KevinG's Autonomous zero-point turn Lawn Mower

You’re already getting RTK corrections over the radio, so leave that alone unless it stops working for some reason.

One of the items that was important to me on my mower was to change the grass discharge to be out the back of the mower to allow it to handle taller grass out here in the country and to make programming easier. It was a chore welding up the side of the mower deck and opening up the back. It also got more complicated than I wanted when I realized I needed to reinforce the back and I had to remove all the veins or baffles directing all the grass over to that one small exit port. Here are a few pictures of how it turned out.

Let us know how it cuts. Only rear discharge mower I have seen in person left windrows of grass due to poor airflow. But it did handle tall grass well. Good luck!

I’d be a bit concerned about blowing all of that debris directly back at the tires. It will probably be ok, but it’s a pretty curious design choice, at the very least.

I have 2 other rear discharge mowers that work just fine on other machinery. I will also adjust air flow underneath the deck a little bit if needed.

Wow, guys! The last month or two on this thread have been really cool to see, now that I’m finally catching up. I’m sorry I haven’t been around. This is a huge source of valuable info shared by so many who know their stuff!

@SJohnson, you have got done some fine work. Thanks for sharing the good photos of your progress.
@jason_miller, we’ll be seeing nice progress photos from you soon, I am sure!

I have been so busy (aren’t we all?) that I have not had a chance to upload any new videos, and that has partly kept me from the forum. With the new Lua scripting and a few super useful scripts written by @Yuri_Rage, autonomous mowing has come a long way. It is truly useful, not just a novelty. I want to report on all that!

But, before I could report on all the good, some bad happend. :frowning: I really wanted to report it on YouTube first, but who knows when that will be.

So, to counteract all the encouraging photos and news you guys are reporting :slight_smile: … here is my mower:

Over the previous 3 days or so prior to the fire, I had mowed 30 acres of my 50-acres of hay fields.

I had a slight leak from the carburetor float bowl a couple of days prior, which was very scary, but I tightened the bowl and I checked it occasionally on the day of the fire, but I still feel that was the cause. The mower set the grass on fire in a line as it mowed for about 50 yards or so before it reached its final end. I think it was streaming burning gas as it went.

I was inside my house, which is in the middle of my fields. I had Mission Planner monitoring the mower, but I was over on the PLAN tab creating the next mission. I had the volume low so as not to disturb others close by. When I checked on the mower by turning up the volume a tad, Mission Pkanner said “No data for 210 seconds.” I looked out the window and the mower was totally engulfed. The fire was too intense and dangerous to attempt to extinguish myself. I let the fire department do that. I hope soon to share more, including video of the scene captured by my wife from the house.

So, I certainly have some lessons learned to share! But, hey, it was just a machine. One of my daughter’s first questions to my wife was “How is Daddy doing? That was his baby!” :slight_smile:

Fortunately, I learned of a really nice used mower for sale nearby that is similar to my original, but a little higher-end. I bought it! So, Version 2 is coming soon! I plan to share the details as I go.

OK, so I hope I didn’t hijack the thread. Please don’t let that happen. I just wanted to let you all know.

I’m actually excited to start over!

Now, you guys continue on with the GOOD stuff! :slight_smile:
Meanwhile, I’ll be cutting my grass the boring way!

EDIT: By the way, the fire happened on May 2. Prior to now, I had told @Yuri_Rage and @Swebre and a couple of others. Our local sheriff broke the news on Facebook that there was a “tractor” fire on my road, and since there are not many farms on my road, the local folks pretty much knew it was me! I had someone ask me if I was going to automate a fire truck… hmmm, an idea…:slight_smile:


Very sorry to hear that Kenny. You and Yuri have inspired a lot of people for sure.


It is great that you are back Kenny, we have all been worried about you. That is a very sad looking mower!

1 Like

Maybe a diesel mower next time ?

you could convert it to EFI, there are a few opensource systems out there like megasquirt and i think there is support in ardupilot for some of them. stuck float bowl is a very common fault. All it takes is a tiny piece of dirt in the float bowl and it usually jams open. fuel injection will solve that problem along with giving you a lot more engine telemetry.

1 Like

Wow! Not much left of that mower. I am glad the fire did not spread. I bet that was a sight to be seen. I have caught a vehicle on fire before and they can make one heck of a mushroom cloud that can be seen for miles. If I am not mistaken this is the second time this mower has caught fire. I think I will be looking at fire suppression options.
Automating a fire truck, funny!
I concur with @myozone, diesel! That is coming from a guy that wants to put a Cummins 6BT in everything. You can go down to the local fast food restaurant and pick up some used vegatable oil and run the mower on veggy! I have done it. Other than smelling like and wanting fried chicken while mowing it works great.
With 50 acres to mow it might be time to get a Toro GroundsMaster with a 12 foot deck. I have way less than your 50 acres and I considered it until I saw the YouTube videos @ktrussell and @Yuri_Rage posted. It took me a long time to find this forum and I know God led me here to find some good people willing to offer their experiences and suggestions to get me off the mower and in the pond for the hot summer days ahead.
@ktrussell, I will be posting pictures soon. We have/had the mower working from the RC remote but it was a bit jerky due to the short throw distance of the levers on the transaxle. We are 3d printing a longer lever today to make sure there is adequate clearance before fabricating one from steel or an aluminum billet.
The electronics enclosure has already been 3d printed but after seeing the damage in your picture I am reconsidering… We might change to a fire safe!
We have also been working on a PCB design for all of the electronics to mount to. Sounds good in principle and should suit my needs well it may not work for others and has been through many revisions as we test and learn from the results.
We started out designing the schematic to understand all of the components required and signal flow but have also been placing the components and routing traces in case we decide to order a prototype from SunStone.
My intention is for anyone with a Kubota ZD326 to be able to use the BOM my stepson and I are putting together along with the CAD files to have what I believe to be a good starting point for putting one of these mowers together or at least, hopefully, understanding what is involved better than I did when I first found this forum.
I also wanted to comment on the professional look of everything @SJohnson does. The craftmanship is perfection. He has a wealth of knowledge as well.


@geofrancis, DIY EFI would certainly integrate with our automation quite nicely! I’m not sure I want to tackle that anytime soon, but it’s a fine idea!

I’m actually working on a Lua+Arduino based I2C temp sensor array that might help put some minds at ease here. There’s a lot of interest in temp sensors, in general, and there are few solutions for discrete sensors that work in the ArduPilot ecosystem.

I’d like to monitor enclosure temperature as well as engine temperature. I think I may further automate my choke using engine temp as a guide for its setting/timing. As such, I don’t need a super robust thermocouple for cylinder head temp or EGT, but rather, just a simple sensor near the engine block, which could likely remain within the sensing range of a DS18B20 - my present vector for the Lua+Arduino project.

Further, a couple of smartly placed temp probes could sense free air temp and warn of overheat or possible fire conditions, as well.

I hope to have an update on that soon!

@jason_miller, I’m not sure a fire safe enclosure is a wise choice. It would have to be well insulated, and your components might be in jeopardy of overheating.


@Yuri_Rage, I wasn’t serious about the fire safe I would not use the fire safe for the same overheating concern you mention. Cutting a hole in the safe to add ventilation defeats the purpose of using the safe to protect the electronics against fire damage.
I will consider using 5052 aluminum for the enclosure to reduce the possibility of the plastic breaking down in the sun, the plastic reaching its melting point and possibly creating a fire hazard, and still allow the compass to read magnetic north even if the compass is not used at the moment.

Has anybody integrated a way to monitor engine RPM or blade spindle RPM? I am sure I can add a microcontroller and crank sensor to the motor but wondering if there is already a solution that exists.
I have been thinking about a way to monitor the engine load in order to slow the mower if it is moving along at a given speed that is too fast for the height or density of the grass for a good cut and could be used to notify when the blades need sharpening, the engine is misfiring, the blades have impacted something, or disable the blades instead of stalling the engine provided there is time.

Wouldn’t even bother considering a magnetometer or making any design concession for one. Mowers have way too much ferrous metal and electromagnetic interference for most magnetometers to be useful. Use dual GPS receivers and a moving baseline config.

I think you’ll need to use Lua and a discrete RPM sensor to monitor things the way you describe.

1 Like

Hey Jason - My guidance system on the tug/mower is in a waterproof plastic box but has a generously sized aluminum sunshade several inches above it. Internal box temps during a 3-4 hr run only get ~10F above ambient on a full sun day.

Ditto Yuri’s comment about the magnetometer. Been there. Not worth the effort. GPS for yaw is
well proven now and works great.

Have not added it yet but engine RPM can be had using simple inductance and wrapping a wire around a sparkplug lead and going into a Pi, Ardunio, etc. Many small engine tachs work this way reliably. Have thought about monitoring the blade RPM but don’t see a need in my case. During mowing, the engine is on the governor and any slowing beyond a certain amount of inherent governor “slop” (150 rpm?) indicates overloading of something and the need to slow down a bit. Serious RPM drop and I’d like the mower to just shutdown and await a set of eyes to survey the situation. Folks say these are easy scripts to write, but that is not my backgound so I’ll have to rely on others for now. Good Luck!


Thank you for the suggestion on RPM sensing.
We have the GPS configuration you describe implemented and so far what is being reported back to the base station/MP is very promising from my trip around the property setting waypoints.
The only 2 issues we have left to resolve are the telemetry radio LoS and distance. I originally ordered the 100mW mRo radios thinking they would be good to go and for the most part they are but there are a couple of dead spots. I am hopefully the RFD 900 1 watt radios will resolve the issue for me when they arrive. The other issue is throw distance of the control linkage with the OEM lever on the transaxle. I am looking forward to reporting back with good news and pictures.


Good to hear from you.
Thank you for the suggestion which could prove helpful to others looking to implement such a feature. Unfortunately or maybe fortunately, my mower is a diesel, no spark plugs or plug wires to read RPM from. I have a few crank sensors laying around from my automotive technician days. I will take a look at implementing this over the winter if I find there is an actual need for this. For now I will manually override the speed if needed and keep a set of sharp blades on.
If that doesn’t suffice, the mower might get a turbo charger or a Cummins! I’m not actually going to put a Cummins on a mower.

I missed that you were dealing with a diesel. Sorry. But I am a bit jealous. Diesel is what I need. Perhaps the next build….

A small crank sensor would work great for rpm readings. And a turbo! What could go wrong?

Over the 4 years I have had the autonomous mower, I have cut many acres, but mostly a relatively small number at a time. I cut the 50 acres about once a year over several days. @myozone I know a diesel would be awesome, but I can’t really justify the cost at the moment.

As to my D845WP failures, all 3 were related to trauma! I had both left and right servos fail at the same time when I ran the mower into a wooden fence. A rail of the fence hit the throttle sticks (which I still had present and connected on the mower) super hard, stripping the gears in the servos. The other failure happened when my home-made linkage between a servo and the mower’s linkage to the transaxle siezed when a nut self-tightened against the linkage. It was a nylon-insert style stop nut, but still tightened itself over time. The next mower will use proper heim joints if I need such a linkage.

So, I can’t fault the D845s for any of my failures. I think pretty much any servo, rotary or linear, would have likely failed in these cases. The only downside, in my opinion, to the D845 is that it requires 8.4V (or lower). This isn’t too much of an issue, but not ideal. On my next mower, I am going to attempt to go all 12V, but price may dictate otherwise. I’ll make the decision pretty soon.