I am using a Quectel L86 GNSS module in a design where the main power (VCC) is switched on/off using a MOSFET, while the backup supply (V_BCKP) is permanently connected to a 3.3 V rail.
The intended behavior is that when VCC is turned off, the module should retain:
- RTC time
- last known position
- ephemeris (valid for hot start)
The main goal is to achieve a hot start (1–2 seconds) when the module is powered again.
How the system works
- VCC is turned on for about 10 seconds so the module can obtain a fix.
- After that, VCC is switched off via MOSFET (module fully powered down).
- V_BCKP remains powered at all times.
- After some time, VCC is turned on again.
However, I am seeing the following behavior:
- RTC survives (correct time is output immediately after boot).
- But ephemeris data are lost, causing a warm/cold start (10–20 seconds) instead of a hot start.
Example logs
After power-up — RMC shows no fix (V)
$GNRMC,221617.189,V,,,,,0.20,199.53,031225,,,N,V*2A
$GNGGA,221617.189,,,,,0,0,,,M,,M,,*57
Over several seconds, it continues like this:
$GNRMC,221618.167,V,,,,,0.26,207.89,031225,,,N,V*20
$GNRMC,221619.165,V,,,,,0.36,231.23,031225,,,N,V*27
$GNRMC,221620.165,V,,,,,0.13,245.25,031225,,,N,V*2F
The module outputs the correct UTC time immediately (RTC is fine), but position fields are empty and fix status is V, which means ephemeris are missing and the receiver needs to download them again.
Expected hot start (which I get only when VCC is never removed)
When VCC is always powered (no power cycling), I get:
$GNRMC,221538.000,A,5107.4687,N,01703.0454,E,...
Fix is obtained in 1–2 seconds, meaning ephemeris are present.
What I already tried
Added backup capacitors on V_BCKP:
- 4.7 µF ceramic capacitor (X5R)
- 0.1 µF (100 nF) placed very close to the V_BCKP pin
This matches Quectel’s hardware design recommendations.
Verified that V_BCKP is always powered
It stays at 3.3 V even when VCC is off.
RTC does not reset
The time reported immediately after start is correct.
Problem summary
Even though V_BCKP is powered and RTC survives, ephemeris are still lost whenever VCC is switched off via MOSFET.
Only when VCC remains continuously powered do I get a true hot start.
This strongly suggests:
- A short voltage drop or glitch on V_BCKP during VCC power-down
- Or current backflow from VCC into V_BCKP
- Or a reset condition affecting the GNSS SRAM even though RTC survives
