I've decided to move spectrum analyzer posts onto their own page. I've been getting complaints from approx. 25% of my readership (hi, Katie) that I'm crossing the line to "too nerdy". So, I'll let you opt-in to hearing about it.The page is available under the title banner, near the "about" and "thesis"page. I'll write sub pages for each module. Unfortunately, I don't think there's a way to make an RSS feed of changes to this page, so if you're interested you'll just have to check in occasionally. I'll also post when it's all done, so there is that to look forward to. Anyway, that's all. :)
Unsuspecting security camera
Well, as a (perhaps welcome 😉) deviation from the spectrum analyzer posts,I've spent a little time working on repairing a security camera I've been hanging onto for a while. I've been toying with the idea of installing it near the radio control flying field as someone in the club knows the owner of a nearby business.
The problem with the camera is that it has trouble switching from nighttime mode to daytime mode. At night, it is sensitive to Near Infrared light (such as what is transmitter from your remote control). During the day, however, this sensitivity makes the colors look strange. To cope with this, they have a filter that slides in front of the lens for the daytime. This filter binds, and because of this, the camera is always in nighttime mode. I decided to go ahead and disassemble the camera to try to fix it. My motto: If it's broke and out of warranty, take it apart! 😀
Removing the lens
You can (kinda) see in this photo the image sensor. This metal mount can move forward and back to accommodate different lenses and adjust their focal length.
Bottom cover removed
Removing the bottom cover reveals the power conversion board.
Removing the top cover
Removing the top cover reveals the PCMCIA slot. This camera is kinda cool in that it can accommodate either a memory card or a wireless LAN card in this slot. Also, there is an ethernet port on the back. It has an embedded web interface that allows the control of camera functions and viewing live video.Apparently, there is the ability to get some information from the serial port, but I haven't found any information about it.
Front lens mount
Like I mentioned before, the front lens mount can be adjusted front-to-back.The black screw near the top of the frame is used to secure the mount. More to the right of the frame, the spring is used to press the mount against a set of wedges that control the depth setting. The 4-pin connector goes to the lens iris, which is kinda like the aperture of a still camera.
Lens mount disassembled
This is the front of the camera disassembled. The image sensor is still attached to the camera frame. The black piece of plastic is the filter module, and the other piece is the metal frame for mounting the lens and adjusting the focal length.
This is the device causing all the trouble. The slight green tint in the left frame is the IR block filter, and the right is clear. The motor is a small gear-head motor attached to a worm gear. The worm gear has very shallow cuts in it. The spring pushes down on a small plastic follower. The whole system is intended to allow the worm gear to continue turning even if the system is jammed. I think this is so that limit switches aren't necessary. I figured that the problems I'm having are due to excessive friction, which would cause the frame to remain static while the worm gear turns. My first thought was to place a little petroleum jelly on the sliding surfaces. I re-assembled the front of the camera and found that the problem remained. However, I noticed that when the lens mount was all the way against the camera it would stick. I could then solve the problem by keeping the lens not-quite against the camera.
Lens mount disassembly
Before re-assembling the camera, I decided to lubricate the adjustment assembly. The black plastic ring on the right includes the adjustment wedges.The middle ring is the precision-machined lens mount with a channel for the plastic ring. The punched metal piece on the left completes the assembly.
Anyway, I put the camera back together. It works great now. I'm a little embarrassed to admit that the problem was simply that the lens can't be in the closest setting for the filters to work. At least everything went back together without a hitch. :)
IF Amplifier cooling down
I've decided to hold off posting spectrum analyzer modules until they're complete. I've been collecting tons of pictures along the way, so hopefully the post for each module will be interesting and visually appealing. With that in mind, I can't post about any of the modules with the exception of this one.
For whatever reason, my kit didn't come with any voltage regulators, and this is the only module besides the mixers from my last post that don't use any.The IF Amplifier, as the name implies, amplifies the Intermediate Frequency on the analyzer. I made a simplified version of the block diagram from Scotty's site. I've also indicated the modules I've finished in green.
Block diagram, green modules are completed as of this post.
As you can probably see, there's a lot to be done. Luckily both DDSs, the LogDetector, ADC, Control Board, and one PLO are done once I get some parts to replace the missing ones. I also need some parts for the Master Oscillator.Once that (actually quite small) order comes in, I'll be almost finished with the boards.
Anyway, back to the amplifier. I was one of the first people on an order of boards back in August '08 because I agreed to look-over a new revision of the design files. Of course, by having me look them over essentially guaranteed that there would be errors. As it happens there was an error on this board.The engineering change order (ECO) is luckily quite simple, the only problem was that a short section of the border was missing. Here is a photo of the completed fix:
Fix for the slight error
You should be able to see that along the border there is some solder wick saturated with solder bridging the gap, and connecting to the capacitor. Also, on (at least) the 2 resistors (R3 and R4) there are tiny balls of solder.These happen during the reflow stage when a bit of solder squeezes out from underneath the device. I haven't gone over the board picking all these out yet in this photo. I think it's important to remove all of these because they may cause shorts. Also, you may be able to see a slight, shiny, residue around everything. This is the solder flux that's included in the solder paste. I remove this later with some 99% rubbing alcohol.
Completed IF Amplifier
This is the completed amplifier board. There are actually 2 amplifiers here, mirrored. In practice, the output of one will be the input of the other. This serves to roughly double the gain. Finally, notice on the right side, that many of the parts have been omitted. On the schematic, Scotty simplified this section. Technically, he has 2 zero-ohm resistors (basically jumpers) in addition to the capacitor. I chose to just jump the capacitor across and leave the resistors out.
Anyway, I enjoyed building this one. I hope you've enjoyed reading about it. I can't wait to finish the other ones!
Mixer 2 completed and verified against schematic
I've verified each of the 3 mixers that I need for the spectrum analyzer. The one above is mixer 2. The schematic for this mixer is available from Scotty's site.
Mixer 1 verified against schematic
I've also verified mixer 1 against the schematic. There have been 2 optional revisions posted that aren't included on mine. I'm going to add them later, but only if necessary.
Mixer 3 verified against the schematic
Finally, mixer three is complete, and verified against the schematic. This one isn't strictly needed for the spectrum analyzer, however it is necessary for the tracking generator. A tracking generator is a module that generates a signal that matches what the analyzer is tuned to. This is useful for tuning filters, or other passive devices.
I'm missing some parts that I need to complete the control board. I hope to get replacements soon, so I can begin testing some of these modules… Stay tuned!
IMPORTANT!! - I've made a special page for my construction and testing of the spectrum analyzer. It's available on the navigation bar below the blog title banner, or here.
A while back I bought a spectrum analyzer kit. It can display waveforms in the frequency domain from 0 to 1000 Mhz. I haven't worked on it for a while, butI've decided to start working on it again. If you're interested in more information, there is a webpage for the analyzer by the guy who designed it, and there's a Yahoo group, also. If you're really-really interested there is a group buy open (until the 28th of march2010) here. Anyway, I'm sure it's of limited use to my usual readers (thanks, friends & family! [oh who am I kidding, most of my family doesn't read this ;) ]), but it may be useful to others working on the kit. I'll post the "annotated" CAD images for each board here, and add images of the completed boards when I finish them. Who knows, it may be interesting for someone. Also, I'll include my thoughts and observations while building these boards. Finally, not all the boards that come with the kit include annotated layouts, so I'll have to dig around and figure out what they do and what to put on them ;)
Direct Digital Synthesizer:
I was able to find some designs online that fit my needs. The major constraint that I had to deal with is the impedance issue. Nearly everything amateur radio related is 50 ohms, and nearly everything video related is 75. The cheap Yagis paper written by Kent Britain, WA5VBJ, has a 75 ohm 421 Mhz antenna intended for amateur television (ATV). My transmitter is 434 Mhz, but I figured it would be close enough. The great thing about these designs is that they can be built using supplies from a standard hardware store. The elements are made from #10 bare copper wire, and the beam is wood.
An interesting characteristic of antennas, and RF in general, is that to get a stronger signal you often have to make compromises. A Yagi works by increasing the directionality to increase the signal. Unfortunately, because my plane is going to be flying around, I can't be too directional. To get better results, without making things worse, I decided to only use the reflector and "driven element". By eliminating the "directors" I hope that I can get the best possible results. (If you're confused by this "director", "reflector", and"driven element" gibberish, the best place to look is the wikipedia article. But all that is necessary for this discussion is that the reflector is behind the driven element [which connects to the transmitter or receiver] and reflects the signal forward, and the directors go in front and focus the signal into a narrower beam).
Feed line soldered
There were some problems during construction that I should mention to help others wanting to try something like this. The antenna designs specify that the feed cable should be soldered directly to the driven element. This should work great on traditional 50 ohm radio cabling, such as LMR or RG-type cables.These cables have copper braid shield around the circumference. With the 75ohm cable used in video, often made as cheaply as possible, a loose aluminum braid is used as the shield. This is a major problem that I had to deal with.It took me a while to even understand why the braid wasn't soldering. I think I assumed that the braid was made of tin. After a few hours of searching, I discovered it was aluminum. Aluminum oxide forms almost immediately and can't be soldered to, so even sanding the wire doesn't help. There are solder pastes and fluxes that help, but I wasn't interested in waiting for something to be shipped. My solution, if you want to call it that, was to mechanically attach the braid to some other wire that can be soldered.
In the first image of the post, I'm comparing the new antenna versus the others I used earlier. When the Yagi was installed, I rotated the antenna 360˚in azimuth to get an idea for how directional the antenna really is. There wasn't much of a change in signal quality, so it isn't very directional. If the transmitter were further away it may have been more dramatic. I am motivated to build a few more antennas, maybe with 1 and 2 directors to see which is better. With that said, I'm pretty satisfied, and I'm hoping for good weather this weekend.
b.t.w: Just to dispel any fears that the shield is shorted to the center conductor, as it appears in the above photo, it was, and I fixed it. Here is a photo of the feed point as it was when I tested it. Also, notice that I got my driven elements and directors confused when I wrote on the board :)
Another view of the feed point
Taking advantage of the crummy weather, I decided to paint the camera module to match my plane. I didn't just do it for the aesthetics, no really, I swear.:) Actually, the real reason I painted it was to prevent stray light from making annoying reflections on the inside of the window.
The first thing I had to do was get an idea of the size and shape of the opening that would be needed. I was able to measure it by looking at the video output and marking with a sharpie just outside of the camera view. Once I had all the sharpie marks, I tried to make the shame smooth and symmetrical. SinceI had to mount the camera low in the module, as discussed in thispost, the opening is low, and very parabolic. I then covered the whole top of the window with masking tape and by shining a light through the bottom traced the sharpie line with a hobby knife. Removing the excess tape left a nice mask for the opening.
With the opening masked, I sanded the window to give the paint the best chance at adhering. It is apparently a little hard to find plastic spray paint (like Krylon fusion) in silver, so I had to use some general purpose kind.
I think the paint on the outside of the window turned out O.K. There is a little dust and lint marring the finish, but it turned out not to matter much.I also wanted to paint the inside with a matte black finish to minimize the chance of reflections from light that is still able to get in through the smaller opening. I used a similar approach to the technique for the outside. I placed masking tape on the inside of the window and shined light in from the other side. I again traced the outline with a hobby knife.
For some reason, I really like the way this turned out. It's all pretty darkin there, and should minimize reflections. I also painted the base of the module to match.
I think everything came together nicely. I really like the finished look, even though I was pretty sure that it was going to look weird. As I mentioned earlier, the lint on the outside paint didn't matter. I put masking tape on the front of the window to protect it while I painted the inside, and it lifted part of the paint when I removed it. This gave it a strange effect, andI'm not sure whether I like it or hate it. Katie says it looks cool, so I'm leaving it for now.
There should be good (enough) weather this Saturday, so hopefully I'll have some video to post!
I've just finished building a camera module for my Kadet. When I was building the plane I knew that I was going to try and put a camera and transmitter approximately where the pilot's head would be in a real plane. That means a couple of things: First, it's pretty cramped up there with the flight pack, speed control, and motor. Second, it's a bit of a dirty secret, but forward visibility is usually not so great in most planes. I tried to mount the camera as high as possible, but the front of the lens keeps hitting the glass. I was able to get most of the view clear, but there is a margin on the bottom of the frame that is obscured by the wooden frame. I will probably install some LEDs there to monitor battery charge, current, etc…
To build the outer frame of the camera module, I traced the shape of the wooden windshield onto a piece of acrylic. After cutting it out I traced the inside of the fuselage onto balsa. I cut out two side pieces and just glued them onto the acrylic. I had to cut a notch into the back side of it to make room for the dowel that the wing rubber bands stretch over. You can see how it fits in the first photo that I posted.
I made a bottom plate by tracing the outer frame onto a piece of balsa. Then,I took some balsa stick and made some uprights. After drilling holes for the screws I soaked some thin CA into the sticks to keep them from splitting.
The image above is the completed camera module. At this point, there aren't any electronics other than the camera module. If you look really close, you can see the hole in the top of the windscreen on the top-left (the plane's right). This is for the video transmitter's antenna.
The schematic for the camera/transmitter circuit is pretty straightforward.The main consideration is that the camera and transmitter require different voltages. The camera works best around 12 volts, and the transmitter uses 9volts. I only wanted to use one battery, so I had to make a voltage regulator.I used my old standby, the LM317. I also wanted to be able to adjust the transmitter voltage, so I added a potentiometer. The transmitter came with aRCA plug and the camera had a BNC plug on a long cable. I wanted to get rid of all this extra bulk, so I cut them off and connected them using the circuitboard, also.
The transmitter works on 433.92 Mhz, which is surprisingly hard to find in consumer products. The obvious reason is that it's in the amateur radio band.But, 433.92 is actually channel 59 on cable. This means that you can use a commercially available USB TV tuner to receive the video. With that in mind, I gathered a few antennas to try. The above photo is an antenna I bought to try.It's a VHF/UHF type from radio shack, I think. I just went for the cheapest available. I also looked for some antennas with "preamplifiers", but I haven't found any that amplify the 400 Mhz band.
I also tried a whip antenna from an old 900 mhz wireless video system. It's not tuned anywhere near 400 mhz, but I've been lucky before.
I embedded a YouTube video of me walking around the house using each antenna.The bunny ears are much better than the whip antenna. The good thing about the whip is that I may be able to make another whip that is tuned correctly. If the weather's good tomorrow, I'll probably try it out. I'm pretty excited!
Ready to fly at the BCRCC PolarBear
I'm back from the third (and fourth) flight, and it keeps getting better! OnFriday (New-Years Day) I went to the BCRCC (Benton County Radio ControlClub) "Polar Bear" event. Basically, the idea is: On new years day, rain or shine, everyone comes and flies something. There is a raffle for all those that fly. I brought the Kadet, mostly because that's the only plane I have, and the tiny Blade mSR that I got for X-mas.
The Kadet's second flight was somewhat uneventful. Flying during an event invariably leads to crowded skies, because I didn't have the runway to myselfI just tried to stay out of people's way. I was still pretty unfamiliar with my plane so it was nerve-wracking. Luckily I did a good job landing with all those people watching :).
Today, I went back to the field with my Kadet, and a new plane that I got at the BCRCC auction for $20 in November. It's a "Funtana Mini" made by E-flite that has since been discontinued. Since the auction, I've been slowly getting the parts I needed to complete it, including a receiver and servos. I reused the motor from another plane that is no longer with us.
My new sport plane
I piloted it on its maiden flight today. It is extremely twitchy, even on low rates. Also, something I didn't expect was how hard it would be to land it. It requires a fairly high airspeed to maintain altitude, and there isn't much frontal area to slow it down, so the landing was fast. I wasn't able to stop it by the end of the runway so it spent some time in the mud. Nothing was damaged, but the mud at the airstrip is very stinky. The second flight was more fun, I actually did some aerobatics. I still had trouble bringing it down, though.
I wasn't satisfied with the prop I was using in the first flight test, which was an 11x7 APC E-series. I bought 2 more, one with higher pitch speed (for the uninitiated, "pitch speed" is the number of inches a propeller would travel forward in one revolution in an ideal fluid) and the same thrust (which how hard it "pulls"), which I think is an 11x8.5 E-series, and one with the same pitch speed and more thrust; a 12x7 E-series, I think. I tried the one with more pitch speed first, and I'm not rushing to try the other one. I'm quite happy with the performance as it is. At full-throttle I can climb-out at45°, and at 1/4 throttle I can maintain altitude and speed. I'm also very happy with the observed endurance. I ran the stop-watch on my radio during the second flight of the day while doing non-stop touch-and-go's, and after 12minutes of flying called it quits. The resting battery voltage on my 3S 2100mAh LiPo was 11.4 volts. This roughly corresponds to 20% of the pack power remaining. With this knowledge in hand I know that I can set the count-down timer on my radio to 12 minutes and not stress-out my batteries too much.
Altogether, it was a good (long) weekend of flying!
I've been developing a RF transceiver for use with my Kadet. I really like these Micrel RadioWire modules, but they basically must be reflow soldered.They're also about $16/ea. in single unit quantities, so I want to make sure that the carrier PCB works before I solder on the module (I've made that mistake already!). With that in mind, I had to figure out how to solder the module on after all the other components were installed. This means hand soldering. Looking on the datasheet, they say that it can be done, using solder paste. So, here's documentation of my attempt.
PCB ready for module
This is the starting point. I cleaned the surface with isopropyl alcohol(IPA).
Solder paste applied
I applied the solder paste as normal, and cleaned the module with IPA as well.
Module seated, waiting for heating
This is just before heating the pads with a soldering iron. I heated each one until the solder melted, but it didn't bond with the module. Before I got too far, I removed the module and examined the results.
Results of the first attempt
I cleaned all the surfaces with solder wick and IPA, and tried again.
Results of attempt two
This time, I held the soldering iron on each pad for much longer. The solder bonded to the module and the PCB, and it held tight. However, I took a multimeter to it, checking for continuity. Nearly every pad was shorted with it's neighbor. This sucks.
Another view of try two
With the module throughly bonded with the PCB, and no way to remove it and try again, I decided to try the hail-mary. I figured that at this point, unless I could think of something, the whole thing was ruined. With that in mind, I decided to try to reflow the whole thing. I really had no idea if it was going to screw everything up, especially the FTDI chip on the bottom of the board.
The reflow went well, I'm pleased with the look of everything. I also confirmed that the pins are electrically isolated. Also, on first glance, everything seems to work. The microcontroller and USB-to-serial chip still work. I really just need to verify that the transceiver module works, and it will be a complete success.
The moral of the story is: Don't attempt to hand-solder the Micrel module. It probably won't work. On the other hand, if you do get it to work, drop me aline, I'd like to know how.