# EKF2 Altitude Offset

**URL:** <https://discuss.px4.io/t/ekf2-altitude-offset/49379>\
**Category:** Uncategorized\
**Created:** [August 18, 2026, 4:31am UTC](https://discuss.px4.io/t/ekf2-altitude-offset/49379 "2026-08-18T04:31:42Z")\
**Posts on this page:** 7\
**Page:** 1

<div class="post-metadata">

**Author:** ![Ananya](https://discuss.px4.io/letter_avatar_proxy/v4/letter/a/50afbb/32.png) [@Ananya](https://discuss.px4.io/u/Ananya)\
**Post date:** [August 18, 2026, 4:31am UTC](https://discuss.px4.io/t/ekf2-altitude-offset/49379/1 "2026-08-18T04:31:43Z")

</div>

# Vehicle Setup

- **Autopilot:** Pixhawk 6C mini

- **PX4 Firmware Version:** v1.17.1

- **Frame:** Multirotor (Quadcopter)

- **Companion Computer / Interface:** Companion computer [Vicharak Axon] running MAVSDK Python (Offboard control via TELEM2)

- **Height Sensors:** Internal Barometer, GPS [CUAV Neo 3 Pro]

- **Flight Log Link:** [https://logs.px4.io/plot\_app?log=be8dbf81-68c4-48e0-b559-cbadcf979396](https://logs.px4.io/plot_app?log=be8dbf81-68c4-48e0-b559-cbadcf979396)

* * *

### Description of the Issue

I am encountering an altitude estimation discrepancy and stability issue.

#### 1. Static / Takeoff Altitude Bias

- Takeoff is commanded to **2 m** relative altitude.

- The autopilot/telemetry reports that it has reached **2 m** , but physically the drone is only hovering higher or lower above the ground.

- The reported EKF2 altitude stays clamped near **~2 m** while the actual drone is much lower or higher.

#### 2. Inconsistent / Erroneous Altitude Display in QGroundControl

- At times, QGC shows the drone as **never reaching the commanded target altitude** , even though the vehicle has physically stabilized there.

- At other times, the altitude value displayed in QGC goes **negative** , despite the drone being airborne and well above the ground.

* * *

I have tried both using GPS and Barometer as the primary altitude source by configuring the following parameters, but it does not seem to work

### Parameters

- `EKF2_HGT_REF = 0 / 1` — Barometer / GPS

- `EKF2_BARO_CTRL = 1` — Barometer enabled

- `EKF2_GPS_CTRL = 5 / 7` — 2D / 3D position

- `EKF2_ABL_LIM = 0.4` — Baro bias estimation limit

- `EKF2_BARO_NOISE = 1.5 / 3.5`

* * *

Any insights or recommendations on parameter adjustments or hardware isolation best practices would be greatly appreciated!

---

<div class="post-metadata">

**Author:** ![gfLv\_amov](https://discuss.px4.io/user_avatar/discuss.px4.io/gflv_amov/32/21809_2.png) [@gfLv\_amov](https://discuss.px4.io/u/gfLv_amov)\
**Post date:** [August 18, 2026, 9:06am UTC](https://discuss.px4.io/t/ekf2-altitude-offset/49379/2 "2026-08-18T09:06:09Z")

</div>

Are there any other high-quality sources that can be utilized, such as external vision, RTK, or altitude radar?

---

<div class="post-metadata">

**Author:** ![Ananya](https://discuss.px4.io/letter_avatar_proxy/v4/letter/a/50afbb/32.png) [@Ananya](https://discuss.px4.io/u/Ananya)\
**Post date:** [August 18, 2026, 2:16pm UTC](https://discuss.px4.io/t/ekf2-altitude-offset/49379/3 "2026-08-18T14:16:34Z")

</div>

No such sensors can be used. Is there any way to make it efficient with only gps?

---

<div class="post-metadata">

**Author:** ![gfLv\_amov](https://discuss.px4.io/user_avatar/discuss.px4.io/gflv_amov/32/21809_2.png) [@gfLv\_amov](https://discuss.px4.io/u/gfLv_amov)\
**Post date:** [August 19, 2026, 3:09am UTC](https://discuss.px4.io/t/ekf2-altitude-offset/49379/4 "2026-08-19T03:09:33Z")

</div>

Could you please provide me with a log file? I would like to view the original data related to the barometer and GPS. If the original data is drifting, it is unlikely that a stable effect can be achieved by adjusting the EKF.

---

<div class="post-metadata">

**Author:** ![Ananya](https://discuss.px4.io/letter_avatar_proxy/v4/letter/a/50afbb/32.png) [@Ananya](https://discuss.px4.io/u/Ananya)\
**Post date:** [August 19, 2026, 4:07am UTC](https://discuss.px4.io/t/ekf2-altitude-offset/49379/5 "2026-08-19T04:07:17Z")

</div>

> **[Flight Review](https://logs.px4.io/plot_app?log=4de95bdc-7e12-4c5a-afcf-5800a03d79e6)**

[https://logs.px4.io/plot\_app?log=25262c03-adbb-47ce-9b3a-7e60ee74a9d6](https://logs.px4.io/plot_app?log=25262c03-adbb-47ce-9b3a-7e60ee74a9d6)

---

<div class="post-metadata">

**Author:** ![gfLv\_amov](https://discuss.px4.io/user_avatar/discuss.px4.io/gflv_amov/32/21809_2.png) [@gfLv\_amov](https://discuss.px4.io/u/gfLv_amov)\
**Post date:** [August 19, 2026, 6:11am UTC](https://discuss.px4.io/t/ekf2-altitude-offset/49379/6 "2026-08-19T06:11:13Z")

</div>

Can the log file be regarded as an attachment so that I can download it?

---

<div class="post-metadata">

**Author:** ![nickfil](https://discuss.px4.io/user_avatar/discuss.px4.io/nickfil/32/21538_2.png) [@nickfil](https://discuss.px4.io/u/nickfil)\
**Post date:** [November 5, 2026, 9:04am UTC](https://discuss.px4.io/t/ekf2-altitude-offset/49379/7 "2026-11-05T09:04:04Z")

</div>

> [@gfLv\_amov](#):
>
> lease provide me with a log file? I would like to view the original data related to the barometer and GPS. If the original data is drifting, it is unlikely that a stable effect can be achieved by adjusting the EKF.

Hi @gfLv_amov. I am also interested in the issue and have a very similar setup.

I am afraid that the .ulg file cannot be loaded. But if you open the px4 log analyzer liink, you should be able to download the log from there
