Raspberry Pi and sprayer pressure controller

Projects Success stories
1

Couple of years ago I wrote some Python code for Raspberry Pi based motor controller. Little over year ago I wrote that on Rust, first ever project on Rust language. I used it for last summer.

Something new for this year. Sprayer pressure controller.

Sprayer control

Sprayer pressure controller is separate feature from AgOpenGPS. Software on Raspberry Pi takes care of sprayer pressure reading and pressure control. One GPIO pin is also used to monitor whether boom is on locked position. Two GPIO pins (relays) are used to control pressure valve.

There are two options for pressure control, constant pressure and speed-based variable pressure.

Sections on my sprayer are controlled with ± 12 V. I use relays to control sections. Pressure sensor takes ground and 12 V and signal is about 0,5 V to 4,5 volts. Pressure control valve is like section valves as it is controlled by ± 12 V. I use relays for that also.

In my setup, this controller replaces controller for my Amazone sprayer.

IMG_20250331_171919
IMG_20250331_1719191370Ă—1027 143 KB

The black box on top of Amazone conroller is just a switch box, auto on/off, manual section control on/off and sections switches. Main controller and relays are on other box (not in a photo). Amazone controller is just for backup as I used it last year for pressure control.

Other (Windows) software is used to monitor sprayer pressure and to control settings.

Software has settings screen for nozzle spacing, nozzle size selector, litres/ha, min pressure, max pressure and nominal pressure. Nominal pressure is used for constant pressure.

Second screeen to monitor target pressure, current pressure and speed. Buttons for controller on/off and button to select variable or constant pressure. Also indicators for min and max speed for selected pressure range, indicator for boom locker yes/no and selected nozzle size.

I have tested the program on Android phone. The problem with Android software is, as I share wifi from my phone, there might be some issues with ip-addressing for Raspberry Pi controller. And lack of interest of tweaking Android app.

settings_screen
settings_screen436Ă—818 13.5 KB

monitor_screen
monitor_screen436Ă—818 18.2 KB

I have not used this on a field yet. It’s not spring yet. But i’ll be ready.

2 Likes
2

Looks a good set up This is what I am waiting to do once I get my autosteer all completed

3

Could you explain more how the speed based pressure control works? Does it assume linear spray volume/pressure ratio or a more complex formula?

Is the sprayer equipped with a flow sensor (or two like many Amazone sprayers do, one for the return flow too)? My point is that as far as I know the EU regulations ask for volume control which OEM Amazone controllers do.

The Amazone implementation (that I know of) is not perfect though. At times I’d prefer simple pressure control (speed controlled). The way Amazone implements volume based control fails e.g. at headlands where pressure control is halted and may stay at a very low or at a very high pressure. Also opening/closing sections causes issues when switching isn’t instantaneous (motor controlled valves) and some liquid passes to return lines (unmeasured volume). Would be much better to have the primary spray control pressure based and just run the volume measurements for longer term calibration using steady state spraying (i.e. ignoring feedback just after any section valve has been operated).

4

Pressure control

litres_min = speed * litres_per_ha * nozzle_spacing / 600.0,
where speed is in km / h, nozzle_spacing is in meters.
pressure [bar] = (litres_min / (NOZZLE_CONSTANT * nozzle_size))^2,
where NOZZLE_CONSTANT = 2.3095, nozzle_size is a number, nozzle number 025 equals to number 0.25.
litres_per_min = NOZZLE_CONSTANT * nozzle_size * sqrt(pressure)
speed = (litres_per_min * 600) / (litres_per_ha * nozzle_spacing)
speed = (NOZZLE_CONSTANT * nozzle_size * sqrt(pressure) * 600) / (litres_per_ha * nozzle_spacing)

I took some time to look at sprayer nozzle charts and figure out formulas and nozzle constant.

Sprayer pressure control

I don’t know exactly how pressure control is implemented on my sprayer (Amazone UF901). It is the most basic model. At least there are no flow-meter. I think it is some kind of 3-way valve. Without active control pressure is rpm-dependent. There are also adjustments for each sector to balance pressure when part of sections are on/off. There are 6 (maybe) pre-settings for different nozzle sizes. I think that is also rpm-dependent. Is is hardly ever right. If settings are off, pressure increases or decreases when turning sections on/off.
My original controller only has digital pressure meter and one switch to manually control pressure up/down.

Section valves are motor controlled. It is still quite fast. Based on motor sound it takes around ½ second to fully turn on/off. Hardly any noticeable delay.

5

OK, sounds like a very good enhancement for a basic sprayer.

I’m just puzzled about applying this approach to a volume controlled (Amazone) sprayer. Perhaps it would be possible to rip off the original pressure control wiring and leave the rest as it is. The sprayer would need to keep the additional bypass valve that keeps the pump volume high enough when only a few sections are open (we have 9 sections on a 20/21 metre sprayer).

Motor valves are both fast and slow. Fast in the sense of section control switching for coverage but slow in the sense that during the sub-second switching time, part of the liquid flow is lost to “unmeasured return”. The sprayer controller then detects higher than true flow and reacts quickly to that but recovers slowly. At every section switch, the volume measurement turns incorrect shortly even if spray pressure remains constant for this short period of time.

This issue concerns a pre-ISOBUS time Amazone sprayer (Amatron+ & GPS Switch), don’t know if the current models are any better.

6

Knowing nothing about volume controlled sprayer. Just thinking out loud. What is idea behind volume controlled system. Well volume is the unit You want in the end (litres/ha). Knowing the width and speed it is just basic calculation to get volume that is needed. Pressure is just a by product that needs to be in some range. Controller knows what sections are closed, so it can just subtract those from volume.
What comes to anomaly on section switching time. That sounds exactly the same effect as section bypass adjustment in my sprayer. And only now I understand Your previous post on this subject.
Are there any adjustments near section valves on Your sprayer?

7

What do you mean with adjustments near section valves?

Yes, volume is the thing that at the end needs to be accurately controlled. Sounds easy with flow meters but at transients it is much easier to keep pressure constant rather than follow an abrupt change of liquid flow (even more difficult if part of the flow escapes from the flow sensor for a moment).

1 Like
8

IMG_20250402_192818
IMG_20250402_1928181421Ă—1066 257 KB

These are motor valves.

IMG_20250402_192827
IMG_20250402_1928271426Ă—1070 216 KB

And these are the things I was talking about. There are actually 5 dfferent presets for different nozzle sizes. Apparently these should let the same amount of liquid to bypass line as one section would when section is closed.
If these are tuned correctly pressure (and flow) should remain the same on closing sections.
Without similar approach there will be some transients.

I think that difference between flow and pressure measurement is that flow measurement is more direct and nozzle size is not affected.

9

I see, completely different valves in our sprayer. As you say, no need to adjust anything for different nozzles.

valve
valve632Ă—473 61.9 KB

10

Maybe Your sprayer doesn’t need that adjustable compensation as it is actively controlled. At least thought not to need. Mine is for manual control so it is more important to keep the pressure constant with some sections closed.

As I tested my new controller, small changes in rpm to pressure are corrected within one second or less. I didn’t expect it to be that fast. At some rpm-pressure combination manual pressure adjustment is really difficult to get desired pressure as it changes so fast.

So… As You said. Maybe Amazone implementation is just not perfect.