My name is Jan and I would like to start experimenting with Ardupilot.
My question is about the best gear (possibly budget to mid-range gear) to be able to create Gps Mower.
I plan on using two servos to be able to convert my Toro 32 mower to an autonomous one. To my knowledge and research, I need a flight controller (maybe Pixhawk but I do not know which cube one) RTK GPS (here3), some radio, power distribution, etc.
Could someone help me pick my gear I am a bit overwhelmed by the amount of stuff I found.
Hey @Jan_Hacek welcome to the Ardupilot website. I am happy to help you get started on your project. I will move this conversation over to another more appropriate “thread” so you will be hearing from me. I would like to set it up so it’s easier for other new users to find and follow the conversations. The information you need is scattered all over the Ardupilot website. It’s tricky to find and some of it has changed along the way. I’ll help you ferret out what you need and share my experiences. @ktrussell@Yuri_Rage
We will just use this thread to talk about all aspects of new mower builds and I expect the conversation to range beyond just the hardware. In building a mower you have to start somewhere so I have just picked the heart of the beast, the flight controller, to discuss on this post…
A big part of this is pouring over the Ardupilot documentation online. I tip my hat to all the extensive work many people have done to learn and document everything. It takes a while to go through all the information, but is very well done.
I will try to resist pushing a single flight controller because many of them would work for this application. Many of the mowers you will see out there use the Orange Cubes and they have all the processing power and memory that the mower requires. One thing that makes the Orange Cube popular is the plug and play aspect of the device (you buy a cable set and plug things in). Some of the other ones like the Mateksys H743-Wing have just as much processing power and memory and they cost quite a bit less. The negative is that they are not packaged as nice and you start pretty much with a loaded board, putting connectors on it, and making your own cables. Making cables sounds scary, but in the end you have to understand exactly where every wire goes and what it does. These cables are small though and it does take some patience and practice to assemble them.
The Matek H743 series have VERY similar processor and memory specs as the Cube Orange and Orange+. What you give up with the Matek boards is heated/vibration damped IMUs (not super important features for us) and the convenient form factor of the Cubes. You actually gain a UART or two. A good compromise might be the ZealotH743 from QioTek, which is packaged into a very nice enclosure at a lower price point than the Cubes.
I do like the H7 autopilots for mowers because they allow for a lot of scripted automation that proves very useful for our sometimes peculiar needs.
If the forum supports such, it might be a good thing to create a “sticky” thread that would stay at the top of the thread list, dedicated to the creation of such a rover mower. I’m sure there’s a bunch of discussion of such, but if it’s not easily accessible, then it’s not helpful. The more of us that get involved, more knowledge will be disseminated!
The wiki (where all the tuning and parameter docs live) is a good place for something like that. Perhaps we should start a page dedicated to the craft. I think it would be well received as a subset of the Rover docs.
@Swebre suggested a while back that we collaborate on something like that. I created an example GitHub repo for it, but it never really grew any legs. Having it live within the ArduPilot docs might be even better than something hosted independently. It’s quite easy to make edits to the official Wiki.
One of the more tricky parts in building an Ardupilot mower is interfacing with the hydraulic control valves because not all hydraulic systems are not the same. I tried to do some detective work and I pulled up a parts ordering system for a Toro 32 mower and it showed that it uses a HydroGear transaxle. If that is really true (and please check them), then your job gets a little easier. That valve is easier to move and several servo actuators will work. I use the HiTec D845WP heavy duty servos on my HydroGear transaxles only running at 5 volts. Other folks have started moving to the 12V servos from AGFRC (I think). Maybe someone using the 12V servos can comment. The 12V servos eliminate the need for a separate 3.5-8.4v power source that the old reliable HiTec D845WP servos required.
I will hesitate to sing the praises of any given servo, since I’ve had failures on all of them.
The HiTec D845WPs are somewhat proven, but I think they are possibly just slightly underpowered for the task (even on the lighter duty HydroGear drives), as they seem to heat up a bit under heavy use, and I’ve had one fail out of the blue, with a second failure that was probably more related to some abuse by yours truly than any flaw or normal stress.
I started using 12V AGFRC servos after my last D845 failure, and they have served me well while eliminating a buck converter from the system. However, I think AGFRC has some quality control issues. After singing their praises initially, a few of us ordered a bunch of them and we experienced several failures either right out of the packaging or shortly thereafter. If you get a good one that lasts more than a few hours, it’ll probably continue to work just fine. But for the price, it’s hard to give them a wholehearted thumbs up. That said, I have 4 of them in various torque ranges on my mower at present, and they are holding up well after more than a season’s worth of use, along with some rather abusive tuning sessions.
I tested and also tore down this 400kg servo from “HappyModel” (what a name!). I was impressed that it actually delivers the advertised torque as well as by its build quality. It’s very large and very precise, but maybe a touch slow to act when run at 12-14V. It gets downright snappy at 24-28V, so any sluggishness can be overcome as needed. The price is right, and it looks like a great candidate, especially for some of the larger/heavy duty transaxles on various machines. However, I only tested one on the bench and cannot vouch for its longevity under regular use.
Thank you and others for providing useful information.
Yesterday I did some research and found Pixhawk orange cube+ standard set for a great price (I got a discount because I am a student) so I will pick this as my flight controller. Also on that website, they sell Hex Aero HERE3 +, Telemetry Radio Set V3 433MHZ from Holybro, and many more things for ArduPilot.
The next thing on my list is probably gps. So my question here is will HERE3+ be enough for my mower to mow straight or do I need an RTK base for it? And if Here3+ is even good for me. I saw that you use two gps one is like a mobile base and the other is rover gps to my understanding, what is the difference between a stationary base and a mobile base for GPS.
So, could someone check this website out and tell me if they have something good or not? I would really like to order from them because of the discount I have and because I am from Croatia and they are close to me.
Purchasing your RTK GPS system will probably be one of your biggest costs so you have to do that carefully. I have never used the HERE GPS systems but I know many others have used them successfully.
I have always used the Ardusimple Simple RTK2B boards. I have 2 Ardusimple RTK2B GPS boards on the mower. Here is a picture of the boards: Yuri's 4.1.0-beta Update (GPS yaw - now 4.3-beta1) - #237 by SJohnson
One is a moving base used for GPS Yaw steering and one is the main GPS for the mower. In addition I have my own base station for sending the GPS correction date over to the mower (RTCM3 signal). The GPS Yaw arrangement accurately determines the heading where the mower is pointed. Normally you would use a compass to measure the heading but they do not work well on a metal mower with a lot of magnetic interference sources. The compasses essentially do not work well enough on a mower to be reliable.
In relation to your other question about whether of not you need a base station of not, the answer is you will need some kind of GPS correction information sent to your mower. A lot of people use an internet based NTRIP service routed through Mission Planner for fixed base to moving base RTCM3. As I mentioned above I built my own base station using a third Ardusimple RTK2B board with a radio attached to the GPS board to send the correction to the mower. The more common way to get the corrections is through an internet based NTRIP service. I went with my own base station because it was actually very easy to do and I do not have fast and reliable internet where I live way out in the country.
Maybe some of the other users who actually use the HERE precision GPS systems will enter their thoughts. I am not the right person to judge if the HERE3+ system is the right one for your build. I am not an expert on the GPS systems and they are a very challenging part of the mower build.
The Here3 and 3+ are great for a single GPS Copter with an occasional desire for RTCM3 corrections. I find them inadequate for moving baseline/agricultural rover applications. The satellite count and ability to maintain a solid RTK Fixed solution are just not good enough.
Okay, I will go with the Zed-F9P-based GPS module.
I looked into internet-based NTRIP service and in my area, I have a CROPOS provider that charges 0.066€ per minute of corrections and a one-time payment for registration of around 40€. Gear for this setup costs around 450€ simpleRTK2B – 4G NTRIP Starter Kit
Another method using simpleRTK2B Heading – Basic Starter Kit
This kit combines one unit of simpleRTK2Blite, acting as a moving base and one unit of simpleRTK2B acting as a rover. The cost of this solution is also around 450€ but it does not require a subscription to my knowledge
I got the impression that the method with NTRIP service is easier to set up, but it is more expensive at first glance. I am interested in what you think is a better method and what will produce better results. My area has good internet and cell coverage so this should not be a problem.
If you maybe know a better method than the ones that I have found please let me know
Something about the proposed configuration doesn’t sound correct to get the precise location and steering we normally achieve.
We need help from others more experienced with the GPS configurations and the options to get NTRIP. I didn’t know there were lower cost ways to get NTRIP, which contributed to my choice to build my own base station.
No, you can’t just magically create subscription-less NTRIP/external corrections by buying the SimpleRTK2B+Heading kit. And, let me reiterate this yet another time:
DO NOT BUY THE SIMPLERTK2B+HEADING KIT FROM ARDUSIMPLE FOR USE WITH ARDUPILOT!
Yes, it can be made to work. And you will hate your life for making the effort to do so.
Unless you plan to mow at various properties where it is inconvenient to set up a fixed base, you should probably forego the NTRIP subscription. You’d only have to mow for about 50 hours before the data cost per minute would overcome the cost of another F9P module and antenna to create your own fixed base.
Here’s what I recommend:
3 SimpleRTK2B modules
3 survey grade antennas
Mount two of them on your mower with the antennas about 1m apart and configure them as a moving baseline configuration. This DOES NOT give you more accurate absolute position. But it does provide a heading reference that is FAR more reliable and accurate than a magnetometer (compass).
Use the third module to create your own fixed base. I used a tripod over a fixed landmark for a while before moving it to a mast on the back of my house where it doesn’t create an eyesore. I use a Raspberry Pi and solar panel+SLA battery to power/connect it. There are many other options that might be a bit beyond the scope of this topic for the moment. It is not a difficult task.
I second Yuri’s opinion about the SinpleRTK2B+Heading kit. It’s very difficult to get it working with Ardupilot and I discourage you from trying. Instead go with 2 simplertk2b modules on the rover.
I also second his opinion about building your own fixed base. It’s easy to provide your own RTK corrections and I think you should build your own local RTK correction source (using a 3rd simplertk2b module and a raspberry PI or other computer) rather than paying for corrections.
The pins on this map are all NTRIP sites that can be accessed freely via RTK2Go. Many of them are privately operated, sometimes unreliable, and sometimes incompatible with our use case(s), but they are free.
Texas runs a robust CORS network, but you have to be a surveyor or gov entity to subscribe. It’s not public, and even a subscription is difficult to obtain.
Good to hear, this tread exists. I started with the idea of automating my zero turn in mid 2021. Bought a bunch of parts, and as quickly as it started the project ended. Now I am back, literally, since yesterday, going through my notes, and relearn what I have bought, and why, and what it does.
My current gear is:
CubCadet ZT2 (Zero turn)
Pixhawk 4
ArduSimple RTK base station (up and running since mid 2011)
two ArduoSimple board for heading
6 servos to automate 2 hydro-static drives, two brakes, throttle, and choke
an Arduino MEGA for the first step of automation to ‘electronify’ the current mechanical controls.
RadioMaster remote
some radios to do the RTK stuff (Holybro 915MHz)
The Arduino code as prototype was put together in March 2022, and I am currently getting my head around what I had written then.
Well, happy automation; let’s see where it leads us
