Please also visit ENSURE I, my first attempt at near-space.
Table of Contents
Fill this in later
When I started the project the first thing I addressed was how the various parts of the payload would fit together. After ENSURE I, I realized how difficult it is to fix and modify parts that are mounted inside a high altitude balloon payload. The "standard" way to design a high altitude balloon payload seems to be stuffing all the parts into a styrofoam box. I wanted to try something else.
The design I came up with can be seen in the diagram to the right. It consists of four 8-in x 8-in homemade printed circuit boards "stacked" on top of each other. Threaded rods and nylon spacers support the stack at each corner, and an eye bolt is attached to each threaded rod with a coupling nut. The eye bolts serve as anchors for the nylon rope that attaches to the balloon.
This design has a number of benefits. First, the design can be compard to an exoskeleton since the internals are supported without any aid from the styrofoam box. The styrofoam box is therefore completely disposable, a handy feature since styrofoam isn't particularly durable. Second, the design makes it possible to field test the payload virtually hassle-free. I was great being able to carry the payload around with one hand while observing the data being transmitted on my netbook. Finally, the design makes it incredibly easy to remove all the parts at once from the payload to make changes. All the internals can be removed by simply pulling the stack out from the styrofoam enclosure.
Image of the PCB "stack" during development.
1. Logic board
The logic board is the "brains" of ENSURE II. Employing an 8-bit microcontroller and a few ICs, ENSURE-2 interfaces with a number of devices including temperature sensors, an on-screen display controller, serial devices, and serial EEPROM memory. I wrote the code in C, compiled with AVR-GCC, and downloaded to the microcontroller via a homemade In-Serial Programming (ISP) cable connected to my PC's parallel port. Below is a flowchard of the logic board:
The logic board contains an Atmel Atmega16 microcontroller and a number of local peripherals, including two sockets for SPI EEPROM ICs, a MAX7456 on-screen display IC, two analog multiplexers, two MAX232 serial transceivers, and a 5V voltage regulator. The logic board's specifications are outlined in the chart below:
|Clock speed||16 MHz|
|Onboard analog inputs||1x LM335 temperature sensor|
|External analog inputs||3x LM335 temperature sensors, 1x battery|
|Nonvolatile memory||2x M95512 512Kbit SPI EEPROMs|
|Serial ports||4 [ multiplexed ]|
|Video character generation||MAX7456 SPI on-screen display controller|
|Video inputs||8 [ multiplexed ]|
|Current consumption||___ mA|
I designed the logic board with a double-sided PCB layout. This proved difficult to fabricate at home, and it was the most complicated PCB I've ever made, but I managed. First, the limitations in the cheap Student version of EAGLE CAD limit the largest PCB you can design to 4-in x 6-in. This was a problem, since the PCBs in my design are all 8-in x 8-in large. I worked around this limitation by designing the logic board as separate PCBs in EAGLE, saving the artwork as a BMP file, photoshopping them together, and then printing the photoshop file with my laser printer to etch the board. I was quite pleased with the result:
2. Radio Modem Board
When I first looked into radio communication methods for ENSURE II, I considered both RTTY and APRS packet radio. I had used RTTY in ENSURE I, so I was familiar with the format. It is attractive because of how incredibly simple it is to implement. RTTY can be used without any extra hardware and can easily be synthesized in software with a microcontroller. However, the downside is that RTTY is slow, typically somewhere between 75 and 300 bps. I was hoping to send all the data gathered by ENSURE II through the radio, not just simple position data. For that, I thought RTTY was not quite ideal.
The other option I considered was APRS (Automatic Position Reporting System). I decided early on not to pursue APRS as an option because commercially-available APRS gear is expensive and I figured it would be a waste of time trying to figure out the protocol in order to design my own hardware.
So I designed my own radio modems. I based the radio modems around the CMX469A radio modem IC from CML Microcircuits. It is an adaptation of the schematic that appears in their product application datasheet. The basic principle of the radio modem is to convert serial data from the logic board to an audio signal that can be transmitted over a normal handheld radio. The CMX469A in companion with a few other ICs does just that:
However, there was one caveat. The CMX469A can only handle synchronous data, and the serial data from the microcontroller and the serial port of a PC are asynchronous. I could have written some software to do this, but I really wanted it to be taken care of in hardware. To overcome this discrepancy, I used a MAS7838 IC to convert the asynchronous data stream into a synchronous format that the CMX469A can handle. I could not find this IC in the U.S., however a quick online inquiry to a few suppliers in China landed me four MAS7838's for under 10$ shipped.
After building the PCBs and fixing a few bugs in the layout, I was able to send ASCII characters from one PC to another. I built two versions of the radio modem, one for onboard the balloon and another in a metal enclosure for on the ground:
ENSURE II radio modem (above), ground station radio modem (below).
|Bitrate||1200 bps (also capable of 2400/4800 bps, however the data could not be reliably transmitted at those speeds)|
|Frequencies||70 cm amateur band [ ~ 440 MHz ]|
|RF power||4W Low / 5W High|
|Packet success rate||90% typical [ measured ]|
|Error correction||XMODEM-type 16-bit CRC checksum|
|Data format||Comma-separated ASCII|
|Packet length||80 char typical|
3. Video Transmitter Board
The ENSURE-2 Video Transmitter Board interfaces two separate RF circuits to achieve a theoretical 2.8-watt RF output at 433.25MHz (NTSC Cable Television Ch 59). Since the board has a lot of extra space, I also mounted the GPS unit to this board as well as the cutdown circuits.
The first stage is a "Mini CH59 Video Transmitter" purchased on eBay. The unit modulates the video signal from the logic board's MAX7456 on-screen display controller and emits approximately 50 mW of RF power.
The signal from the mini video transmitter is then sent to a second stage, an RF linear amplifier, which increases the RF power 17.5 decibels from 50 mW to an estimated 2.81 W. The actual power output is probably a bit lower due to cable losses, impedance mismatch, the mini video transmitter might not actually be emitting 50mW, and the fact that I am purposefully underpowering the amplifier with 12V instead of 15V. Regardless, I calculated this theoretical figure as in the graphic below:
The final signal is then sent to an external J-pole type antenna which relays the video signal to earth. I initially used SPST relays to control the video transmitter and cutdown circuits, however they drew too much current and created too much heat. I later replaced them with n-channel MOSFETs (not pictured). The entire video transmitter board can be seen below:
At ambient room temperature (20°C), the video transmitter amplifier heated up to around 60°C. Since 100°C is the absolute maximum temperature specified in the RF amplifier's datasheet, I thought it would be a good idea to provide some ventillation to the video transmitter board. I installed a small 50 mm PC fan in the styrofoam payload box right at the level of the video transmitter's heatsink and cut a hole in the box on the other side, and the temperature dropped to around 30°C. I run the fan on the 6.2 V bus to reduce the current draw; a little airflow goes a long way.
4. Power Supply Board
The power supply board is a simple design that employs 20 AA lithium batteries and two LM2941 low-dropout linear regulators to provide multiple regulated DC outputs.
The 20 AA lithiums are split into two banks of 10 in parallel giving a nominal 15V. Since each cell is rated at 2900 mAh, the total battery capacity is 5800 mAh giving about 3 1/2 hours of runtime with all systems running. There is an unregulated 15V output, as well as two regulated 11.5V and 6.2V outputs supplied by separate LM2941 ICs. The LM2941's are described very ambiguously in the datasheet as capable of supplying "over 1 A." They datasheet seems to imply that they are rated at 1 A but can deliver larger currents when properly heatsinked. With the aid of the large aluminum heatsink I used, I found that I could indeed pull well over 1 A on both regulated outputs with no issues. A photo of the board and some info on the power distribution is shown below:
|15 V (Unregulated)||Cutdown circuits|
|11.5 V||Logic board, radio modem, video cameras, video transmitter, GPS receiver|
|6.2 V||UHF radio, digital camera, video transmitter fan|
Chart of current draw as a percentage of the total:
For those of you wondering why I needed so many batteries, the above chart quickly shows you the burden of powering hungry video systems.
1. Making the PCBs
In order to save money, I decided early on to fabricate the PCBs at home using traditional toner transfer methods. Unfortunately I was wrong about saving money and I was definitely wrong about saving time. On the other hand, I did have some fun perfecting my methods as I am a do-it-yourself type of person.
First I designed the PCBs in EAGLE CAD. I then tweaked the PCBs in Photoshop and printed them off with my cheap LaserJet 1100 onto "Jet Print Multi-Project Photo Paper" which has a water-soluble gloss finish. I provide these details because they are very important to a successful transfer. I first tried some HP brand paper, which turned out NOT to be water-soluble and stuck to the PCB like glue and completely ruined the board.
Next, I transfered the artwork from the paper to the copper with a heated laminator. The laminator I purchased has a spring-loaded roller which allows very thick objects to pass through it. This is a necessity for PCB thermal transfer. Using a tiny piece of Scotch-brand tape I secured the edge of one side of the cut-out paper to the copper and inserted it that side first into the laminator. I did this repeatedly about 20 times in a row; since the area of the copper is so large for an 8 in x 8 in board, it takes MANY passes for it to fully heat up before the toner will adhere.
After that, I immediately dunked the PCB into a cold pan of water with the paper attached and let it sit for 10 minutes to let the glossy layer dissolve in the water. After that I could remove the paper without any effort. I then carefully rubbed the board clean of any paper or gloss residue.
Finally, I etched the board with a homemade etching solution. I tried all types of etchants from Ferric Chloride to Ammonium Persulfate and they did not work desirably either because of cost or poor results. I finally found out online about an etchant that you can make that seems to work very well and is very cheap.
By mixing 1 part hydrochloric acid (available at most hardware stores, mine was 2 gal for 7$, WAY more than I needed) with 1 part hydrogen peroxide (available at virtually every drugstore, the local Walgreens was having a closeout on their 16 oz bottles at the time and I bought all they had at the store. I got a LOT of strange looks from people.) you get an etchant that is comparable in speed and etch quality to ferric chloride at a much lower cost. (I apologize for the completely unscientific explanation of this etchant; Chemistry is not my strong science.)
The result, a very nice etch:
To finalize the PCB I removed the toner with acetone, drilled out the holes with a Dremel tool and carbide bits, and put each PCB into a pan of "MG Chemicals Liquid Tin", a product which chemically tins copper in 3-5 mins. This protects the PCB against copper corrosion and makes it easier to solder. After soldering I removed the flux from the solder with rubbing alcohol.
Double-sided PCBs take a bit more effort. While in theory I could have etched two sides at the same time by transfering the toner on both sides at once, the need to carefully align both sides required doing them separately by covering up the side I wasn't etching with duct tape. Luckily I only had to make one double-sided PCB (the logic board) and I think I got really lucky on that one.
The cutdown subsystem for ENSURE-2 consists of two independent nichrome wire coils wrapped around the load line. Having two independent means of detaching the payload from the balloon is a requirement imposed by the FAA in their Part 101 regulations. Not only that, but it is also just a good idea in general to have redundant systems.
Electrically, the cutdown system is really simple and consists of the nichrome wire and a large wirewound resistor in series. When the cutdown MOSFET is turned on by the logic board, 15V is fed through the wirewound resistor which imposes a current limit and a voltage drop (easily calculated with Ohm's law) and the nichrome wire heats up enough to burn through the nylon load line in 6 seconds as I found out through thorough testing (see video below).
I decided that the resistor needed to be lower in value, so I put two wirewound resistors in parallel. I scrounged one resistor from the power supply of an IBM PS/2 from the late 80s and the other from an ancient 1200 baud dial up modem.
3. Radar Reflector
Another FAA requirement is that every unmanned free balloon must have a device to reflect radar signals so that the balloon shows up on Air Traffic Control's radar. To do this, I built a quad reflector using mylar, tape, and two thin dowels from the hardware store. I built it the night before, so it was really rough, but in principle it should have worked ok.
Left: TV and radio antennas for the ground equipment. Right: the location of the onboard antennas
Roll-up j-pole antennas were used for all communications.
I wrote the payload software in C and compiled it with WinAVR (avr-gcc for Windows). There was a lot of trial and error involved, and a lot of serial port probing, but eventually I had everything tweaked.
I wrote the PC software in Microsoft Visual Basic 6 a) for lack of better knowledge of PC programming environments and b) because its really really quick and easy to do, provided you can get the stupid MSComm control to talk to the serial port reliably. The result of a day or two of tweaking was a rather slick graphical program to capture data from the radio modem, display it graphically, and send commands remotely to the balloon.
VI. Ground Station
A bunch of equipment were required for the launch, including:
Questions/Comments may be sent to adwiens AT gmail DOT com.