Propwash: what it is and what helps

You chop the throttle after a flip or a fast turn, punch back in, and the quad shakes? That is propwash: it falls into its own downwash. Here is how to recognize propwash, in flight and in your blackbox log, and which levers the Betaflight docs name against it, each with its direction and its trade-off. Checked against the docs and source code of Betaflight 2026.6.2. Values for PIDs and filters are left out on purpose.

Short answer

  • What: after a throttle chop the quad sinks into the turbulent air its props just pushed down. Thrust drops unevenly and it wobbles.
  • When: after flips, dives and hard turns with zero throttle, worst in vertical descents.
  • Biggest lever: according to the docs, flying style. On the technical side, what counts most is how fast motor and prop build up thrust; filters are only a small part of it.
  • In Betaflight: more D, Dynamic D, less filter delay with the RPM filter, Dynamic Idle. Almost everything has a price, usually hot motors. The overview shows the direction and trade-off of each lever.
  • Rates: are not the cause. They set what you ask for, not how well the quad follows.

What propwash is

The most detailed explanation is in the Betaflight 4.0 Tuning Notes: “Propwash happens when, after throttling down and turning very fast, the pilot then throttles up, and the quad falls backwards/downwards into dirty air. Reversed airflow generates tip vortexes that greatly reduce thrust.” The same page explains why it gets so rough: motors that almost stop at zero throttle spin up much more slowly than the others. Thrust builds up unevenly, the axes affect each other, and with such uneven motor responses the PID controller can hardly keep up. The page dates from Betaflight 4.0; some parameter names it uses no longer exist, so this guide uses the names from 2026.6.2.

The docs page on Dynamic Idle describes the same situation from the motors’ side: in a flat drop or after a quick 180 degree reversal, air pushes into the props from below. That slows them down so much that they spin too slowly to start quickly, stop, or even spin backwards. The PID Tuning Guide in the docs, one freestyle pilot’s notes, puts it more briefly: propwash happens when changing speed, and the turbulence makes the quad unstable while steering.

How to recognize propwash

In flight, the quad wobbles or shakes shortly after you get back on the throttle after a fast rotation, worst in vertical descents. The 4.0 Tuning Notes describe the shaking like this: “During propwash, the quad shakes at relatively low frequencies that are easily seen in video, eg 20-30hz. These shakes cannot be filtered out.” So it is a visible wobble, not a fine buzz like the noise from motors and props.

Not every wobble is propwash. These cases look similar, but according to the docs they have other causes:

  • Bounce-back at the end of a flip or roll: according to the PID Tuning Guide, too little D, too much P or even too little P.
  • Oscillation only at high throttle: for this the same guide names TPA, see below.
  • Wobble right when you chop the throttle to zero: according to the PID Tuning tab docs, too much I can cause it, especially with too little P; so can too much anti-gravity.
  • Shaking without stick input, hot motors: more likely mechanics or noise. The guide Twitchy quad: causes and fixes walks through those causes.

Propwash in the blackbox log

To evaluate propwash in a log, the PID Tuning Guide names one yardstick: how gyro follows setpoint. Setpoint is the rotation rate you ask for with the stick, gyro the one the quad actually flies. According to the same page, “propwash also generates PID error”, a gap between the two. And according to the 4.0 Tuning Notes, propwash typically happens at low stick inputs. This is how you look for it in the Blackbox Viewer tab of the Betaflight App:

  1. Find the spot: set the timeline at the bottom to “Average motor throttle” and look for places where throttle drops hard and then rises again. The RC Command [throttle] trace shows your throttle directly. How the timeline and zoom work is in How to use Blackbox Explorer.
  2. Overlay the traces: with “Graph setup”, put Gyro [roll] and Setpoint [roll] in one graph, and the same for pitch. If gyro swings back and forth after you get back on the throttle while setpoint stays almost still, the quad is not following you: that is the picture the docs describe as propwash.
  3. The gap as its own trace: the viewer computes PID Error [roll] itself, setpoint minus gyro (flightlog.js).
  4. Look at the D-term: PID D [roll] shows what the D-term is doing. The PID Tuning Guide says: “D-term should be soft (not noisy), otherwise you cant handle Propwash”.
  5. With video: since propwash is visible in video, it helps to load the flight video into the viewer and line it up with the log, see More tools in the Explorer guide.

Which tools besides the viewer analyze setpoint and gyro, for example as a step response, is in Blackbox analysis: which tool for what. How to record a log in the first place: Blackbox logging: setup and downloading your log.

What helps: the overview

The table sums up what the docs say about each lever: what it does, which way they point against propwash and what it costs. It also gives the parameter name from settings.c and the state out of the box in Betaflight 2026.6.2. Numbers are left out on purpose; what fits depends on your quad. Change one thing at a time and feel the motor temperature after each short test flight; the docs advise that after every change to D and to the filters.

What helps against propwash, according to the Betaflight docs; settings and defaults in Betaflight 2026.6.2
LeverWhat it doesDirection per the docsTrade-offIn Betaflight 2026.6.2
Flying and build
Flying styleSource: 4.0 Tuning NotesKeeps the props in clean air and the motors spinningAvoid cutting throttle to zero, fly smooth arcs instead of hard 180s; the biggest lever according to the docsnone, takes practiceno setting
Props and powerSource: 4.0 Tuning NotesDecide how fast the quad can change thrustlight, easily spun props, don’t over-prop; more cells; heavier center mass, battery on topnew parts, a rebuildno setting
Clean buildSource: Dynamic DStraight props, good bearings, a solid frame: less noise, so less filtering neededcheck first; per the docs a noisy build needs more filtering and less D boosttime and spare partsno setting
Props outSource: Motors TabThe front props spin away from the camera instead of towards itthe docs say it helps against propwash on smaller craftevery motor has to spin the other way; debris goes into the frame instead of the camerayaw_motors_reversed, Motors tab: “Motor direction is reversed”; off by default
In Betaflight
D-term (“Damping” slider)Source: PID Tuning GuideDamps any motion, including outside disturbances; per the docs the only PID term fast enough for propwashmore D helps, but only as much as needed, in small stepshot motors, amplified noise, duller stick responsed_roll, d_pitch; slider simplified_d_gain; sliders on by default
Dynamic D (“Dynamic Damping” slider)Source: Dynamic DRaises D towards D Max in fast moves and propwash; only the gyro part (Gain) reacts to propwash, not the stick part (Advance)for propwash, D Max above D and the Gain part switched on; higher only on clean builds, lower on noisy onesD jumps up with every vibration, hot motors on noisy buildsd_max_roll, d_max_pitch, d_max_gain, d_max_advance; by default D Max on, Gain off, only Advance on
P (“Tracking” slider)Source: PID Tuning TabHow firmly the quad follows the setpoint; with too little it bobbles and goes off course in propwashno simple more or less: the docs speak of a sweet spot and also suggest lowering P and raising Dtoo much: oscillation, hot motorsp_roll, p_pitch; slider simplified_pi_gain
Gyro and D-term filtersSource: PID Tuning TabRemove noise but delay the signal; more delay means more propwashless filtering (sliders to the right) can help, only on clean builds; never switch filters off entirelymore noise in P and D, hot or burnt-out motors, flyawayssliders simplified_gyro_filter_multiplier, simplified_dterm_filter_multiplier; on by default
RPM filterSource: DShot RPM FilteringRemoves motor noise exactly at each motor’s speed; on clean quads less of the other filtering is needed afterwardsswitch it on; since 4.3 the docs report less propwash at low throttlethe ESC must support bidirectional DShot; without valid telemetry the quad won’t armdshot_bidir (Motors tab: “Bidirectional DShot”), rpm_filter_harmonics; DShot telemetry off by default in the firmware, the board configuration may differ
Dynamic IdleSource: Dynamic IdleKeeps every motor above a minimum rpm, even at zero throttle; the motors spin up again fasterswitch it on; per the docs a high minimum rpm helps avoid propwashflips turn into small circles, shorter inverted hang time; needs bidirectional DShotdyn_idle_min_rpm, PID Tuning tab: “Dynamic Idle Value”; off by default
TPASource: PID Tuning GuideLowers D (in PD mode also P) above a throttle point; meant for oscillation at high throttleonly a brief note on propwash in the PID Tuning Guide, without explanationless D, so less damping at high throttletpa_mode, tpa_rate, tpa_breakpoint; on by default, D only
I-term and anti-gravitySource: 4.3 Tuning NotesI holds the attitude; anti-gravity briefly boosts I and P on fast throttle changesnot named as a fix where the docs discuss propwash; keep anti-gravity close to defaulttoo much of either: wobble on throttle chopsslider simplified_i_gain (“Drift - Wobble”), anti_gravity_gain; anti-gravity on by default
Thrust LinearizationSource: 4.0 Tuning NotesMore control authority at low throttle, no effect at high throttlecan help on some quads, according to the 4.0 Tuning Noteswatch for wobbles at idle; rescaled in 2026.6, old values behave differentlythrust_linear; off by default

First: flying style and build

The 4.0 Tuning Notes are clear: “Be aware that the root cause of propwash is flight style. Of the technical causes, the biggest factor, by far, is the delay arising from motor/prop thrust generation delay. Only 10% of the problem is filter delay related. Even if there was zero filter delay, there would still be propwash.” In the air, according to the same page, it helps to:

  • keep the props moving forward in clean air,
  • avoid sharp 180 degree reversals and fly smooth arcs,
  • keep some throttle on at all times and never cut it to zero, except during flips, inverted drops and the like.

For the build, the same page says the quads that cope best with propwash are those that can change thrust quickly: more cells, light, easily spun props that are not too big for the motors, plus a heavier center mass and the battery on top. The Motors tab docs also mention “props out”: if the front props spin away from the camera instead of towards it, “It also helps against prop wash on smaller crafts”. For that, every motor has to spin the other way, and debris goes into the frame instead of the camera.

A clean build matters just as much. Bent props, worn bearings or a resonating frame create noise; then the quad needs more filtering, and more filtering means more delay. According to the PID Tuning tab docs, the gyro filter slider belongs further left with such problems, meaning more filtering. How to check props, motors and bearings is in Twitchy quad: causes and fixes.

How well do your rates fit your stick? The free check measures it on your blackbox log. No account, the file stays with you.

Check my log for free

No log yet? See a real example · Turn on logging, with pictures

For your next flying day
  • Betaflight App, Blackbox tab: logging device onboard flash or SD card, logging rate 1 kHz
  • RC Commands, Setpoint and Gyro on, then “Save and reboot”
  • Fly 3 to 5 packs of normal acro
  • Then get the log off the quad and drop it into the free check

D-term and Dynamic D

D damps any motion of the quad, including outside disturbances such as propwash or gusts of wind. According to the PID Tuning Guide it is “The only PID-term that can respond quickly enough to handle fast-changing situations like propwash oscillation”. With too little D, the same page says, you get the worst propwash in vertical descents. So the direction is: more D helps. The price: D amplifies high-frequency noise, the motors get hot and can burn out, and the quad responds more sluggishly to the stick. The page advises raising D in small steps, only as far as it reduces bounce-back or propwash after a descent, and doing a short flight and checking the motors after each step.

With the sliders, which are on by default in 2026.6.2 (pid.c), you set D in the PID Tuning tab with “Damping”. The tab docs describe the trade: high D dampens stick response and may make motors hot, but improves propwash; low D responds faster but copes worse with propwash. Too little “Tracking” (P and I), the same page says, makes the quad bobble and go off course in propwash. Under the heading “Propwash”, the PID Tuning Guide names three ways: change the ratio between P and D, lower P and raise D, or use feedforward, the last one without explanation. The 4.0 Tuning Notes speak of a sweet spot for P, neither too low nor too high, that you have to find by experimenting.

Dynamic D raises D towards D Max only when it is needed. According to the docs there are two triggers: the gyro part (d_max_gain, in the App “Gain” under “Dynamic Damping / D Max settings”) reacts to actual movement of the quad, including propwash oscillations. The stick part (d_max_advance) reacts to fast stick moves; the PID Tuning tab docs state explicitly: “Advance does not respond to gyro or propwash.”

Important for 2026.6.2: in the source code the gyro part is switched off by default and only the stick part is on (pid.c, pid_init.c). Out of the box, Dynamic D therefore raises D on fast stick moves, not on the shaking in propwash. Commit b6bf0f5 set it this way in July 2026. The Dynamic D docs page still lists the 2025.12 defaults, and the 2026.6 release notes say the stick part is off by default; that was the state after commit d340f79, which b6bf0f5 reversed three days later. If you want Dynamic D to work against propwash, you have to switch the gyro part on yourself. The docs advise higher values only on clean freestyle builds and lower ones on noisy builds, because otherwise D shoots up with every vibration. With debug_mode = D_MAX, Betaflight logs how much of the boost comes from the gyro part and how much from the stick part (pid.c).

Filters and the RPM filter

Filters remove noise from the gyro signal and from the D-term. But every filter also delays the signal, and according to the 4.0 Tuning Notes, the greater the delay, the more the quad tends to oscillate in propwash. So less filtering can help, but only if the quad runs clean enough. The PID Tuning tab docs warn about the “Gyro Filter Multiplier”: moving it right for better propwash lets more noise into P and D and risks flyaways or burnt-out motors. Moving the “D Term Filter Multiplier” right is not recommended; on clean builds it may improve propwash, but it can also cause motor grinding on arming, sudden resonances in flight and serious motor overheating. Switching whole filters off is no shortcut, say the 4.1 Tuning Notes: “We do not recommend turning entire filters off to reduce filter delay. We strongly recommend using the sliders and keeping all four filters on at all times.”

The RPM filter places narrow filters exactly on the speed of each motor. According to the docs, motors run cooler with it, bent props are tolerated better, and on clean quads the delay of the remaining filters can usually be reduced afterwards. Since Betaflight 4.3 it fades out at low motor speed, which according to the 4.3 Tuning Notes means less propwash at low throttle. It needs bidirectional DShot: the ESC must support it, and you switch it on in the Motors tab with “Bidirectional DShot” (dshot_bidir). In the firmware it is off by default (motor.c); a board configuration can set it differently. If an ESC does not deliver valid telemetry, the quad won’t arm.

Dynamic Idle

Without Dynamic Idle, the motors get a fixed minimum at zero throttle. When air comes from below, they can still slow down so much that they don’t spin up again fast enough. Dynamic Idle measures each motor’s speed over bidirectional DShot and keeps the slowest one above a minimum rpm. According to the docs, the motors then start up quickly: the quad responds better at zero throttle, wobbles less across the other axes when you stop a flip quickly, and stays calmer in flat drops. The 4.3 Tuning Notes add: “In both cases, high minimum RPM helps avoid propwash since you tend to always move through clean air; this make a lot of sense when flying freestyle or fast racing.”

Dynamic Idle is off by default, in the docs and in the 2026.6.2 source code (pid.c). You switch it on in the PID Tuning tab with “Dynamic Idle Value” (dyn_idle_min_rpm), only with working bidirectional DShot and, according to the docs, not together with 3D mode. The price, according to the 4.3 Tuning Notes: with a high minimum rpm even the slowest motors provide thrust in flips, flips become small circles and turns small arcs. If you want the longest possible inverted hang time, the same page says to use the lowest value that avoids desyncs.

TPA, I-term and anti-gravity

TPA (Throttle PID Attenuation) lowers the D-term above a throttle point, in PD mode also P. According to the PID Tuning Guide it is meant for fast oscillations at high throttle. In 2026.6.2 it is on by default and acts on D only (tpa_mode, tpa_rate, tpa_breakpoint; pid.c). On propwash, the PID Tuning Guide only has a note: “Better propwash (before move D gain up, move TPA up)”. The page does not explain what it means.

I-term and anti-gravity do not appear as fixes where the docs discuss propwash. They still matter for telling things apart: according to the PID Tuning tab docs, too much I, especially with too little P, can cause a wobble when you chop the throttle to zero. Anti-gravity briefly boosts I, and since Betaflight 4.3 also P, on fast throttle changes. The 4.3 Tuning Notes warn that very high values risk significant P and I wobble precisely during throttle chops, and advise staying close to the defaults. If the nose bobs on throttle chops, the docs point to the “Pitch Tracking” slider and to anti-gravity.

Thrust Linearization (thrust_linear, off by default) gives more control authority at low throttle and has no effect at high throttle. The 4.0 Tuning Notes name it as an option for some quads and advise watching for wobbles at idle. Since 2026.6 its curve has been re-derived, according to the release notes, so old values behave differently.

What about rates?

Rates are not the cause of propwash. They set which rotation rate you ask for at which stick position, that is the setpoint. Propwash is something else: the quad does not follow the setpoint because it is sitting in turbulent air. With the stick centred, Actual Rates ask for no rotation at all; in the formula from rc.c, stick zero gives exactly zero degrees per second. If the quad shakes in a descent while you hold the stick still, it does not come from your rates curve. In the docs passages on propwash linked here, rates appear neither as a cause nor as a fix.

What fpvrates measures and what it doesn’t

fpvrates does not measure propwash and sets neither PIDs nor filters, not even as a suggestion. From your blackbox log, fpvrates measures your rates: how much stick travel you really use, where your normal flying ends and how steep your curve gets towards the edge. With the tuning pass you get a rates suggestion from that. If you want to tune PIDs and filters against propwash, you will find tools that analyze blackbox logs in the comparison table in the autotune guide and in Blackbox analysis: which tool for what.

Sources

As of November 13, 2026, checked against Betaflight 2026.6.2 (firmware and App of September 16, 2026, the latest versions on that day):

Last updated: Nov 14, 2026 · checked with Betaflight 2026.6.2