Fixing and Improving My DIY Adaptive Headlights
HTML-код
- Опубликовано: 13 сен 2024
- I built some adaptive headlights for my car some time ago. The concept seemed promising but the project was never really finished and it had some issues. In this video, I demonstrate how this lighting system works and share the process of getting the thing fixed up for an upcoming road trip where ability to see well at night will be very important. If you are interested in making your own DIY adaptive lighting setup, I'm hoping this video provides you with some insight on things to consider and ways to avoid issues that I encountered.
The amount of happiness i have by finding this video❤❤❤❤❤
This is awesome
THIS is Exactly what i was looking for. THANK YOUUUU❤, Also i subbed
Glad to hear! Are you working on something similar?
@@CurtBagneNot currently but i want to do something like this.
why in the world didn't you use a draw wire encoder for the steering angle? It would have been less prone to problems, and it wouldn't need anywhere as much engineering to get the thing attached. I am in the process at this time of adding 2 HID projectors to each of my headlamp assemblies. When done they will be curve active and also horizon active. All of the electronics are inside of the housing. You are also going to run the risk of cracking your windshield because of the heat from the lamps. I don't see any kind of a limit switch system involved in there, how is the design supposed to know if say you stripped a gear in the servo? You might also consider using RC car shocks that have the springs internal with this type of shock you can cut a small piece of spring to put on the opposite side of the plunger from the large spring. Fill with a heavy oil and now you have something that gets rid of servo slop
That came down to cost. I looked into using a string pot but at the time when I first built this, any decent looking options were on the order of a couple hundred dollars. The style of pot that I used can be had for ~$10 and integrating it really wasn't that bad. A string pot would have eliminated the gears and any lash concerns but the gears have proven reliable so long as the lash is set properly and they're sufficiently well protected to keep the wire for my steering wheel controls from getting sucked in. I had more problems early on when a couple of pots that I sourced from Amazon failed but when buying what was allegedly the same model from DigiKey the problems went away so I'm suspicious that the Amazon ones were counterfeit.
Are you documenting your project anywhere? Sounds pretty cool!
I'm not worried about the windshield cracking since it doesn't get very hot inside the housing. The housing has a sufficient surface to volume ratio to dissipate the heat effectively. No worse than running a defroster.
There aren't any physical limit switches but there are software limits that, to date, have been successful in keeping the components from reaching mechanical limits.
I've actually considered incorporating RC car dampers and might still do it if/when I revisit this project. The springs that I have now soak up any slop in the servos but the servos can still be kind of jittery in their movement. I might be able to address this to an extent with software filtering too.
@@CurtBagne I haven't done any real documentation yet on my lamps. I have 2 of them to do so after I iron out the kinks in the first assembly I will then document what I do on the second. Mine is pretty involved, I had to cut the back of the housing and extend it to fit the additional pieces in there. I ended up building a vacuum forming machine and I bought a sheet of the same plastic the housing is made from and I formed the piece I added onto the housing. I then used 1/8" rod made from the same plastic and put it in a Dremel tool leaving only 1" sticking out. I set the Dremel on it's lowest setting 5000RPM and I pushed the rod right at the seam where the housing and the new piece met. I went around the entire seam "welding" the 2 pieces together I have some photos and some information here.
www.solsticeforum.com/threads/dual-projector-hid-retrofit-attempt.162037/page-2#post-2223275
I am using the RC Car shocks as dampers, the ones I got have internal springs and they also come with 4 different plungers that have a different number of holes in them. This is to set how much dampening the shock has. You can get a draw wire encoder that is well made for about 100 dollars., the thing that drives the price way up is the length of the string. get the diameter of the steering shaft, do the math to get the circumference and times that by the number of revolutions the wheel can make (typically 3 turns 540° degrees each direction). It's a really short string that is needed. add in some some length for mounting and also so you don't hit the mechanical limits.
I am using a 3-axis gyroscope, 3-axis accelerometer, 3-axis magnetometer (to correct drift) coupled to a Digital Motion Processor (DMP) to get the orientation of the car. I am using madgwicks quaternion equations to break down the orientation readings into pitch and roll of the vehicle in degrees. My vehicle has stability control so I was able to reverse engineer the CAN frames for the stock steering angle sensor. Because I do not want to add any additional load to the car I also added in a 3.5 farad super cap so when power gets cut I am able to "park" the lamps, this way I do not have to go heavy on the springs. The cap is super small maybe 1" x 1" x1/2" there will be one for each headlamp assembly, it's enough to power 4 servos at 2 amps of draw and all of the electronics for close to 3 seconds.
If you got decent metal gear servos that are digital jack up the PWM frequency to 200, this will give you a far greater precision for the servos. It will eliminate most of the jerking. Another thing you can also do is for loop between every 10 PWM settings between the last position and the new position adding a small delay between the changes.
@@CurtBagne Oh I also found a company that makes a worm gear servo that has the same footprint as the ones you are using. A worm gear servo will have 0 slop and it will also lock in position when powered off. They come at a high price tag of 70.00ish USD a piece and they have a lead time of 2 months to get them. the model number of them is EVO-P4W-SiC
@@kevinschlossser7858 This is really awesome--I'll definitely be following your build thread. Especially curious with how everything will turn out with the IMU and using the feedback from it. A couple of thoughts though... Have you looked into getting vehicle position info off of the can bus? Most vehicles will have a gyro to measure the roll axis. They use a "leaky integrator" to avoid drift. To my knowledge, most vehicles don't have a gyro for pitch; however, as it's not so crucial for avoiding rollovers.
You might also be interested in adding something like this so that your setup can respond to traffic: pixycam.com/pixy2/ I wanted to incorporate one of these in my setup to aim my lights down and/or dip the high beams to respond to oncoming traffic or to be more courteous to vehicles in front of me but ultimately decided to just keep it simple because time.
I'll look into some of those servo ideas, especially if they're fairly simple to integrate. Thanks for sharing.
@@CurtBagne On most vehicles the accelerometer and gyro are located in the passenger presence module that is apart of the airbag system. Typically the EBCM and the air bag systems are tied together on their own databus. The information can be requested from the EBCM, in order to get that request frame you would have to use an OE diagnostic tool on the car. I would have to have access to one and set up a man in the middle type of system to know what is being sent out of the diagnostic tool.. at to be honest at 5 bucks for the MPU9250 i wasn't worried about it. I am running an ESP32 to handle everything. these things are dual core and have a 240mhz RISC processor. so they are fast and do not have problem doing complex floating point math. They have gobs of ram compared to the Arduino's. and there are exactly enough pins to do everything I need to for a single housing. I run MicroPython on it so developing is a breeze. and I can push software changes to it over bluetooth or wifi. It does have a can interface built into it as well tho this is not supported in MicroPython.
I had to come up with some pretty crafty ways to take care of some of the issues I ran into, because of how the lamp is directly connected to the servo and there is no use of a rod I quickly discovered there is actually a large amount of play in a servo's spline gear where the servo horn attaches. and my time that play gets up to where the lamp is the lamp would wobble about 1/4" total. I ended up getting aluminum servo horns and made a plate that foes over the bottom servos, it essentially joins them together, this play sits under the servo horn and is such a tight fit that the aluminum servo horn slides on it. It acts like a slide bearing and removes the wobble from the spline.
I am also trying to build this thing with cost in mind. I want to make it as cheap to do as possible. If I didn't have to build the vacuum forming machine I am about 200 bucks per lens.