Consistent high "Estimates without confidence" values

[swart-j]

Hello,
We received SnapperGPS units from the Arribada Initiative (thank you for selecting us!) and selected the ECHO 27 Antenna for each unit. Each unit is powered either by 40 mAh or 110 mAh lithium-ion battery. After multiple tests with each unit, our lowest “Estimates without confidence” is 23.2%. We typically get 90 – 100.0 in our tests. Do you have any recommendations on what we can do differently to achieve Estimates without confidence under 1.5%?

Background information:

• Testing Locations: southern Clark County, Nevada, United States

• Google Maps

• All units are placed outdoors, away from buildings, with clear views of the sky. The units have been placed in parking lots, backyards, and out in the open desert.
• We verified the connection between the antenna and SnapperGPS receiver. We have not disconnected and re-attached the antenna’s more than 3 times.
• All units pair to the computer and successfully configure
• All units run Firmware version: 0.0.7
• Uploaded the data 1 – 3 days after it was recorded.
• We upload the data via the SnapperGPS app and the browser version using Microsoft Edge
• The LED patterns are consistent with the info found on SnapperGPS - Home
• Max velocity has been left as null or 1 meter/second for our stationary tests
• We've uploaded the data without changing the Start and End Date/Times and we've also re-uploaded the data with trimmed dates to verify that the GPS units had a clear view of the sky the entire time.

[amanda-matthes]

That sounds unusual. I suspect that the low "max velocity" may be the culprit. Could you send the raw data from one (or multiple) of your tests to [email protected] so we can take a closer look? You can download it from the tag using the "transfer data" button on the upload page. The .json file should be enough. If it's a large file consider placing it on a shared cloud and just sending us the shared link.

[swart-j]

Thank you for getting back to me so soon. I sent a couple files your way.

[amanda-matthes]

Thanks for sending them!
I had a look through the data and most of the recordings do not show any signs of satellites. A few do but they have fewer satellite signals than I would expect. Some of the recordings also show signs of corrupted data. I have never seen anything quite like it before.
I checked the serial number of those SnapperGPS boards and they all worked well when we tested them before sending. So unless they were damaged during shipping, there is a problem with the test setups. There are two things we should rule out first:

1. Was the antenna facing the right way? The side with the solder dot needs to point to the sky and the metal side needs to face down.
2. Was there any metal touching any of the electronics that could have maybe caused a short?

If those are not the issue, could you take a picture of the test setup and a close-up of the board? Maybe we can see what's wrong from the picture.

[swart-j]

1. Yes - we kept the solder dot pointed to the sky.
2. I don't believe any metal touched the boards besides the li-po battery and the metal side of the antenna. Should we insulate the antenna and battery?

I attached pictures of our setup and boards. We keep each unit inside a clear bag, with the antenna facing the sky. Then, after the test runs for a few minutes, we will sometimes move the units under shrubs or a rock overhang.

[amanda-matthes]

Putting some insulating PVC tape on the battery and the back of the antenna is not a bad idea when there is a chance of them touching the board. However, looking at your pictures it seems unlikely that that's the cause of the problem. I haven't been able to replicate it.
Can you try another test with none of the parts touching? A snapshot every second for a minute should be enough.

I checked the dates on your recordings, and the "data corruption" I mentioned before only shows up on tests from October 9. I've never seen anything quite like it. Is there anything different about those tests that you can think of?

[JonasBchrt]

Hi @swart-j

I am sorry to hear that it isn't working as expected, but thanks for posting about it here. This makes our conversation visible to others who might face similar problems in the future.

I 100% agree with Amanda that there is something unexpected going on since your receivers worked well when we tested them in Oxford. So, it is likely related to the specific setup. I hope that Amanda's recent suggestion(s) might help.

I have nothing to add, just want to illustrate her comments with some visuals.

Here are some recordings from your receivers that we did ourselves:

• https://snappergps.info/view?uploadid=d8c147a0f6
• https://snappergps.info/view?uploadid=51de1d9d57
• https://snappergps.info/view?uploadid=1f747b6707
• https://snappergps.info/view?uploadid=823c5bd406
• https://snappergps.info/view?uploadid=15aff85cac

I would expect qualitatively similar results when you deploy them.

Sky visibility for these experiments was good, but not perfect (building on one side). The setup was very similar to the one below from the getting started document with all the components being placed on a table and apart from each other.

[swart-j]

Thank you for the suggestion! I'll try the insulating PVC tape this weekend and see if we can get the desired results.
I'll keep experimenting with setup arrangements to see if I can identify the issue.

I tested the SnapperGPS units this morning at one of our field sites. I placed the units in the open, with a full view of the sky. I then put a few small rocks to keep the units in place and shade the li-po battery from the sun.