As I wrote in the previous post, I built 2 GPS loggers with same parts. Satisfied to finish creating my own GPS logger, I brought both outside, and took some GPS logs. Then, analyzing logs, I noticed that the stacked one consistently reports a little weaker signal reception. I wondered why and did some experiments.
Experiment 1. Stacked vs Flat
(A) Stacked GPS logger
(B) Flat GPS logger
First I evaluated two receivers: (A) stacked and (B) flat. As described in the previous post, both are made of exact same parts. Photos of both receivers are shown again in the right. I took a walk for about 5 minutes in my neighborhood, putting each receiver on each palm of my hands.
The evaluation result is below. By the way, according to the "FGPMMOPA6H Data Sheet", SNR value reported from this receiver is C/N0 [dB Hz], and so are the numbers in the table. "x" in the table means the satellite was never reported from that receiver.
| Sat# | A | B | Diff. (A-B) |
|---|---|---|---|
| Total | -67 | ||
| Mean | -5.2 | ||
| 10 | 38 | 43 | -5 |
| 12 | 41 | 45 | -4 |
| 14 | 20 | 18 | 2 |
| 15 | 33 | 36 | -3 |
| 18 | 33 | 39 | -6 |
| 19 | 5 | 11 | -6 |
| 20 | 0 | 17 | -17 |
| 24 | 36 | 42 | -6 |
| 25 | 36 | 40 | -4 |
| 29 | 0 | x | 0 |
| 31 | 0 | 12 | -12 |
| 32 | 36 | 39 | -3 |
| 193 | 33 | 36 | -3 |
As we can see, receiver A reports almost all satellites' signals are weaker than receiver B with only one exception: satellite #14. Total difference is -67 dBHz and average is -5.2 dBHz.
Discussion
Following the above result, I did some more experiments as follows.
- Switched hands.
- Put both receivers in same Zip Loc container both with covered and uncovered.
But the results are consistent: receiver A has worse signal reception with almost the same difference as the previous experiment. Possible reasons left are (1) variance of GPS module and (2) stacking has somehow negative effect. So I did the next experiment.
Experiment 2. Flat vs Flat
(A') Unstacked GPS logger
I disassembled the receiver A, put GPS module and CPU module on a large breadboard and wired them. Wiring is same as the receiver B. See the right for the picture. I will refer this receiver as (A') (a-prime) in the following.
| Sat# | A' | B | Diff. (A-B) |
|---|---|---|---|
| Total | -28 | ||
| Mean | -2.2 | ||
| 10 | 41 | 43 | -2 |
| 12 | 42 | 45 | -3 |
| 13 | 0 | 0 | 0 |
| 14 | 12 | 15 | -3 |
| 15 | 39 | 39 | 0 |
| 18 | 36 | 40 | -4 |
| 19 | 2 | 11 | -9 |
| 20 | 26 | 28 | -2 |
| 21 | 0 | 0 | 0 |
| 24 | 39 | 41 | -2 |
| 25 | 34 | 36 | -2 |
| 32 | 35 | 33 | 2 |
| 193 | 34 | 37 | -3 |
As we can see, signal receptions of receiver A' and B came to much closer than the previous experiment. Although A' may be still a little worse than B, but the difference is small compared to one in the previous experiment.
Thus, I would conclude that stacking CPU module and GPS module has some negative effect in signal reception.
Since further investigating the cause requires more expensive equipment, I will not go further, but the cause might be due to noise from CPU module or SD card circuit.
Anyway, with this result and the fact that stacking hides some LEDs, I will stick to the flat structure hereafter.
Notes on QZSS
As we can see in the tables, there is only 1 QZSS satellite reported (193). However, according to the QZSS operating status, other satellites (194, 199) are operational (see L1 C/A row for the consumer grade GPS functionality). Part of my motivation of this project is to utilize new QZSS as I write at the beginning. So, I may look for other GPS unit later, but replacing GPS unit may require updating wiring and modifying software. So I will continue to use the current Adafruit's units for a while.