Looks great, I think you contacted me on YouTube last week? But yeah it looks like we are building similar machines for similar purposes. Hopefully we can share our ideas and come up with something cool. I am currently debating going electric ie. Wheelchair motors or using an old Toro greensmaster lawnmower with a petrol motor driving a hydraulic drive and steering.
1 Like
What is the max speed on your servo?
Hi friends,
Iām trying to Building something similar to your drone. In particular, my drone will have a hybrid powering, I mean a petrol engine that will power a 2 Electric motor track.it will be used for spraying a 40 hectars wheat field. It will be driven through a Cheap but efficient RTK system and should be equipped with a lidar for obstacle avoiding.
Iāll be glad to participate to this cool project.
What about the controllata unit? How do you think to set your field contour and headings?
Hi @azadani,
Might I make a humble suggestion. I donāt know much about your build, but if you donāt need all of that distance between your drive wheels and your steering wheels, why not shorten the length of the vehicle and have it steer by differentially driving the motors? You could then throw a swivel caster on the front to replace your steering motors.
You can get a motor controller like the Sabertooth from Dinension Engineering for ~$100. Again, donāt know your constraints but this would be a little simpler.
I think @ktrussell tried using some actuators to control steering on a zero turn mower and had a heck of a time getting things tuned weāll. The desired steering versus achieved steering can deviate significantly if you donāt design and tune your system well. He succeeded from what I can gather, so maybe he can give some guidance.
Best of luck, looks like a pretty cool build 
1 Like
Gāday Neil,
This is looking very promising. Nice an neat!
I feel your pain with that linear actuator youāre using for steering. Iām using 2 of them on a zero-turn mower. The potentiometer feedback and the DC power for the actuators is fed through an Actuonix linear actuator controller. Took care of being able to interface the actual actuator position back to the Pixhawk Cube. Sing out if youād like more info. As others suggested, a couple of castors and skid steering from your motors could be the go,
I was just looking about thinking of using a scooter like yours to do some slow, steady mapping with a lidar in orchards. Love your progress. Sing out if you like. Just up the road near Ballina.
Cheers,
Ben
Hi @azadani,
Iām not sure why I didnāt see your post sooner and the suggestion from others that I might be able to help. I do have very slow linear actuators on my zero turn mower, but amazingly it controls pretty well in auto. It is tricky to drive in manual because you have to anticipate when to release the steering before the rover has finished the desired turn. I do have some slightly faster actuators that I will try soon. They are still relatively inexpensive. If they donāt do the trick, I will try faster ones, perhaps just a heavy duty servo. The other issue is that there is a pretty significant deadband at the center of the left and right steering handles. I think that causes a little difficulty. But, I have a usable autonomous mower. (https://www.youtube.com/playlist?list=PLIsYv3gzZOt9N9yZmpI_WYiMaBQ6HazOG)
The only real issue in auto is that pivot turns seem to overshoot a good bit, but everything is repeatable, so the mowing area gets covered. I am blaming the overshoot on the slow actuators. I will know once I get faster ones.
Iāll be glad to help you in any way!
Kenny
That was to me a very inspirational doco.
Whats the plan with this?, would you be driving it remotely and identifying the weeds via a video camera. Would it be driving the entire fields and using software to identify and spray weeks automatically? Or would u be boom spraying the whole area using gps?
1 Like
Hi Kenny
How is the farm, im having trouble connecting my ardusimple to my black cube via Ardupilot RTK/GPS inject someone told me on that you are the man when it comes to this part.
are you able to help out here
That is a very clean design im looking at the same accept a potential rotary geared actuator
Hi Rooster,
Well, I donāt know about being āthe manā! 
I am not injecting RTCM3 corrections via MAVLINK if that is what you are doing. I have never done that, although it is the most common way, it seems. I have my F9P base GPS communicating directly to my F9P rover GPS via an RF link. In my case, I am using Adafruit 915MHz LoRa modules. You could use other telemetry modules with success, perhaps even easier, as long as you get the rate and range you need.
Is this the kind of setup you are attempting?
1 Like
Hi Kenny,
Thankyou so much
It seems the Xbee radios have somehow stop working. Let me read up on these Lora radios.
Are you achieving 1cm accuracies?
If so Iām in all the way
I have been following you and Robo on YouTube
Can you send me a schematic on how you configured the boards if possible.
And I really appreciate your help in advanced
I also liked how you wrote a script to mow a polygon throwing the cuttings outside.
I have a few priority projects right now but I will be glad to send you a sketch of how I have my base and rover connected via the LoRa modules. It might be a few days, though.
1 Like
If any of these projects are related to Pixhawk and f9p please I would be very interested
They certainly are and I have been documenting as I go with video. I am very close to having a complete new layout of my controls that I will look a lot better and hopefully be more reliable than the messy, wires-everywhere system I have been using! I am replacing the Pixhawk with a Kakute F7-AIO module, however. Iāll update everyone when I get it operational.
1 Like
Hi Kenny
Iām quite new to Ardupilot and also plan to make my own lawn mower using RTK GPS. Iām currently looking for the ārightā auto pilot hardware and are wondering, why you choose Kakute F7-AIO?
Hi Raphael. Welcome to the community!
My decision process wasnāt really very scientific. I had heard a good bit about it here and it was only $50, compared to the $100+ for a Pixhawk 2.4.8 from a reputable company (and Iām not sure how one knows what is and isnāt reputable, since the original manufacturer no longer makes them.)
I have 2 Pixhawk 2.4.8ās. One worked very reliably for over a year.
The other was a little unreliable after a short while. I honestly canāt remember what kind of trouble it gave. The original one that had been so reliable started failing recently (rebooting during a mission, which is a very scary thing on a big machine) and failing to retain waypoints after a power cycle.
BUT, I really torture a flight controller on my mower in the Georgia (USA) heat for hours on end. So, I am NOT badmouthing any of them.
So, I thought I would try something else. I sort of bought it on a whim but so far I am happy with it.
Here are the pros and cons that I know of at the moment FOR A BIG ROVER WITH NO LIMITATIONS ON SIZE AND WEIGHT:
-
PRO: It has 6 serial ports. There are some 30 dollar controllers with just 2 serial ports, which is not enough in my opinion.
-
PRO: Has ability to inject On Screen Display information into a video stream if you install a camera and transmitter. I have not put a camera on my mower yet, but one dayā¦
-
CON: Does not have any of the normal connectors that are found on the Pixhawk, so you have to solder them on in some way or wire directly to peripherals. For small aircraft, where every gram counts, this is probably a plus, but it is a little bit of a pain otherwise.
-
CON: Some of the signals are brought to a .1 inch dual row header, so you can wire to them pretty easily. Others, such as the I2C signales, SCL and SDA, are not on the connector. Instead, you have to solder directly to a pad. I wish they at least had thru-holes in them to insert a wire. The SCL and SDA, in particular are ultra tiny. But, it wasnāt really that hard to wire to them.
-
CON: Only has 1MB of Flash. I had not thought about this until I got it all ready to go. I was worried when I saw that some features are left out of Ardupilot on the builds for the 1 MB controllers per this https://ardupilot.org/copter/docs/common-limited-firmware.html. Under Ardurover, it reads āN/Aā. I wasnāt sure if that was āNot Applicableā or āNot Availableā but I will say that the one item left out of Copter that I really use a lot is Guided Mode. I was happy to discover that Guided Mode does work on the F7. So, I have not run into an issue with memory. I can load a 500+ waypoint mission just as I could with the Pixhawk.
-
PRO: Has an SD Card slot. Some of the low cost controllers do not. I donāt know if those still store binary logs that are accessible in some way or what. But, I do know you will want to be able to store binary logs, and having a removable media makes it faster to load into your computer.
-
Small CON: You have to load a bootloader on the device before you can use Mission Planner as you normally do to update firmware. This took me a while to figure out for some reason, but I would not let this sway my decision in any way.
-
CON: No CAN bus, but I donāt have a need for that at the moment.
Here is a short video showing my new control box: https://youtu.be/3ohdYgGsTd0. You will notice I made a custom carrier board for the F7 that brings out the signals I need to both the standard DF13 JST connectors and to .1" pin headers do I can use whichever makes sense. I plan to organize the wires a little better with tie-downs, so this is a preliminary video!
I have been running this F7 AIO for a few days now and probably logged 20 hours or so with no problems. So far, so good.
Iām sorry to be so long-winded in my replyā¦
Good luck with your build!
Kenny
3 Likes
Greetings,
Iām also looking to build a farm rover that does not use skid steering. Iām wondering about your choice of a linear actuator to control the steering. This idea had not occured to me, and I have been searching for a servo motor to be centrally positioned between the front wheels and connected with motor arm extension and tie rods to the wheels, much like a go-kart setup. However, I am having trouble locating a line of servos that function like the RC scale servos, but have the power to turn something larger. It sounds like the linear actuators have the power and Iām wondering if they are a better fit for this application.
Steve,
Perhaps others will give you more direct application help with your question. I will share a few things from my limited experience with linear servos/actuators. They can be quite strong. This table includes some with 180 lbs of force: https://www.servocity.com/heavy-duty-linear-servos/. They are a nice fit mechanically in a situation where you are trying to move something linearly (kind of obvious!).
Some gotchas to watch out for:
- Speed - many of the stronger ones are also quite slow. Notice the 180lb ones in the table above move at 0.3"/sec. That is extremely slow. I would say that is too slow to drive the steering directly and have any hope of tuning it unless your vehicle is going to move super slowly. And even then, tuning will be hard, I believe.
- Life of the servo: I do not see a spec on how many strokes the servos above are rated for, but I have used linear actuators before on an industrial project and was disappointed in their life. The actuator I used back then had a rating of 25,000 strokes. I really didnāt think about it. That sounded like a lot of strokes. BUT, the application called for the actuator to stroke 3 full strokes once every 6 minutes. After the actuator failed in a few weeks, I did the math: 25,000 strokes was only about 34 days! Now, in a steering application where the movements are mostly small with an occasional full stroke, I do not know how you determine the life. So, you may just have to learn by doing (or someone else here may can advise).
- Duty cycle: The servos normally are rated for a 20% to 25% duty cycle, meaning that you canāt just keep them moving all the time or they will fail prematurely. I used small linear servos on my zero turn mower for a year or so and did not have a problem with duty cycle, but Iām just mentioning it. You likely wonāt run into a problem with duty cycle in a steering application.
- Canāt manually move the servo: You cannot force the servo in or out. The lead screw inside and the gearing just wonāt allow it. So, plan for a way to steer your unit manually by pulling a pin if you need to be able to do so.
How big is your planned vehicle? I also want to build a larger farm vehicle, however, I like zero turn so much that I am still hoping to make it zero-turn when I do get to that project. Right now, it is on the far distant horizon, I am afraid.
I am very interested in your project!
Kenny
Thanks Kenny for all the additional info on linear actuators. This is a vehicle that will generally need to go quite slow. At times, we will need it to creep along, and I canāt imagine a top speed of more than 10mph. 5mph will likely be adequate, as the ground it will travel on is bumpy and contains rocks here and there, so anything too fast could cause the payload to bounce around.
The vehicle will be 66" wide, and probably 60" to 66" in length. Its main purpose is to carry produce around the farm, from the field to our pack shed. It has to have this width in order to be able to travel down our vegetable beds in the same paths that the tractors do. It will also need to be two to three feet tall so that it can clear the plants. Iām trying to build something fairly light to start off and develop it, but I eventually hope that it could carry several hundred pounds.