I’m about to go CAN GPS. What’s a couple more bucks to save some UARTs for obstacle avoidance and future Lua sensor mods?!
mRo F403 CAN nodes modify nearly any serial GPS module for use with DroneCAN.
Show no fear. 
Someone has to take the lead!
1 Like
I want to start off by saying there is a lot of good information here. I have a Kubota ZD326 I would like to and after this past weekend WILL be converting to an autonomous rover. There are so many things I could be doing with my life instead of wasting 5 hours every weekend bouncing around on a mower getting sunburnt and getting covered in grass clippings. I have seen the ArudPilot project and did a fair amount of reading but have no idea where to start with a project like this. I have watched several of the @Yuri_Rage videos on autonomous mowers which eventually brought me here. I have a couple of basic questions that I know will lead to more questions so I will apologize in advance. The first question I have is what servo should I use? I have seen the D845WP model that a couple here are using and the recommendation from DeepSouthRobotics ASME-04A by WingXine. Is there a preference and if so can someone list the pros and cons? Once I know what servos to use I will get them mounted and share the pictures. I will likely draw the parts I am fabricate in AutoCAD so if anyone else has the same machine and servos they can use my files. The next step will be the electronics which I will have many more questions about. If there are suggestions on a complete setup, ideally like the cm level of accuracy that I have seen, could help save me a lot of time getting this project done. Thank you in advance for any assistance this forum can supply!
4 Likes
Welcome to the autonomous mower club! If you make CAD for a WingXine servo cover I would love a copy.
There is no single, comprehensive “how to” because we all have different mowers and potentially slightly differing goals.
The HiTec D845s are great servos for HydroGear transaxles with fairly low actuator torque required. I have moved to AGF 100kg 12V servos, since they can operate directly from the mower’s charging system without a buck converter. They are holding up well except for one failure that I think I can trace to a high D gain and ensuing oscillation during SCurve alpha testing.
If you have a transaxle with a piston-style valve actuator, they may require more force than all of these recommended servos can provide.
Jason,
I have spent a lot of time figuring all this stuff out and building my mower with help from the folks here on this website. It has been a long haul and I am just now starting to activate my mower but I have enjoyed the project. Anything I can do to help others build these mowers, I will gladly do. For me it seemed harder than I thought it would be and I struggled a lot, but it did force me outside my comfort zone to learn new things. The servo selection is the first question I struggled with also and I am using the D845WP large water proof servos from Hitec. They are used by a large number of mower builders, but they might not be certain to work for for your application if your mower has hydraulic control valves that are hard to move. I know the D845WP servo works with the HydroGear transaxles. This is one of the more popular hydraulic control valves on these mowers. Some people have been forced to use more powerful servos in different situations. If you take a picture of your control valve I can at least tell you if it looks like the type I have. I also think Yuri _Rage also can help identify the control valve. The D845WP can push or pull a 40lb load (keep you fingers out of the linkage). You will find designing the linkage very tricky also in order to get the correct travel for forward and reverse. Where you have available to mount this presents a serious geometry problem. I fabricated 3 sets of interface brackets and linkages before I got mine to give me the correct travel.
Feel free to ask me questions through the blogsite or through the associated message system. I am happy to help others build mowers.
Here’s my latest linkage configuration. The Heim joint hardware is from McMaster Carr. That’s a HiTec servo in the pic. It’s now an AGFRC 12V model.
1 Like
@SJohnson Thank you for that bit about HydroGear and servos. I just adopted a z246 Husqvarna that I planned on converting in the next few weeks. I didn’t know what type of transaxle it had until I just looked it up. It scared me for a second.
1 Like
Steve, thank you for the information and offer to help identify the transaxle. I got the mower stuck pretty good and haven’t had a chance to clean it off after pulling it out of the pasture. I will get it cleaned up tonight and take a few pictures.
1 Like
Attached are the photos I took of the transaxle and linkages after an hour of cleaning up from getting it stuck in the cattle pasture. I can remove the linkage and measure the torque required to move the lever with a beam style torque wrench if anybody thinks that is worth the effort. I like the waterproof body of the AFG and the S845WP unlike the WingXine ASME-04A which appears to have an unsealed PCB and connectors. The 125kg/cm (9ft/lb) of torque the AGF 100kg offers as compared to the 50 kg/cm (3.62 ft/lb) of the D845WP. For those budget conscience the D845WP ($104 on Amazon) is roughly 1/3 the price of the AFG 100kg ($279 on Amazon). Thank you for pointing me to these two servos. I am fairly confident I will opt for the AGF 100kg servo as @Yuri_Rage pointed out the AGF runs on 12V-16VDC where the D845WP runs on 4.8-8.4VDC.
1 Like
I’m not familiar with those transaxles. Might be smart to at least estimate the torque required to operate them (engine running).
I am with @Yuri_Rage, I have never seen transaxles that look like that. Measuring how much pull is required move them might save you some time and money. I am not that experienced on the transaxles, but I would be suspicious that the torque required to move them might be different with the hydraulic pressure applied.
Glad to hear you are building another mower.
How has the reliability been long term when using the HiTec D845 servos. I am only running 5v on mine and they move the transaxles very easily with the engine running driving the mower around. It seems like I am not working the servos very hard.
I had one failure with the D845s. @ktrussell has had one or two fail as well. They aren’t bulletproof, but they do seem to get the job done reasonably well. For the price, I’m not sure you’ll do much better.
The AGFRCs are a little more expensive and seem to be slightly stronger with the added advantage of running on 12V.
1 Like
The best I can do for now is to measure how much torque it takes to move the lever with a beam style torque wrench at the point where the control push rod bolts on.
. With the engine running and the controls (stick and push rod) disconnected the in/lb torque wrench is showing about 80 in/lb of force needed to move and hold the lever just short of max wheel speed forward or reverse.
With the controls connected it take about 105 in/lbs. If the method I have used to measure the force needed to control the transaxle is correct the AFG 100kg will be working pretty hard at 8.75 ft/lbs. The lever at the transaxle is pretty short and there is room for the lever to be extended to lower the torque required by the servo provided the servo has the throw necessary to move the lever at the transaxle limit to limit.
1 Like
And there’s your problem with these hobby grade servos. The D845s and AGFRC 100kg models are about the max you’ll find at that price point, and the available throw can be pretty limited (mine are maxed out for throw and just barely cover the full range required).
You may need to step up the game to a more industrial servo.
1 Like
Hey Jason - Another option is to use servo actuators. Great choice for high load applications but a bit pricey. I put two on our original mower and it was extreme overkill with our HydroGear transaxle but worked great. 4” of travel, 25+ lbs stall, 500 lbs static load, 12V power, about 2.2”/sec travel and has a built in BEC so you can power the servo rail.
1 Like
Steve, thank you for the suggestion and link! The price is not so bad at $299. I was just looking at gear reduction servos as well. Any idea what the mean time between failure is for this linear servo?
Also remember that you likely will not be moving the controls to their limits when mowing autonomously. Most mowers will hit 8-10 mph but your machine will likely be mowing under 5 mph forward in auto and you will seldom need more than a crawl in reverse. You can plan your linkages accordingly. Good luck!
1 Like