PANDA is the system that the teensy boards use that synchronize the IMU and F9P so for each GPS reading there is a perfectly timed IMU reading sent along with it. This means that your roll compensation is actually compensating for the correct roles at the same time.
As I understand PANDA is just the code. PANDA needs a Teensy since the arduino is to slow to synchronize IMU and GPS data. Perhaps I got it wrong?
alright thx, that already cleared it up so much for me.
I got into the group buy and that discount does look good.
So guess Iām going micro in this case!
next to get the board ordered, looks like JLCPCB got a lot of shortaged on the parts, is there an alternative site maybe? looks like a lot of the headers are out of stock, would this maybe be something i could solder on myself? Or is the new V4 most surface mounted and isnt this really an option?
also looking for a Teensy now but hope to find one on a European webshop so i dont have to pay all the taxes.
I ordered the AOI board without headers since they were out of stock. In the local shop I purchased 40-pin 2.54 mm and 2.0 mm headers. My plan is to cut the headers to the correct number of pins and solder them to the AIO boards myself when they arrive. Hopefully it will work.
Back to the RTK correction. I could read that it seams that using so called NTRiP protocol is a way of transfer corrections over internet. If I understand correctly you could transfer corrections from either your own base station or neighbor, community etcā¦
This seams to be the way to got in order to avoid problems with weak radiosignal transfer.
I would appreciate if someone could confirm if NTRIP is the way to goā¦
This is my first season using ntrip or any rtk corrections for that matter, and I would just say make sure you have good reception everywhere if youāre relying on ntrip. If you do I canāt imagine anything being better or easier.
For me as a beginner it seams that this DIY project will grow. This feels tempting:
So tempting 
Thanks B_bimmer, NTRIP seams the way to go. We have great 4g reception in our areaš
Hey @minorplanets,
I saw your setup with the WAS on the T5 New Holland. it looks really good!
what kind of WAS did you use here?
Are the 3D casing files on this forum? Would love to get this on my tractor aswell.
update: ordered the PCBās, the Teensy, the BNO and the GPS stuff. So next stop is the hardware on the tractor. hopefully there will be a good RTK signal nearby.
Iām using this:
The cover is a bit over engineered as I was learning Onshape and 3d printing at the time. It fits on the hub with magnets and the sensor is upside down rotating around the hub axis with fixed part attaching to the grub screw a little ways onto the axle. Printed in PETG and hasnāt been busted off yet. It covers a grease nipple so I need to either simplify removing the assembly or get an angled grease nipple.
Itās not a particularly effecient use of the sensor travel compared to a linkage based setup (relatively small travel angle) and Ackerman is significant. U-turns at higher speeds turns one way better than the other which probably means I need to fiddle the Ackerman a bit more.
haha it looks really nice, can tell you put a lot of thought into it!
Iām not sure yet what kind of setup i want to do for my WAS, but just thought yours looked really nice.
did give me some food for thought tho, never really thought about the Akkerman effect on the WAS before.
the placement right on the steering knuckle will increase the Akkerman effect right? Am i right when I say that this effect should be less when you place the WAS with a link onto the steering cilinder?
Other question: what would be the main reason to go for a 24V motor, is it just to have some extra power? seems to me that a 12V should be able to do the job and it saves you a converter.
update: teensy is in (man thatās tiny) looking up all other hardware parts now for the autosteer.
When you adjust Ackermann correctly, then steering around the middle (straight ahead) is not affected. You should also be able to set Ackermann so U-turn work same to both sides.
The Prius WAS is a potentiometer not a Hall effect sensor, like many other WAS used with AOG. But I admit the low price is tempting.
This is the inside of a Prius WAS that got water inside.
Yeah they can be considered a consumable. I mounted it upside down to avoid dust getting in through the seal. I keep a spare and the one I had fail had a dicky seal, so maybe the cheap ones have failed QC.
Ackerman is adjusted in the software and adds a factor to the steering in one direction and that works just fine, however itās not quite that simple in real life as it isnāt linear. Your tractor has a tierod across the front so it would be fairly straightforward to mount the WAS to the axle closer to the pivot point and have linkage connected to the centre of the tierod eliminating Ackerman altogether. You ca also use linkage to "amplify " movement to use the full range of the sensor.
Thatās how I did my last tractor and it was literally tied to the axle with fencing wire. The new setup, while much prettier, is arguably less functional. It also tends to pickup a bit more mud in wet conditions.
Is everyone using this cost-effective but non durable WAS setups? Any proposals of any durable variants?
Try the link in post above
to be honest, just to get started I did order the lexus/toyota sensor off of Ali.
might not be super durable but it will get me started to set my systems up. Iāve seen a few very nice Elobau sensors but they are quite expensive, maybe Iāll go to a car breaker or something in the area to look for a OEM sensor for cheap.
Good way to get started, I got my BMW sensors from a car scrapyard
Thanks larsvest, I missed that linkā¦
By the way are you from Sweden? I was thinking about your name 
I actually bought som cheap ordinary potentiometers for my bench trails. I also have a 12v motor for the trails. Got informed that the v4 bord has arrived at our local post office. Things are getting forwardš
For test the will work, but might be noisy (when you turn them, was signal might jump)
Not so far away. Jutland DK
Thanks for the input. If I get problems on the trails I could suspect the potentiometer.
Seems like hall based rotational sensors are to be preferred.