower it into the instrument motherboard where it makes a connection. Basically the whole fixture becomes a safety shield when it's in a unit under test. If a power supply board shorts to ground or if a capacitor explodes, anybody operating the setup will be safe from injury. It's kind of a hassle to use though because it's awkward and bulky. A third power supply board is placed in the fixture on it's extender board and then you close the lid. You're ready to test the damn thing. From there the process is software controlled by the test rack. I can either take a little break while it runs or I can jump back to build another chassis. Sometimes when I'm in kick-ass mode I like to jump back and forth all night long from building to testing. Keeps things from getting boring.
Not much can go wrong at power supply test. It either works and passes, or something screws up and you fail it. Troubleshooting a problem is easy. The third power supply PC board has a bunch of error LEDs for the various line voltages so if one comes up red I know there are probably a couple of pins on the motherboard that are making contact and shorting out. It's usually the +5v or +8v supply that goes out when that's the case. I fix it by carefully scanning over the board to locate the culprit pins and straighten them out with needlenose pliers. Occassionally though I do end up with a bad circuitboard which will kill the power, shit happens. I take notes and get a printout of failed data and dump those defective boards into a box and forget about 'em. They're not my problem at that point.
If it passes power supply test the box then goes over to Final Assembly. That's the job around here I like the most. Compared to my old line the final assembly here is easy as hell. The whole box really comes together and it's all about dropping in a dozen circuitboards, adding the front panel display, and dropping the RF deck. Then you cable it all up and shove the unit over to the technicians. It's a breeze. The front panel keyboard and display are built by our sub-assembly group so all we have to do is bolt it to the front frame and we're done. We build the RF decks here on the line, but when I get them at Final someone else has put it all together. So again all I have to do is drop it down the side of the unit and bolt it in place. Quick work.
For some reason I really dig the front panel displays on these boxes. I've always liked display technology, I don't know why. Guess I'm an uber-nerd for that. I have some favorites. Nixie tubes from the 1950s are super cool. In the 1970s the first 7-segment LEDs were a rich ruby red color and I've always liked those alot. These units use a pretty light blue-green VFD screen that looks especially cool in low light. The Japanese manufacture our VFDs for us. Besides a few instrument lines here that use VFDs the only other place I've seen that kind of display technology is in some stereo equipment.
RF Decks are the one place I am going to stay away from. I've tried it a few times and I can positively say that job sucks. It sucks real bad. The RF Deck is a collection of microcircuits that are linked together by rigid metal cables. It's a pain in the ass to build because you have to use a rotating fixture and the instructions didn't take into account the fixture position an assembler would have to place everything to put it together. So I'm sitting there looking at the instructions and everything is upsidedown or backwards from the way it's laid out on the workbench. It's like I suddenly became dyslexic and the shit gives me a splitting headache in no time. It's a mess. Thankfully, Okie Carol is so good at building RF Decks that she's stuck putting them together every night.
posted by factory_peasant at 11:23 AM
3 Comments:
ol' smokey carol
hehe, ya she trained on decks when i
was an h.p. temp lol.
i wonder how she is doin? i'll hafta
dig up here number and meet her at the
bar.one of the FEW women that could drink me under a table!
R_T
yea she was a rowdy ol' bag...
i had an erector set when i was a kid. it was a little like how you describe building those chasies, only when i was done the things would usualy self destruct, amd i would go play with my action figures.
Post a Comment
<< Home