My Gaming Laptop Called Its Own Charger "Non-Original" and Throttled Everything — an EC Reset Fixed It
HP gaming laptop flagged its original charger and throttled CPU + GPU. Measuring it with PowerShell and nvidia-smi, and the EC reset that fixed it.
My laptop is an HP Victus-class gaming machine: 13th-gen Intel H-series CPU, a laptop RTX card, and the chunky 200 W HP power brick with the round barrel plug. Original brick, came in the box. One evening Windows pops a notification that the charger is not an original HP adapter, and the whole machine starts stuttering. Mouse fine, everything else feels like wading through mud.
Nothing had changed. Same brick, same outlet, worked fine that morning. The only thing that wasn't original was the mains cable going into the brick (the kettle-style lead), and that one is irrelevant for what the laptop sees.
TL;DR: HP gaming laptops identify the power adapter through the thin center pin in the barrel plug. If that ID handshake fails, the firmware assumes a weak third-party charger and hard-limits CPU and GPU power. In my case the brick was fine and the embedded controller (EC) was just confused. Shut down, unplug the charger, hold the power button for 30 seconds, boot. Warning gone, full power back. Measure before and after so you actually know.
Why a laptop cares who made the charger
The barrel plug on these HP bricks isn't just plus and minus. There's a center pin that carries an ID signal, so the laptop can ask the adapter "how many watts can you deliver?". A 200 W gaming laptop pulling 200 W through a 65 W office charger would be bad, so if the answer doesn't come back, the firmware plays it safe: it assumes the worst, shows the "non-original" warning, and clamps power hard.
That's sensible behaviour with a bad charger. With a good charger it just means your gaming laptop now runs like a netbook. The question is whether the problem is physical (pin, jack, cable) or just the EC holding onto a bad state.
Step 1: measure, don't guess
"Feels slow" isn't a diagnosis. On Windows, the counter I care about is % Processor Performance. It's the current performance relative to the base clock: 100 % means base frequency, above 100 % means turbo.
Get-Counter '\Processor Information(_Total)\% Processor Performance' -SampleInterval 1 -MaxSamples 10
If you want MHz instead of a percentage, multiply by the base frequency:
$base = (Get-Counter '\Processor Information(_Total)\Processor Frequency').CounterSamples[0].CookedValue
$perf = (Get-Counter '\Processor Information(_Total)\% Processor Performance').CounterSamples[0].CookedValue
"{0:N0} MHz" -f ($base * $perf / 100)
Idle numbers don't tell you much because the CPU is supposed to clock down. You need load. I don't install a stress tool for this, I just burn a few cores with PowerShell jobs:
# 4 busy-loop jobs = 4 cores pinned
1..4 | ForEach-Object { Start-Job { while ($true) {} } } | Out-Null
Get-Counter '\Processor Information(_Total)\% Processor Performance' -SampleInterval 2 -MaxSamples 10
# clean up — don't forget this, they run forever
Get-Job | Stop-Job; Get-Job | Remove-Job
What I saw while the warning was active:
- right after the warning:
% Processor Performancestuck at 22 %, roughly half a GHz. No wonder it stuttered. - after unplugging and re-plugging the brick: idle 22 % (which is normal at idle), but under load only ~110–128 %, where this CPU normally sits at 180–200 %.
So re-plugging helped a bit, but the firmware was still clamping. The warning stayed too.
Step 2: check the GPU as well
It wasn't just the CPU. I fired up a game (Apex) and watched the GPU with nvidia-smi, which ships with the NVIDIA driver:
nvidia-smi --query-gpu=utilization.gpu,power.draw,power.limit,clocks.gr,pstate,temperature.gpu --format=csv -l 2
And to see why it's slow, the performance section shows the active clock limiters:
nvidia-smi -q -d PERFORMANCE
Throttled state:
- GPU at 100 % utilization, but only 22–26 W power draw against a 60 W limit
- P4 performance state, about 1.9 GHz
SW Power Cap: Active
The game ran "fine", but it felt off. A GPU at full utilization that only pulls a third of its power limit isn't busy, it's being held back. SW Power Cap means software (driver/firmware) decided on a lower power budget, which matches the "unknown adapter, play it safe" theory.
Gotcha: re-plugging is not a reset
The obvious first move is to pull the plug and push it back in. That re-runs the handshake, sure, but the embedded controller doesn't necessarily forget its state. For me it got the CPU off the 22 % floor and nothing more. The EC is a tiny microcontroller that runs independently of Windows. It handles power, charging, fans and keyboard, and it keeps running while the laptop is "off" as long as it has battery or AC. Restarting Windows doesn't touch it.
Step 3: the EC / hard reset
This is the HP hard reset for laptops with a built-in battery. You can't pull the battery on these anymore, so you drain the EC instead:
- Shut down Windows completely. Not sleep, not hibernate. If Fast Startup is enabled, hold Shift while clicking Shut down, so it does a full shutdown.
- Unplug the power adapter. Also unplug USB devices and the dock if you have one.
- Press and hold the power button for 30 seconds. (HP's own instructions say about 15 seconds; I did 30, it doesn't hurt.)
- Plug the adapter back in and boot normally.
While it's unplugged anyway, look at the plug. Is the center pin straight and clean? Any burn marks or gunk in the laptop's jack? Is the cable kinked or frayed at the strain relief near the plug? Those are the spots where the ID line breaks first. On mine everything was clean.
After the reset: warning gone.
Step 4: measure again
Same tests as before. A warning that disappears doesn't count as fixed until the numbers say so.
CPU, plan "Balanced":
| Load | Before (throttled) | After EC reset |
|---|---|---|
| 4 burn jobs | ~110–128 % | 130–167 % |
| 16 burn jobs (all threads) | — | ~130–140 % |
The all-core number being lower than the 4-job one is normal. With every core busy the CPU hits its package power limit and spreads the budget over all cores, so per-core clocks drop. That's how the turbo budget is supposed to work, not throttling.
GPU, same game, same settings:
| Before | After | |
|---|---|---|
| Power draw | 22–26 W | 45–58 W |
| Clock | ~1.9 GHz | 2.4–2.5 GHz |
| P-state | P4 | P0/P2 |
| Temp | — | 70–77 °C |
| SW Power Cap | active all the time | only briefly, right at the 60 W limit |
GPU utilization afterwards was only around 50 %, which looked weird at first. It wasn't a problem though: the game was hitting its FPS cap or the CPU limit, so the GPU simply had nothing more to do. Power cap only touching the limit occasionally is what you want to see.
I ran all of that without HP's performance mode in the vendor tool. If you want more, switching the fan/performance profile is the next knob, but that's tuning, not fixing.
If the reset doesn't fix it
Then it's probably real hardware. Work through it in this order:
- Plug and pin. Bent or dirty center pin, wobbly jack on the laptop side. Gently clean with compressed air, don't stab metal into it.
- Cable. The DC cable between brick and plug is the part that gets bent every day. A break in the thin ID wire gives exactly this symptom while power still flows fine.
- Swap test. If you can borrow a known-good original HP adapter with the same wattage, that answers it in two minutes.
- Replace. Get an original adapter with the right wattage. If the laptop is still under warranty, go through the vendor, chargers are usually covered.
Don't buy a cheap "compatible" 200 W brick to solve a "non-original charger" warning. You'll most likely get the same warning, plus a power supply you don't trust.
Side find: the monitor was on the wrong GPU
While I had the tools open: my ultrawide (3440x1440) was running at 59 Hz instead of its max 75, because it was driven by the Intel iGPU and not the NVIDIA card. Which port the monitor is plugged into decides which GPU drives it on many gaming laptops, and the Windows display settings show you the refresh rate under Advanced display. Not related to the charger issue, but worth checking if you're measuring performance anyway.
What I'd do next time
- First thing:
Get-Counterunder load. 180 % or more under a few-core load on an H-series chip means the charger detection is fine. Stuck far below that, plus the warning, means it isn't. nvidia-smi -q -d PERFORMANCEwhile gaming.SW Power Cap: Activewith low watts is the GPU-side tell.- EC reset before buying anything. It's free and takes a minute.
I keep a small set of these PowerShell checks around anyway. If you want more of them, my PowerShell sysadmin cheatsheet has the everyday ones, and how I use AI for PowerShell scripts covers how I build throwaway diagnostics like the burn-job loop above. And if you want a check like this to run on a schedule, the Task Scheduler + PowerShell post shows the setup.
