As far as I remember you were using UDP on your tractor, are the units powered the same way as before? which pcb version are you using? (not that it matters if v1 or v2)
Is it erratic because of vibration? Or is it because of RF interference?
its a tiny toy motor, i put my hand on the imu to stabilize - but made no diff
Here you see my MMA mounted in the center of the image. It is 1 meter from my steering motor but maybe too near power cables. But I could test sometime in the following weeks with another MMA unit and will report back when I have the chance.
Not done yet, but my test rig for aog. All this stuff worked perfectly in the tractor, but by the motor the MMA is wonky doodle.
The motor turns a 10 turn pot that is the same as a complete steering with WAS.
I had a test setup a year ago with a 1 mega ohm pot, and no mma. That setup had spikes in steering value when pot was turned. I thought it vas lack of power as setup was only powered vith USB. Have to put setup together again an try once more with hall sensor and mma also.
The black box on the left side is a 3A 12v supply.
Seems vibration is an issue with the mma. When it was solidly mounted in the tractor to the cab frame there was no problem.
If the standard Arduino Uno isn’t powerful enough to do this kind of integration, there are quite a few Arduino-compatible boards out there that run 32-bit ARM processors. I’m a bit partial to the Teensy boards. A lot of muscle in a small package. I’ve got some Teensy 4.0 boards on order to do some CAN processing, and they are 600 MHz! Crazy.
The F9P can be talked to with ublox protocol, instead of waiting for and parsing NMEA. Can work over I2C or serial. I haven’t had any luck with the I2C stuff with Sparkfun’s library, though.
A simple ESP32 is approaching two orders of magnitude faster than an arduino in some situations.
Nano 33 with wifi, onboard imu and the same pinout as a Nano. Only 3.3V though.
Everything is getting to be 3.3v these days. Can level shift logic voltages fairly easily if need be (another PCB revision perhaps). Most things like relays and motor controllers can probably work with 3.3v just fine.
All the faster boards and processors are 3.3v. The esp32 is probably 5v tolerant on input pins due to them being open collector. Also the esp32 has fantastic floating point speed for doing coordinate transformation and roll integrations.
My MMA is mounted half way down on the RH side (fuse side) of the white component box. Hot glued to the box, wired with ribbon cable which also allows one with a different pinout to be used (Adafruit). It seems to produce a rock steady output. My BNO is mounted centrally at the top of the box. Again, hot glued in place. Also very very stable output. This area is at the rear right hand side in a Fendt 716 cab. Power and reset switches are on the front edge of the Fendt box. The whole thing is wired to the IBT, WAS and steer switch with RJ45 terminated Cat 6 cable. 8 conductors being just enough.
I’m really impressed with the general stability of the system. My test setup was far less well secured (basically on top of the fendt box) and the MMA output was much less stable. I guess this points towards vibration too.
That would be nice just put in Nano 33iot or 33 BLE or 33 BLE Sense , and connect via bluetooth or wifi.(including IMU and even magnetometer for the BLE types, and the extra speed available )
Two changes needed A and B:
A: Supply 12/15v (or 5v from WAS power supply) to v-in on pin 15 , which is not connected on PCBv2. Nano 33 wants between 4,5v and 21v on v-in (Nano 33 has no connection to pin 12 or as they say, normally NC). Pin 12 position will have 3.3 v on PCBv2 when IC3 is changed as follows.
B: shift IC3 to a 3,3 v regulator in order to supply ADS1115, it can work on 2v to 5v. R-783.3-1.0 Recom Power - Voltage Regulators - Switching - Distributors, Price Comparison, and Datasheets | Octopart component search
Alternative B: cut away 5v connection line to ADS1115 and feed it with 3.3v from nano +3v3 power out.
If DOGS2 is used it must be connected as Brian suggests in the connection schematic.png, png is supplied with the AOG download.
Here they tell that md13s is 3,3v compatible. 13Amp 6V-30V DC Motor Driver
Whilst I understand the benefits of having the motor controller on the PCB I do wonder whether, given the sensitivity of the MMA and BNO, if it is not better to remotely mount it nearer the motor / valve?
I know the BNO isn’t recommended but I’ve found it rock steady mounted within the PCB enclosure. I do however use an IBT2 controller mounted well away from the PCB, and I’ve also given the IBT and electronics completely separate direct power lines.
This then means only one cable to outside (antenna) and lightweight signal wires from the PCB enclosure to WAS and motor controller. CAT6 in my case. No heavy power wires running back and to either.
Here is a pcb that can use a Nano or a Nano33.
With I2C there are limits on how long the wires can be. At the highest data rate, length is really limited to about a metre. At lower data rates, it can be up to 10 metres which is more than sufficient for AOG’s purposes. I’m not completely sure what transfer rate the MMA code in the sketch is using, but probably slow enough for longer distances?
Completely Agree Alan. Mine is located in much the same place firmly attached to the cab frame. The MMA is just as stable as the DOGS2 expensive one. I have both on for testing.
In a couple days i’ll take the DOGS of the tractor as well and put it on the test platform, see if there is a difference with tapping vibrations
I’m sorry during the conversation
Can the diyAutoSteer PCB designed by BrianTee be replaced by Arduino Nano 33 BLE?
And Is it possible to modify AutosteerPCBv2.ino and connect it to a PC as a Bluetooth virtual serial port?
I have a Windows tablet with one USB port ( as it is used for charging ),and no Ethernet port.So I want to connect with AutosteerPCB via Bluetooth.
Not sure, can you compare the pinouts?