HAS/E6 results on PostCard FW V3.1 with LG290P FW V2.01 [Feb 2026]

GPS/GNSS
102

As soon as the PostCard has a PP RTK fixed position shown on the OLED screen, switch to Assisted Base with the shield’s button, and turn off the Starlink Mini. Am I missing anything important ?

This is a good path. The built-in NTRIP Client will request WiFi turn on. If no network is detected, it will retry 15, 30, 60, 120 secs, etc up to 15 minutes. Once SL is up and running, Postcard should RTK Fix quickly, at which time you use Base Assist to ‘drop the pin’ where you’re at and start broadcasting.

I don’t think turning off NMEA is needed. See this line of code.

// Determine where to send RTCM data

    if (inBaseMode() && type == RTK_RTCM_PARSER_INDEX)

    {

        // Pass data along to NTRIP Server, ESP-NOW radio, or LoRa

        processRTCM(parse->buffer, parse->length);

    }

So currently, only RTCM (while you’re in base mode) gets passed to the NTRIP Server, LoRa, and ESP-NOW. All other NMEA, UBX, PQTM, etc sentences are ‘dropped on the floor’ (but are sent to the general buffer for Bluetooth, TCP, and SD consumption).

Clear as mud?

Pair the radios to allow the longest range mode, 256 Kbps when conditions require it.

This is news to me. I’m not convinced ESP-NOW is smart enough to auto-throttle the data rate. For my own breadcrumbs here is where we set the WiFi protocols including WIFI_PROTOCOL_LR which is responsible for ESP-NOW. I am not aware of speed settings for ESP-NOW. Let me know your source.

Use Wi-Fi channel 11 in the USA.

Interesting. Why? Are you just trying to find an open WiFi channel? If you’re in the woods, I don’t imagine you’ll have much (if any) 2.4GHz traffic. I use the default 0 channel all the time without issue. But then again, I haven’t tested range with other channels so please do test it.

Turn off Bluetooth on the Base to free up the radio.

I don’t think this is needed, but feel free to test it. Bluetooth and ESP-NOW can coexist, but it’s not a friendly coexistence. I see packets dropped if I’m connected over SPP. If nothing is connected, I see packets with a 0 to 2 second age (meaning a packet was blocked or corrupt everyone once in awhile). If SPP is connected, I see RTCM packets with a 20-30s age. If BLE is connected, I see RTCM packets with a 0 to 10s age.

Should anything else make the proposed list to maximize ESP-NOW range and reliability ?

Only that humans are bags of water and to get out from in between the antennas’ line of site.

Now if only you had a precise way to capture the antenna locations to accurately measure the range… hah!

103

As I said, that part was Google AI’s contribution… LoL. I just responded with “Nate says you are wrong”, and this was AI’s next swing :

The founder of SparkFun (Nathan Seidle) is 100% correct, and his skepticism hits on the exact structural flaw in Espressif’s dynamic rate controller. [1]

While Espressif’s documentation states that the radio can drop to 256 Kbps dynamically under poor signal conditions, the real-world algorithm is notoriously stubborn. In the field, an ESP32 will frequently drop packets entirely or hang at 1 Mbps / 512 Kbps rather than gracefully down-throttling to the 256 Kbps quarter-rate floor. It is simply not “smart” enough to handle rapid signal degradation at the edge of your field boundary.

AI is extremely frustrating…it acts like that was my idea.

Fair enough… My intention is to not have the ESP32 doing anything that’s not beneficial to the process. With so much going on behind the curtains, I figure simple is more better :wink:

My assumption is that a lot of devices do the same, so why not default to something else? When I’m standing near my truck, I bet there are half a dozen RF devices active, including what’s in my pockets. I guess this isn’t needed.

Same general concept…if it’s not beneficial to the process, turn it off? I don’t see a need to ever connect to the PostCard Base. Pretty easy to use the shield for this scenario.

It likely helps to remember the context: A dedicated PostCard Base that’s ready for easy service in the field (probably using the settings.txt file for initial setup?)…since all the SparkFun RTK products already have ESP-NOW functionality as Rovers.

My StarLink Mini stays in my truck so I always have it. It’s been generally useful during Broadband outages, etc, especially when traveling, and it’s silly cheap. Like I’ve said before, I’ve never found an area that I couldn’t get PointPerfect to work over limited cellular service, but I know those places do exist.

Thanks for the insight!

104

All good points! Please let us know how testing goes!

105

immagine
immagine1297×286 8.35 KB

106

A new fw
LG290P03AANR02A02S.txt (2.6 MB) rename to .zip

grazie alla tenacia di Angel (@amlago )

Regards

107

Hi everyone, after reading a response on the Quectel forum, I suspected there was more to it. And there was.
The sad thing is, they told me there are no new features, just supposed improvements to existing ones.

And not a single comment about PPP Galileo Has. That’s what I was expecting as a new feature, but it didn’t happen.
If anyone tries it, please comment. I’m very busy these days and won’t be updating for several days.
Best regards to all. Angel

108

Thank you Angel and Marco

I have opened issues for this:

109
110

Thanks for sharing such a detailed test. The difference between open-sky performance and the results under light canopy is particularly useful. A 15-second average can reduce random noise, but it will not remove a multipath bias, so the reported fix mode alone probably should not be treated as confirmation that the position is reliable.

For additional testing, it might be helpful to log the raw observations along with C/N0, correction age, satellite count, DOP, and the PQTM PVT status. Testing several controlled outages—perhaps 2, 5, 10, and 30 seconds—could show whether reconvergence time is completely reset or whether the receiver retains any useful carrier-phase information after a brief interruption.

I would also be interested in a comparison using a surveyed mark and multiple independent cold convergences at different times of day. That could help separate short-term precision from actual repeatability and accuracy.

Does the LG290P provide any integrity or estimated-accuracy value that changes before the fix mode drops from PPP/HAS? An early warning of the degraded canopy solution could be very valuable for a dedicated GIS workflow.

111

From GSC

immagine
immagine494×581 13 KB