Development history

This story is more sad than the previous one, because it does not have that big good part in it as the previous one :).

In spring 2012 I spent about two month for developing and testing the MiraField program, this year I spent a few weeks modifying it. Last year I did not do enough tests and I thought there is no problems with kinematic methods in open areas. This year I tested it on a real project and realized the the kinematic is a total disaster. The problem is not with MiraField program, there must be a problem with the raw data, I think it is with the overflow of cycle counter which jumps by 1575420 every 15 minutes, similar happens to preudoranges - the counter jumps by one light millisecond. But, on the other hand, this is not a big problem since there is no cycle slips, this is just counting that starts from a new number - it is easy to fix.

I tried to fix those jumps in my Mirаpract Tools program but I did not have any positive results yet, and it might be that I will not have solution ever. I processed data in Trimble Business Center and GNSS Solutions, and in both cases the results are bad. I hope to find the problem when I develop my own RTK program, this is described bellow in new section.

If you have nothing to do and want to play with my kinematic data in some other program - the raw data are bellow, and if you could share your results with me that would be great :)

In order to get rid of a laptop, which was part of my system, I had to build a compact circuit that would control gps, write data to a memory card, signal system status with LEDs, monitor batteries charge level etc. I found out about microcontrollers, purchased a development kit and a compiler and started to learn this new world. I added to my system a chip that monitors batteries charge level, checks power button, signal level converter, signal lines switch chip etc. When I build my first prototype on a card board and it worked that was quite a feeling:) Turning a cardboard into a pcb was not easy either. Found program called eagle, learned how to use it, designed PCBs, eventually generated gerber files, sent them the to the factory and in a few weeks received my PCBs with components soldered. Because of my numerous errors each version did not work properly. Every year I was developing a new version. Only third version works fine.

1. Development kit	2. First cardboard prototype, no laptop is needed

3. First pcb version, (top two boards)	4. Second version

5. Third version, all on one board

Now I wanted to make a nice plastic enclosure. So first I had to design it. I learned 3d commands and drafted pcbs then 6 batteries underneath, and than wrapped everything with ellipsoid shaped surface. It took me 1.5 month. Now I had to materialize this enclosure in real world. I was going to cut the model in thin layers, print each layer on a plastic sheet, cut each sheet along printed lines and clue cutouts together building up real enclosure - that was my crazy idea. But I was lucky to find out that 3d printers exist :). That was quite a discovery for me. I found several vendors, sent them my files and received a few quotes. Ordered and in a few weeks I received my materialized parts. That was cool to hold in your hands what was only in you dream a short time ago. Now I had to make a mold and cast copies of my parts. For this I build a vacuum and pressure chambers, purchased liquid plastic and rubber from smooth-on and started. It took 6 month to learn mold maling and casting and cast 35 enclosures

1. Learningthread fundamentals and drafting took 2 weeks	2. CAD model sent to the vendor

3. Parts received from the vendor, cost me 800$	4. The company sanded off more plastic than was needed while finishing. I was fixing this by myself for like 20 days. Was not very happy

5. Same story with these edges	6. Attaching holders for battery clips

7. Converted paint pot into pressure chamber	8. Home made vacuum chamber

9. Making mold for a part.	10. External wall shifts up, clay is removed and second half of the mold is poured

11. Making vents	12. Plastic injection

13.Removing the part	14. Almost ready

15. I tried this bugger 12 times before I got it fault free	16. The end of six month story

I did not want to depend on the antenna manufacturer so decided to build it by myself, just one more crazy idea :). Started from here www.antenna-theory.com since I did not know anything about antennas. Decided to build a patch antenna because they are made in similar way the PCBs are made, so I could order them at the same factory. Tried 4 probe patch antenna but realized that the EM field of the antenna was affecting a hybrids bellow it, a device that makes 90 degrees delay in phase, which I was designing by myself as well. To keep hybrids intact I needed 14mm spacing bellow the antenna. Did not have enough room in the enclosure to place hybrids this way so build a single probe gps antenna. Right Hand Circular Polarization in such antennas is achieved by adding a slot in the radiator located at 45 egrees to the probe. The best horizontal phase center stability is achieved by using 4 probes (Trimble zephyr) or a spiral array (Novatel Pinwheel). Antenna took me 5 month to learn design and build.

Isolagranted me 2 sheets of high frequency laminate	Modeling the antenna, added second slot and that improved axial ration of the antenna a lot

Materialized antennas	Soldered coax cables

Seems like easy but not at all, again spent lots of time designing the switch and lots of money for making it. When a vendor sent me switches the layers did not stick together bellow LEDs because my design did not leave enough room. Manufacturer agreed to fix it by cluing but the clue spoiled the graphic layer on 20 switches so only 30 look relatively good. These 50 switches cost me 1500 US$ (which is a good price though, 30$ per switch). Design is very important stage, the design errors can not be always fixed !, I learned that hard way - from my errors.

My schematic and explanations sent to the manufacturer	My artwork

Materialized switches	Lemo granted me this socket, water-pressure -dust prof

After wires soldered to battery clips, bluetooth modules and lemo sockets it takes 2 days to assemble 1 receiver.

battery clips and wires cut for soldering

programming bootloader into my processor, after that my firmware can be updated via serial cable, no programmer will be needed	left to right: main board + mem card board, bluetooth, lemo socket, battery termnals, antenna, memory card cable, enclosure

	shrinkwrapping wires

	soldering bluetooth and lemo socket wires to my pbc

	checking if bootloader works before closing the enclosure

antenna positioned in place, antennas in all receiver are positioned in the same direction	final power check

I purchased an oem gps module, voltage regulator, signal level converter, as it was specified in the message and started to experiment. I had no knowledge in electronics so I started from scratch. Most of the information about how electronics components work I took from data sheets, wikipedia and google. Eventually I connected the module to a pc and when I saw first several lines of binary messages on terminal screen I thought I achieved something extraordinary :), for me that was quite a step.

oem board (left) transceiver with caps and DB9 connector (right)	All in a metallic box

My next step was to write a program that would control gps, show info about tracked satellites, and write data to hard disk. After that I had to write a program that would generate RINEX files from the data stored on disk. After I spent entire summer learning C++ programming and serial communications I developed a data collection program and RINEX converter. Parsing binary files and forming RINEX files turned out to be a very laborious task. I spent a lot of time studying GPS interface manual before I decoded ephemeris. Eventually a had a system that I could survey with:

data collection program

As there is no better flashlight than maglight, there is no better case than pelican :). So I took the appropriate one, than I built a device for foam cuttingsimilar to this oneand in three days finished

Sells like this	Template attached

.... It has been a while since I wrote here, have not been able to find any time for this yet. I just wanted to update this page before I start developing my own RTK program. In last 1.5 years I have had my antenna calibrated, wrote a datacollection app for java phones and also spent about seven month learning math necessary for RTK program development.... so I will start with antenna calibration: Any GNSS antenna is not geodetic if it does not have calibration - a table which has horizontal and vertical electrical phase center offsets in relation to some reference point, (center of bottom of ant mount), same table also has vertical variations of the phase center as a function of satellite elevation, from 10� to 90�, with 5� step. One of the organizations that does antenna calibration isNGS. Even dreaming about your home brew antenna being calibrated by such company is already naive :). I wrote them a letter describing my gps story and to my biggest surprise the organization agreed to calibrate my antenna and they also did it for free!. I would like to mention names of the people who did the calibration but I do not know if it is appropriate. I really appreciate the wonderful job they have done and this is the best gift I have ever had during the system development. I can not imagine something like that happen in my country. So I have chosen 4 receivers, got the antenna RF cables out of them and shipped the receivers to USA and sometime in February 2012 I got the results.

On the background the receiver has lemo connector installed, which I had to remove in order to get the antenna RF cable out of the receiver	I built special adapter cable between MCX connector type and N type

Now the antenna inside my receiver can be connected to an external GNSS receiver used by NGS	Everything is ready for shipment for calibration, shipped by the cheapest and lowest priority mail option by Ukrposhta
Bellow is the relative calibration for my receiver. The first part contains average values of phase center variations from three antennas. In the second part of the table there are errors calculated from the average for each variation value. The NGS tolerance is 1.0mm. Unfortunately for my antenna the errors are 1.4mm (marked in red). Because of this my antenna is not published on NGS website. It is sad, but what can you do..., 0.4 mm out of tolerance. I should have centered the antenna better inside the receiver and should have positioned it vertically better as well, but I was not fussy enough. In the next version of antenna and plastic enclosure I will design better antenna holding system, if I ever get to that point.
MRAN1+NONE -1.2 0.0 75.3 0.0 0.4 1.2 2.3 3.4 4.4 5.1 5.5 5.5 4.9 3.9 2.3 0.3 -2.1 -4.8 -7.7 -10.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 rms - 3 measurements 0.8 1.4 0.7 0.0 0.1 0.2 0.1 0.1 0.1 0.2 0.2 0.2 0.1 0.1 0.2 0.2 0.3 0.5 0.8 1.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0