Why does my voltage sag under throttle?
Every LiPo sags. That is Ohm's law, not a fault. The only question worth asking is whether yours sags more than it should — and here is how to answer it.
The symptom
You punch the throttle and the voltage on your OSD falls off a cliff. Maybe it triggers the battery warning. Maybe the OSD flashes at you halfway through what felt like a perfectly normal flight. Maybe, in the worst case, the aircraft browns out and reboots in the air.
Before anything else, understand what you are looking at, because most of what gets called a "sag problem" is not a problem at all.
Every LiPo sags under load. This is not a fault; it is Ohm's law.
V_under_load = V_open_circuit − I × R_internal
The pack has an internal resistance. Current flows through it. The voltage you measure at the terminals is the pack's true, unloaded voltage minus whatever that resistance eats. Pull 100 A through 10 milliohms of pack and you lose a volt, right there, before the ESCs see a thing. There is no battery chemistry, no brand, no C-rating that escapes this.
So the question is never "does it sag." It is "does it sag more than it should." Everything below is a way of answering that.
Check this in 60 seconds
- Watch what the voltage does when you let go. Punch, then centre the throttle and hover. Does the number spring back up? A pack that dives to 3.4 V per cell under a punch and recovers to 3.8 V at hover is a healthy pack. A pack that sags to 3.4 V and stays there is not sagging — it is empty.
- Land and measure per-cell voltage with a cell checker, not pack voltage. Pack voltage hides everything. One cell at 3.2 V and five at 3.9 V averages out to a perfectly reassuring number, and you have a dying pack.
- Compare the OSD number to a multimeter on the same battery. If they disagree, the aircraft is fine and the sensor is lying. Fix the vbat scale before you fix anything else.
Loaded voltage and resting voltage are two different measurements. Confuse them and every conclusion below goes wrong.
The numbers you should be flying to
Per cell, for a normal LiPo:
| State | Volts per cell |
|---|---|
| Fully charged | ~4.2 V |
| Nominal | ~3.7 V |
| Where you should be landing, at rest | ~3.5 V |
| A hard punch may briefly pull it here under load — normal on a healthy pack | ~3.0 V |
| Below this at rest, you are damaging the pack | ~2.8 V |
Resting means after the aircraft has been sat on the bench for a couple of minutes. A pack that reads 3.5 V under a hover will read considerably higher once it has recovered, and that is normal. Judging a pack's remaining capacity from a number taken mid-punch is how people land packs at 3.9 V resting and wonder why their flight times are terrible.
The ranked causes
1. The pack is simply old
This is the answer most of the time, and nobody wants it to be. Internal resistance climbs with cycles, with age, with heat, and — most avoidably — with being stored at full charge. A pack left at 4.2 V per cell for a fortnight has aged more than one flown twice. Storage charge exists for a reason.
An old pack does not announce itself. It flies, it just sags harder, gives you less usable capacity, gets hotter, and puffs. The definitive test is internal resistance on a decent charger — most half-serious chargers will report per-cell IR. Measure your packs when they are new, write the numbers down, and measure again every few months. A pack whose IR has doubled is done, whatever it feels like in the air.
2. The C-rating is too low for what the build draws
Extremely common on heavy builds, and the mistake is nearly always in the arithmetic. A pack's continuous current capability is capacity × C. Your build's peak draw at full throttle is what the motors, props and voltage actually pull — which on a 5 kg airframe is a much bigger number than most people estimate, and the manufacturer's C-rating is an optimistic number to begin with.
If your peak draw is anywhere near the pack's rating, it will sag, it will get hot, and it will age fast. Buy pack headroom. It is the cheapest performance you will ever get, and it makes every other problem on this list smaller.
3. Cold weather
A cold pack has dramatically higher internal resistance. This is not a small correction — it is the difference between a pack that flies fine in July and one that trips its low-voltage alarm on the first punch in January. Cold lithium chemistry simply cannot move ions as fast.
Warm your packs. Keep them in an inside pocket or a heated bag until the moment you plug in, fly gently for the first thirty seconds to let the pack self-heat, and stop treating a winter sag as a failing battery. A warm pack is a different pack.
4. A bad connection — the genuine fault
Everything above is physics. This one is a defect, and it is the reason to take sag seriously rather than shrugging at it.
A cold solder joint, a worn or arced XT60/XT90, thin or unnecessarily long battery leads, an undersized power wire, a connector that has been crushed in a crash — every one of these adds resistance in series with the pack, and the voltage you lose across it is invisible on the battery and very visible on your OSD.
The tell is heat at a specific point. Land after a hard flight and put a finger on the connector, on the solder pads, on the leads. Resistance dissipates power as heat. A connector that is warmer than the pack is a connector that is failing. Reflow the joint, replace the connector, shorten the leads.
5. A single failing cell
One cell in the pack has gone high-resistance and drags the whole thing down. It sags harder than its neighbours under load and recovers to a lower resting voltage.
This is why you check per-cell voltages. An imbalanced pack is the tell — if one cell is consistently a tenth of a volt or more below the others at rest, and the gap widens over a flight, the pack has one bad cell and it is not going to get better. Retire it. A pack with one dead cell will keep working right up until the moment it does not, which will be at full throttle, a long way from you.
6. The build is simply overloaded
Over-propped, over-KV'd, or heavier than the power system was chosen for. The aircraft is not being asked for anything unreasonable and is drawing far more current than intended anyway — so of course the pack sags. Hot motors after a mild flight are the corroborating symptom; if you have those too, read why your motors get hot, because you have one fault, not two.
7. Noise being mistaken for sag
A missing or badly placed capacitor on the power leads does not make the pack sag, but it can make the FC's ADC read like it does — the measured voltage gets noisy and the minimum-voltage reading craters. If your sag looks spiky and implausible rather than smooth, and your video is dirty too, suspect the capacitor before the battery.
8. The sensor is lying
The FC's voltage reading is a scaled ADC value, and the scale is a configurable number that is very often wrong out of the box on a new board. If your OSD says 3.2 V per cell and your multimeter says 3.7 V, nothing is sagging. Calibrate the voltage scale against a meter and the problem evaporates.
Why any of this matters operationally
Sag is not an aesthetic complaint. Your battery warning and your low-voltage failsafe both trigger on the loaded voltage, so a sagging pack fires them early — and a hard enough punch on a tired pack can dip the FC below its own brownout threshold and reboot the aircraft in the air.
And here is the cruelty of it: sag is worst at the end of a pack (state of charge is low, so open-circuit voltage is already down) and worst at high throttle (current is high, so I × R is large). Those two conditions arrive together at precisely the moment you are furthest from home, lowest on margin, and most likely to need full throttle to get back. A build that is marginal on paper will fail you exactly there, and nowhere else.
How to know you actually fixed it
Not "it feels better." Sag is a number, so measure it.
- Same pack, same manoeuvre, same throttle, and the minimum loaded cell voltage in the log has come up. Blackbox records voltage. Use it.
- Per-cell voltages after landing are within a few hundredths of each other, and the pack lands at roughly 3.5 V per cell at rest rather than being flown into the ground.
- Nothing on the power path is hot. No warm connector, no warm solder pad, no warm lead.
- It holds in the cold, and it holds at the end of a pack. A fix that only works on a fresh battery on a warm day is not a fix.
Simulating it first
Sag is one of the few faults you can genuinely predict before you own the hardware, because every term in it is a number you already have: the pack's capacity and internal resistance, the current your motors and props draw at a given voltage, and the state of charge. The equation is not a mystery. The arithmetic is just tedious, and almost nobody does it.
A simulator that models the discharge curve, the internal resistance and a real motor and ESC current draw will do that arithmetic for you across an entire flight — including the awkward case of a full-throttle demand at 20% remaining, which is the one that actually bites. That will not replace measuring your real packs, and it will not tell you that your XT90 has a cold joint. But it will tell you whether the build you are about to fly has the electrical headroom to survive its own worst moment, which is a much better thing to learn on a screen than at the far end of a field.