Alan Fistwood Posted April 17, 2021 Posted April 17, 2021 1 hour ago, propeler said: This is a better solution indeed Another advantage that comes to my mind on the CoreXY kinematics is that every workload will be shared always by two motors working together to obtain a movement on a singular axis, so by example if you keep pulling hard on a dive recovery you will not stress only one motor for obtaining the same force output but that force will be obtained by both A+B motor working together. This could benefit a lot smaller motors to avoid the overheating issues that you mentioned earlier from your testing. I'll try to design an do some motion simulation on something based on this sketches in the following days, but I am not sure I will ever be using Hoverboard motors, cause they take up to much space and I am constrained by space so i'll try to design something compact that can be upscaled just as a concept.
RealDCSpilot Posted April 17, 2021 Posted April 17, 2021 Prototype finished. Worked right from the start. 4 i9 13900K @5.5GHz, Z790 Gigabyte Aorus Master, RTX4090 Waterforce, 64 GB DDR5 @5600, PSVR2, Pico 4 Ultra, HOTAS & Rudder: all Virpil with Rhino FFB base made by VPforce, DCS: all modules
RealDCSpilot Posted April 18, 2021 Posted April 18, 2021 (edited) Here is how i solved the implementation of the G940's MLX90333 magnetic position sensor (Melexis): It's basically a "mini joystick-within-a-joystick" to scale the huge movements of the big gimbal to a much smaller one. The magnetic sensor is turned by 180° to mirror the X/Y input from the main stick, the contraption "slides" within the big gimbal's T-connector to compensate moving on the z-axis. The ball joint has a little friction so the magnet's position is always very precise. Due to new motor arrangement their polarity had to be reversed. Edited April 18, 2021 by RealDCSpilot 1 i9 13900K @5.5GHz, Z790 Gigabyte Aorus Master, RTX4090 Waterforce, 64 GB DDR5 @5600, PSVR2, Pico 4 Ultra, HOTAS & Rudder: all Virpil with Rhino FFB base made by VPforce, DCS: all modules
Alan Fistwood Posted April 18, 2021 Posted April 18, 2021 My Progress so far, did some changes on the arrangement to be able to mount it on a desk mount like Monstertech, changed the gear ratio from the original 0.20 to 0.23, this sacrifices some potential torque but allows me to have a compact enough gear, I'll add a gear and belt cover later on when I am sure of the layout. From the original design made by Hotpretzel I converted the size of the parts to metric measures, so the bearings now are easy to find 6006zz and the shaft for the axis is just a 2020 profile cut to length. With this configuration I can get at least 60° of motion in all direction (30° from center) and the stick is really compact and small enough that it can be printed on a 200x200 printer.
RealDCSpilot Posted April 19, 2021 Posted April 19, 2021 @Alan Fistwood Don't forget to include an active cooling solution, especially for your x-axis motor. An open air vent with a 40 mm fan directly underneath the motor should fit your design. i9 13900K @5.5GHz, Z790 Gigabyte Aorus Master, RTX4090 Waterforce, 64 GB DDR5 @5600, PSVR2, Pico 4 Ultra, HOTAS & Rudder: all Virpil with Rhino FFB base made by VPforce, DCS: all modules
walmis Posted April 19, 2021 Posted April 19, 2021 It looks a bit disbalanced because of the upper motor weight, one axis would probably need a constant motor torque added to counter balance it.
Thadiun Okona Posted April 19, 2021 Posted April 19, 2021 (edited) 10 hours ago, walmis said: It looks a bit disbalanced because of the upper motor weight, one axis would probably need a constant motor torque added to counter balance it. The roll axis motor is at the bottom of the carriage below the pitch pivot and the pitch motor is on the fixed frame, looks sensible to me. The roll motor will not be easy to cool in a pocket like that, also no provisions for belt tensioning yet. Edited April 19, 2021 by Thadiun Okona
RealDCSpilot Posted April 21, 2021 Posted April 21, 2021 Finished the active cooling design: 3 i9 13900K @5.5GHz, Z790 Gigabyte Aorus Master, RTX4090 Waterforce, 64 GB DDR5 @5600, PSVR2, Pico 4 Ultra, HOTAS & Rudder: all Virpil with Rhino FFB base made by VPforce, DCS: all modules
TomVR Posted April 21, 2021 Posted April 21, 2021 What is the firmware/software that everyone is using to drive these setups? Is there an arduino handling the code to emulate the directx ffb effects or some external software calculating the forces?
Dirt_Merchant Posted April 21, 2021 Posted April 21, 2021 This is my favorite thread on this forum!! 1
finch Posted April 21, 2021 Posted April 21, 2021 (edited) 48 minutes ago, TomVR said: What is the firmware/software that everyone is using to drive these setups? Is there an arduino handling the code to emulate the directx ffb effects or some external software calculating the forces? I guess realdcspilot is using his logitech G940 hardware, I'm going to mod one of my MSFF2, the only problem I have is cut the case panels Edited April 21, 2021 by finch
RealDCSpilot Posted April 21, 2021 Posted April 21, 2021 (edited) @TomVR Like finch said, for this mod i'm using just the main PCB of the G940 base, it's motors and the magnetic sensor. fred41 made a special firmware modification for me, to disable the "deadman sensor" functionality that is implemented in the grip of the G940 stick. For the grip i used the PCB of a WarBRD base. Edited April 21, 2021 by RealDCSpilot i9 13900K @5.5GHz, Z790 Gigabyte Aorus Master, RTX4090 Waterforce, 64 GB DDR5 @5600, PSVR2, Pico 4 Ultra, HOTAS & Rudder: all Virpil with Rhino FFB base made by VPforce, DCS: all modules
TomVR Posted April 21, 2021 Posted April 21, 2021 (edited) Got rid of both my MSFFB2 right before the pandemic unfortunately. I have an OG MSFBB1 collecting dust in the junk pile, but unfortunately even with a gameport adapter the joystick uses a friggin camera for x/y/z position data, so no way to wire in external hall effect sensors. Is there any work on a separate standalone board that handles the generation of FFB effects? I know there has been talk for years of Freejoy doing something with FFB, but no movement in awhile there. Edited April 21, 2021 by TomVR
finch Posted April 22, 2021 Posted April 22, 2021 (edited) 14 hours ago, RealDCSpilot said: @TomVR Like finch said, for this mod i'm using just the main PCB of the G940 base, it's motors and the magnetic sensor. fred41 made a special firmware modification for me, to disable the "deadman sensor" functionality that is implemented in the grip of the G940 stick. For the grip i used the PCB of a WarBRD base. how do you adjusted the motors torque? have enough power to feel the force feedback with a stick so long? In the MSFF2 you can add resistors to increase the power, I did with mine when I added a longer stick Edited April 22, 2021 by finch
vogel25 Posted April 22, 2021 Posted April 22, 2021 16 hours ago, TomVR said: What is the firmware/software that everyone is using to drive these setups? Is there an arduino handling the code to emulate the directx ffb effects or some external software calculating the forces? I just came across this thread, I'm building a ffb joystick with hoverboard motors and a odrive, I'm using this (https://github.com/jmriego/Fino) for firmware, it runs on an arduino leonardo.
RealDCSpilot Posted April 22, 2021 Posted April 22, 2021 (edited) 4 hours ago, finch said: how do you adjusted the motors torque? have enough power to feel the force feedback with a stick so long? In the MSFF2 you can add resistors to increase the power, I did with mine when I added a longer stick I'm currently running forces set to max at 150% in the driver settings. I'm sure i also could do an electronics mod like you, i already have a new power adapter with 24V/5A connected. But i need advice from a pro first. On the other hand i don't use "effects" on the stick (turned it to 0), i only use trim and control surfaces effect force at 100. For all the other airframe effects i have a Forcefeel pad on my seat (Realtus Forcefeel) and use Simshaker for DCS, a much better way to have haptic feedback of the airframe on the whole body. Edited April 22, 2021 by RealDCSpilot i9 13900K @5.5GHz, Z790 Gigabyte Aorus Master, RTX4090 Waterforce, 64 GB DDR5 @5600, PSVR2, Pico 4 Ultra, HOTAS & Rudder: all Virpil with Rhino FFB base made by VPforce, DCS: all modules
propeler Posted April 22, 2021 Author Posted April 22, 2021 3 hours ago, vogel25 said: I just came across this thread, I'm building a ffb joystick with hoverboard motors and a odrive, I'm using this (https://github.com/jmriego/Fino) for firmware, it runs on an arduino leonardo. Actually, you even do not need separate controller for it. ODrive is capable to work as FFB device. My first prototype was using ODrive with custom firmware and is wisible in system as FFB joystick without additional manipulations. I use odrive for DD wheel as well.
TomVR Posted April 22, 2021 Posted April 22, 2021 (edited) 3 hours ago, propeler said: Actually, you even do not need separate controller for it. ODrive is capable to work as FFB device. My first prototype was using ODrive with custom firmware and is wisible in system as FFB joystick without additional manipulations. I use odrive for DD wheel as well. Any plans to put that firmware on the github? Would be nice to have an entirely standalone FFB device not requiring donor hardware, especially with the joystick costs on ebay right now. Edited April 22, 2021 by TomVR
vogel25 Posted April 22, 2021 Posted April 22, 2021 15 minutes ago, TomVR said: Any plans to put that firmware on the github? Would be nice to have an entirely standalone FFB device not requiring donor hardware, especially with the joystick costs on ebay right now. The github I posted is Opensource and works on a standalone arduino leonardo.
propeler Posted April 22, 2021 Author Posted April 22, 2021 2 hours ago, TomVR said: Any plans to put that firmware on the github? Would be nice to have an entirely standalone FFB device not requiring donor hardware, especially with the joystick costs on ebay right now. Yep. Now one axis firmware already is on github https://github.com/o-devices/bldc-ffb-wheel-firmware for two axis will be available as well soon here https://github.com/o-devices 2
Phantom_Mark Posted April 24, 2021 Posted April 24, 2021 First moment one of you guys has a working system available to purchase that is a worthy replacement for my MSFFB2 consider me onto the purchase list Sod the Brunner, would be much happier to fund guys here. 1
RealDCSpilot Posted April 24, 2021 Posted April 24, 2021 (edited) Thanks to fred41 it was possible for me to make a special firmware patch for the G940 to set new axis ranges for my configuration with stick extension, a nice side-effect was a slight increase in motor torque. I also redesigned the air tubes for the cooling fans today, to let the heat get pulled out instead of pushing cool air in. Also installed less noisier noctua fans now and this makes my project feature-complete. Yay! Big Thanks to @propelerand @fred41 !! Without them this would not have been possible. Edited April 24, 2021 by RealDCSpilot 3 i9 13900K @5.5GHz, Z790 Gigabyte Aorus Master, RTX4090 Waterforce, 64 GB DDR5 @5600, PSVR2, Pico 4 Ultra, HOTAS & Rudder: all Virpil with Rhino FFB base made by VPforce, DCS: all modules
TomVR Posted April 24, 2021 Posted April 24, 2021 On 4/22/2021 at 12:58 PM, propeler said: Yep. Now one axis firmware already is on github https://github.com/o-devices/bldc-ffb-wheel-firmware for two axis will be available as well soon here https://github.com/o-devices Great, looking forward to a 100% open ffb stick design. I've got a huge ass cast aluminum Tornado grip that I really want to built onto a FFB base, and it's very heavy so will need all the torque I can get. https://streamable.com/7sfttr 3
walmis Posted April 24, 2021 Posted April 24, 2021 (edited) Did a force test on my machine. Currently I get ~5kg @ 18-20A motor current, motor dissipates around 60 Watts at that current. (Motor phase resistance is around 0.135ohm) Edited April 24, 2021 by walmis 3
Recommended Posts