Marius Board Setup: Circularity and Advanced Calibration – Battle Beaver Customs

30 DAY MONEY BACK GUARANTEE!

Project Marius Setup & Calibration Guide: Part 2

Welcome to Part 2 of the Project Marius Setup & Calibration Guide! With initial setup out of the way, the next step is to fine-tune the controller so it feels exactly the way you want.

On this page, we’ll cover:

  • Circularity
  • Circularity Error Rate
  • Advanced thumbstick calibration methods

Haven’t done the initial setup yet, or still need help understanding all the options? Start with Part 1: https://battlebeavercustoms.com/pages/marius-calibration-instructions


Circularity, Error Rates, and You


Now that the initial setup and calibration are complete, let’s dig into a hot topic in the community: Circularity.

When you test Circularity, the tool takes measurements of the total range of motion of the thumbstick, then maps that range onto a perfect circle. The Circularity Error Rate (CER) is a measurement of how often the thumbstick range misses the boundary of that circle.

The short version: Circularity is a measurement of how circular your circle is.

For in-house calibrations, Battle Beaver targets a CER range around 6% to 9%, with diagonals slightly extending past the boundary of a perfect circle. We believe this target provides the best balance of input consistency and responsiveness.

Six random OEM controllers pulled off the rack and tested for Circularity without adjustment.

A common instinct is to want the CER to be near 0%. It’s called an Error Rate, after all. Why wouldn’t we want that to be as low as possible? Errors are bad!

Without going too deep into game design explanations, developers build their games with OEM controllers in mind. OEM controllers typically have a CER around 5-10%, slightly overshooting on the diagonals, but even coming directly from Sony, these numbers can vary wildly.

So when a thumbstick range is undershooting or too close to a perfect circle (ultra-low CER), certain directions may not be able to reach as far as the game expects a controller to reach. Depending on the game and how it’s coded, this can lead to sluggish analog movement or even an increase in issues like sprint-canceling.


If the thumbstick range is too square, on the other hand (extremely high CER), it can potentially have a minor impact on the consistency between cardinal directions and diagonal directions. Note, though, that many older games actually expect a square input field, and some players actively seek it out. (It’s common for certain styles of Rocket League play, for example.)

When Battle Beaver ships JDM-5X controllers, we target a Circularity Error Rate around 6% to 9%, with each diagonal slightly extending past a perfect circle. If your controller fits that profile, it should feel normal. If it’s much lower or much higher than that, or especially if one or more corners undershoots the boundary, continue to the next guide for help adjusting it.

Note:
In real-world testing, we’ve noticed little practical difference from 4% up to 13%, though sensitivity to this will obviously vary person-to-person. Like we said at the beginning, every set of sticks will feel different to every set of thumbs.

Testing Circularity Error Rate

To test your Circularity after completing setup or calibration:

This will activate the new Circularity Tester. To run the test, rotate both sticks slowly, a few times in each direction until all 32 sample points fill up. You may need to go over stubborn segments a few extra times.

As the 32 sample points fill up, the Tester will draw the shape of your input field with a green line. It will also calculate your Circularity Error Rate numerically, shown under Sector RMS. Ideally, your CER will show somewhere around 6% to 9%, and the drawn shape will look almost like a rounded square. Diagonals should extend slightly past the boundary of the circle.

If your test shows this, you’re probably good to leave the calibration alone. Below, there are instructions and tips for how to handle low and high CER readings.


Adjusting Circularity Error Rate

At Battle Beaver, we call 0% - sub-5% Low CER, and anything higher than 10% High CER. Adjusting your CER changes based on whether it’s low or high CER. However, CER is not a particularly consistent or scientific measurement, and you should always try your controller ingame first, before starting any lengthy adjustment processes.

If you do dislike how it feels, return to the Circularity Tester to complete a fresh test. With that done, click the ADVANCED button and choose between the Adjusting Low CER or Adjusting High CER instructions below.

Fixing Low Circularity Error Rate

Look at the green shape drawn by the Circularity Tester. Areas that fall inside the main circle are undershooting. Diagonals that match the edge of the circle are also likely undershooting. If
you see either of these, you should modify the calibration to ensure full controller functionality.

Extremely low CER can negatively affect your gameplay. Examples:

  • Certain directions can struggle to reach a maximum value (slow or sluggish movement).
  • Issues like sprint-canceling in Call of Duty can become a more common issue.

Fixing low CER is a three-step process:

  1. Identify Issues
  2. Fix Undershooting Cardinal Directions
  3. Manual Calibration Adjustment

Each of these Circularity Test results came from the same controller with the same calibration on the same thumbstick. All that changed was the way the test was performed.

Fixing High Circularity Error Rate

While the ideal range hovers around 6-9%, in practical usage, we've found that CER up to 13% (and honestly, beyond that, too) still feels fine to many gamers. However, we recognize that some are more sensitive to this than others, and extremely high CER is something we avoid in-house and fix prior to shipping. For Project Marius boards, our QA ceiling is 9%.

If you recalibrate your controller and your unmodified CER is higher than that, we recommend repeating the calibration, focusing on maintaining even (not extreme) pressure against the thumbstick holes as you rotate them.

High CER is extremely uncommon on Project Marius boards with TMR or MIDAS mechs. It may be more common on boards with Alps (OEM) mechs.

Be aware that the physical way you test Circularity can also impact the reported CER. Differences in pressure around the circle can have a major impact, and the longer you test, the more the CER will increase due to these inconsistencies.

Even something as simple as turning the controller around can lead to different results.

If recalibration doesn't help, contact Customer Support for further assistance.

Final Note


In the end, the numbers don’t always reflect the feeling. Chasing perfection on something like Circularity Error Rate before even trying the controller in-game is a common pitfall that prevents people from simply enjoying their new controller. Sometimes, all you need is a few minutes for your brain to adjust to the new equipment.

So before spending a ton of time making minor calibration adjustments, we strongly recommend loading up a game you’ve played a lot. Give yourself a chance to adjust and see how it feels in the context of a full match.

If you still aren’t satisfied, and you’re still struggling with dialing in your thumbstick configuration even after trying everything on the page, go ahead and contact Battle Beaver Customer Support. We’ll try to help you out.

And if you’d rather talk to other gamers and enthusiasts who are deep into calibration experimentation, Marius has his own Discord community filled with them. They’ll be happy to give you advice.

Join them here: https://discord.com/invite/QcCkfbkp3S

Troubleshooting

My LED is flashing green and I can't set up the controller

Update the controller's firmware by visiting https://update.mariusheier.com and
following the on-screen instructions.

Support

For support and additional info:

Marius Heier's Available Platforms

Website
YouTube
Instagram
X / Twitter
Patreon
Discord

Frequently Asked Questions

Why order the JDM-5X from Battle Beaver instead of Marius?

The benefit of purchasing a JDM-5X board from Battle Beaver means receiving a ready-to-play controller.

We remove all of the required equipment and technical soldering experience to make use of the parts you'd get from Marius, and offer our own patented and signature modifications that professional and casual players enjoy to build a uniquely competitive PC-Only controller.

Setup Configuration & Update Links

To setup, recalibrate, remap, and more:

https://setup.mariusheier.com/
For update notifications please use Marius' Official Channels:
Firmware Updates: https://update.mariusheier.com/

How does it connect?

The board requires a wired connection to a Rear I/O port with a USB-C to A cable. Connections to a case USB port or a USB hub are incompatible.

What can this board NOT do?

As this is a custom board, it is important to note that controllers built with this WILL NOT HAVE the following:

  • Console connectivity
  • Bluetooth connection
  • Macros / Keybinding
  • Rumble / Vibration
  • Touchpad Tracking
  • Battery

What are MIDAS Mechs (HE)?

MIDAS Mechs (HE) are the custom 5 pin Hall Effect thumbstick for the PS4 JDM-5X by Marius Heier.

MIDAS stands for Magnetic Integrated Digital Array Sensors, and are only compatible with the PS4 JDM-5X board. Unfortunately the 5 pin connectors are not compatible with 1st Party boards from Sony or Microsoft.

These unique Hall Effect thumbsticks overcome the issue you see in traditional 3 pin HE sticks other 3rd party controllers have by utilizing a fully circular magnet sensor for wider signal coverage, and an additional 2 pins of signal transference for improved accuracy and stability.

Why isn't Battle Beaver using the same Ginful Thumbstick mechanisms as Marius?

All of our modifications were built and developed on the ALPS platform, but after testing the Ginful stick boxes we believe that they have a significant enough difference in the return-to-center feeling that we decided to stick with what we know.

What is a JDM-5X board? What are the best use cases for a controller with a JDM-5X board?

The JDM-5X board is a custom PC only controller board developed by Marius Heier, matching the form factor of a PlayStation 4 board. It's one of the fastest boards available on the market: GPDL testing shows a button latency of 1.43ms at 8KHz, which is faster than an overclocked 8KHz DualSense controller (2.11ms) and is roughly twice as fast as a max-overclocked PlayStation 4 controller (2.89ms).

Click here for Average Latency Comparison by Controller.

The JDM-5X also reaches that 8KHz polling rate without having to install separate overclock drivers like hidusbf. With this board in your hands, you'll never have to doubt whether or not your controller is keeping up with you.

What are the differences between the MIDAS HE 5-pin thumbstick mechs and the two 3-pin options (Magneto TMR and Alps OEM)? Which one is better?

Alps OEM thumbstick mechs are the standard mechs used on PlayStation controllers. They'll act and feel just like a stock controller, because they're stock parts. However, due to how carbon-track potentiometers work, they will almost always develop stick drift over time.

This is an unavoidable limitation of the physical hardware. If you're okay with that risk, nothing will perform or feel more like the controllers you're used to than OEM thumbstick mechs.

If you want to avoid thumbstick drift for as long as possible, though, MIDAS HE 5-pin and Magneto TMR 3-pin both fulfill that role.

They both have upgraded sensors that use magnets to generate thumbstick input values instead of a carbon track that wears down over time. Both offer superior performance compared to OEM, and both are highly drift-resistant.

Where they differ is in the technology:

MIDAS HE 5-pin
MIDAS Mechs use an array of Hall Effect magnetic sensors combined with two extra pins for signal validation, so they're capturing more reliable data from more areas of the magnetic field as the thumbsticks move. Hall Effect technology is older than TMR, but the sensor array setup is fundamentally different than most Hall Effect mechs, which usually only use a single sensor.

Magneto TMR 3-pin
Magneto Mechs use TMR (Tunnel Magnetoresistance) sensors, the highest-precision magnetic single-sensor technology available in a controller. They lack the extra validation pins and sensor array of the MIDAS mechs, but precision is still high and drift is still rare. Common feedback is that they feel closer to OEM handling.

So are MIDAS mechs or Magneto mechs the better magnetic sensor option?
This will fundamentally come down to personal taste. Both magnetic sensors are extremely accurate and highly drift-resistant. TMR is generally better technology than Hall Effect, but the MIDAS Mechs have the extra sensors and pins for a cleaner and more complete signal. Magneto Mechs, on the other hand, can feel closer to an OEM controller, and they perform at a similar level for a lower cost.

Either choice is excellent.

Will the controller come ready to plug and play?

Yes, your controller is ready to go out of the box, set up and calibrated in-house by our technicians before shipping.

Note: We calibrate every controller to match our optimal settings, but those settings may not match your preferences. Customize at https://setup.mariusheier.com.

How does Battle Beaver configure the controllers?

Stick Configuration
Inner Deadzone: 2%
Outer Deadzone: 5%

Raising the Outer Deadzone helps push the Circularity Error Rate to a range closer to OEM, reducing the likelihood of sluggish diagonals. We also fine-tune the thumbstick values in Advanced mode to make sure each cardinal direction (up, down, left, right) easily reaches a maximum value. Exact fine-tune numbers vary.

Trigger Configuration
Mode: Analog for all trigger types. (See "Should I choose Analog or Digital for
triggers?" for more explanation.)

Rest: When triggers are at rest, we make sure the Raw value under the trigger meter is slightly higher than the Rest value. Exact numbers vary.

Press: This value controls how early or late the triggers reach their maximum value. Exact numbers vary.

Hair Trigger: 100% - This slider doesn't behave as expected, so we set it to 100% to effectively disable it.

LED Control
Color: #0099FF for Battle Beaver blue.

External Input Mapping
Button assignments vary depending on customer selection. Our most common layout is Left - X (cross), Right - O (circle).

Polling Rate
Slider: 8KHz.

My board should be running at 8KHz, but it isn't. What can I do?

First, check these common factors:

  • The setup configuration: make sure the polling rate slider didn't reset itself to a
    lower value, then re-save the calibration and check again.
  • Your cable specs: USB-A to USB-C, USB 3.0+ for best performance.
  • The USB port you're using: connect to the motherboard directly, not through a USB hub or to a front panel USB header.

If those aren't the issue, join Marius's Discord server at https://discord.com/invite/QcCkfbkp3S and ask for help there. Include as much detail as you can.

Note: to check the polling rate of your JDM-5X board more reliably, use the DEEPPOLL tool found at https://tools.mariusheier.com.

I’m suddenly having performance issues, stuttering, jittery aim, etc. on my PC. What’s wrong?

8KHz polling rate may be too much for your system (or the game you’re playing) to handle. Older CPUs, some motherboards, or certain USB ports can get overloaded by the higher polling rate.

If you don’t want to drop to a lower polling rate, try the easy options first: different cable, different USB port, different game. After that, use Marius’s CPU Direct tool to make sure you’re connected to a CPU-direct USB port.

If the issue continues, drop the polling rate to 4KHz and check again. If that isn’t it, drop to 2KHz and check again. And if that isn’t it, the most effective next step is contacting Marius via his Discord for deeper hardware troubleshooting: https://discord.com/invite/QcCkfbkp3S.

Note: Battle Beaver is unable to assist with individual motherboard-level connection issues or JDM-5X hardware issues beyond the basics.

What’s the point if I can’t have it at 8KHz?

We’re currently building an XLAT latency measuring tool from Finalmouse to provide our own robust testing numbers. However, early community testing has shown that even all the way down at 1KHz, the JDM-5X can perform faster than 8KHz OC’d DualSense/DSE controllers.

As Marius himself puts it, asking the pizza guy, “Is it done yet?” more often is less important than making sure the pizza is cooked faster.

Can I have the hidusbf software running? Can I use that software on a JDM-5X controller?

Yes, the software can be running without causing an issue. However, using it to try to
overclock one of these boards will cause performance to degrade. Instead, make sure no overclocks are applied to the JDM-5X board or the USB port it's plugged into, then use Marius's Setup page to apply the 8KHz polling rate.

Do I need to use DS4-Windows or a similar program to use this controller on PC?

No. These boards share the PlayStation 4 form factor, but they are not a PS4 board
replacement. The JDM-5X has no console compatibility and is PC-only. Because of
this, it's plug and play on PC — no extra software required.

Should I choose Analog or Digital for triggers?

For all Battle Beaver JDM-5X controllers: calibrate the triggers as Analog, regardless of whether you ordered stock or Smart Triggers. During initial setup, do not click the SKIP THIS STEP button.

Smart Triggers initial setup: follow the on-screen instructions as closely as you’re able to. Since there’s no way to actually slow press a Smart Trigger, however, getting the counter to go down during that step will require some experimentation with the timing.

If you did accidentally select SKIP THIS STEP during initial setup, don't worry:
repeating the Setup or Calibration process will allow you to fix it.

In the Trigger Configuration section in the Settings Menu, make sure the triggers are set to Analog.

The reason: Selecting Digital tells the controller to look for triggers on the Eagle Board. We wire Smart Bumpers and Triggers to a separate chip.

Note: this information may change with a future board revision or software update, and it may not match if you purchased your controller from a different company.
Last updated: February 27, 2026.

Can I swap from a 3-pin to a 5-pin thumbstick mech, or vice versa? How much would it cost?

3-pin and 5-pin thumbstick mechs and boards are not cross-compatible. The boards themselves are different where the mechs connect.

Swapping from 3-pin to 5-pin requires purchasing a new board of the correct variant. To purchase one from Battle Beaver, you have three options:

Costs vary depending on which option you choose. Base cost is $150 for a 5-pin board, and $120 for a 3-pin board, not including installation or other modification costs.
(Pricing info is current as of February 27, 2026, and is subject to change.)

What TMR sensor does Battle Beaver use in Magneto Mechs?

AKNES V5/V6 TMR sensors, retrofit onto an OEM Alps frame for a drift-resistant upgrade that keeps the familiar OEM feeling under your thumbs.

Note: this information will change over time if/when AKNES releases new versions and
we receive new stock.

Last updated: February 27, 2026.

How can I verify my polling rate?

Use the DEEPPOLL tool located at https://tools.mariusheier.com.
Note: this tool does not measure input latency itself, only polling rate.

For DEEPPOLL support, join Marius's Discord server at https://discord.com/invite/QcCkfbkp3S.