Hazelnut Combine Development

Development of a complex piece of machinery like a hazelnut combine begins with the establishment of major design specifications that control/limit design options. For example, major design specifications for this project included a maximum height in the transport position of 10.5 feet, the ability to transport the machine on a deck-over trailer with a width of 8.5 feet or less, the ability to quickly transfer a bin of nuts into hopper trailers with sidewall heights up to 11 feet, a picking tunnel width of 5 feet, and the flexibility to harvest hazelnut plants at a variety of plant heights. Guiding design principles (i.e., design selection criteria) for this project (in overall order of importance) were/are (1) functionality and compatibility (i.e., how the does a particular option fit with other systems on the machine), (2) safety, (3) cost, (4) KISS (Keep it Simple, Stupid) which is related to cost, (5) interchangeability/flexibility, (6) repair-ability by owner, and (7) ease of maintenance.

Basic Unit Operations and Material Flow

Before any detailed design work can begin, it must first be established exactly how basic fundamental unit operations will be performed and how material will be moved or will flow between these individual operations. 

Often, an equipment development project results when an individual visualizes a different/unique combination of unit operations after observing various machines in operation, and then decides to develop a machine featuring this unique combination of unit operations.  In this case, the harvester combine resulted from UW-Madison research that showed (1) a variety of fruit harvesters could effectively shake hazelnut clusters off hazel shrubs, (2) the cleaning/winnowing system on a Hasatsan 2200 (a Turkish-built hazelnut harvester used to vacuum nuts off the ground) was simple and effective, and (3) the threshing unit built by John Bashaw was capable of effectively freeing hazelnuts from high moisture content clusters.

For various reasons (many that are articulated in the following sections), it was decided to construct an over-the-row harvester featuring side-by-side rotary shakers, with material removed by the shakers subsequently moved to the rear of the machine with screw conveyers (a.k.a. augers), where vacuum is then used to suck the conveyed material into a vacuum chamber near the top of the machine.  A single screw conveyer at the base of this chamber then feeds the material into a rotary air-lock, that in turn feeds the material into a John Bashaw designed threshing unit.  Material exiting the threshing unit then drops through a moving stream of air that diverts (and hence removes) any material with a high specific surface area (i.e., a high surface area to mass ratio).  The velocity of this airstream is controlled by a variable speed winnowing fan. Material not removed in this winnowing operation drops into a rotating drum-in-drum sorter virtually identical to that incorporated into the Hasatsan 2200.  The inside drum of this sorter features round holes each with a diameter slightly less than 1 inch. The outside drum contains slots with a narrow dimension of about 3/8 inch. Material entering this drum-in-drum sorter takes one of three paths. Material that does not fall through the inner drum is sucked back into the vacuum chamber as it exits the inner drum.  Material that gets trapped between the inner and outer drums is made up almost entirely of nuts free of their involucres. This material is moved by air to on-board storage.  Material falling through both the inner and outer drum is allowed to drop to the ground.

At this point, it is important to note that cleaning operations for harvesters that remove clusters directly from plants (instead of picking or vacuuming them off the ground) are typically less complex as they do not need to deal with stones, dirt, mud, and various other items that are present and/or that collect on the ground.

Self Propelled vs Pull-Type Combine

The first major decision on this project came after initial brainstorming and evaluation of existing fruit harvesting equipment. This decision was to develop a self-propelled combine instead of a pull-type unit as implied by the title of this SARE grant. The reason a pull-type unit was initially proposed was largely because it was felt that adding a power unit (i.e., engine) and wheel drive system to the harvester would significantly increase cost. Additionally, we also had picked hazelnut clusters with the Johanna 4 (a pull-type aronia harvester) without major problems.

After some digging, we became concerned about side-draft issues with an over-the-row pull-type unit (note here that the Johanna 4 we tested was not an over-the-row harvester). Side draft is the tendency of an implement to move or be forced in a direction at right angles to the direction of forward motion. Such movement increases the greater: (1) the trailing implement is offset to the side of the tractor, (2) the force at which the crop and/or soil pushes back on the trailing implement, (3) the rolling resistance of the trailing implement, and (4) the incline of the land on the side of the tractor to which the implement is offset. In the case of straddle harvesting of hazelnuts, soil/crop pushback occurs when the plants are forced into the tunnel, plants scrape along the sides of the tunnel, plants strike the shaking system, plants force the collector plates open, and the machine frame strikes the ground. The only effective way to counter side-draft when it is affecting a pull-type implement is to power (i.e., drive) the wheels of the trailing implement. The drawback of this approach is that a certain amount of energy gets wasted, and perhaps a bit of turf gets torn up (and unnecessarily so) when tractor ground speed and the ground speed of the driven pull-type implement do not independently match.

Given that any pull-type hazel combine will likely require one or more hydrostatically-driven wheels to prevent side-draft, the cost advantage of a pull-type harvester over a self-propelled is measurably reduced. Other advantages of a self-propelled over a pull-type combine include:

1. Maneuverability. A pull-type unit requires a much wider headland to align the harvester with a plant row prior to engaging the crop. Self-propelled units can be designed with a very tight turning radius.
2. Horsepower. A pull-type hazel combine will likely require a tractor with at least 70 hp (with 100+ much more preferred). Not all farmers have such a tractor.
3. Orchard Suitability. Most tractors are not suitable for orchard conditions as the ideal tractor for a pull-type hazel combine would have a lower profile, feature turf tires, and be narrow enough to move between hazel rows without damaging the plants or the tractor.
4. Transport. A self-propelled piece of equipment is typically easier to load on a transport trailer than a large, offset pull-type harvester.
5. Operator View. A self-propelled combine can be designed so that the operator is ideally located for proper maneuvering of the unit. This is likely to be at a location that is at the front of the machine and at a low enough position to see the base of each plant entering the harvester.  When harvesting single-trunk trees with large canopies, the operator compartment must be below the tree canopy if there is not an automatic guidance system to keep the machine centered on the tree row,
6. Control Accessibility. Controls for pull-type units are either on the harvester where they are not accessible to the operator, and/or in a box that must be clamped to the tractor when hitching up to the harvester. It follows that such controls are not ideally located for operator use/comfort like they would be in the compartment of a self-propelled machine.
7. Alley Cropping. When other crops are grown between hazel rows, it is highly desirable to minimize/avoid driving on these crops. Eliminating a tractor likely minimizes the width and/or number of wheel tracks associated with harvest.

Wheel Numbers

Once the decision to switch to a self-propelled harvester was made, the advantages and disadvantages of a 3- versus a 4-wheel harvester were examined. Three principle factors led to a decision to go with a 3-wheeled machine (1 front, 2 rear): simplicity, cost and maneuverability. Simplicity is associated with the realization that three points determine a plane, and thus when you have three wheels, each of them is in contact with the ground regardless of what surface you are on. With a 4-wheel machine, you must add special linkages, connections and/or suspensions to keep all four wheels on or near the ground, especially when traversing rough terrain. Not only are these special linkages, connections and/or suspensions costly, they get more complex when design specifications include the ability to independently raise/lower each wheel. With respect to cost, we were told during a 2022 visit to a fruit harvester equipment manufacturer in Oregon that the cost for them to go from three to four wheels on a particular harvester was around $30,000. As far as maneuverability is concerned, suffice it to say that it’s much easier to obtain a tight turning radius with a single, driven, steerable wheel than with two, driven, steerable wheels.

Wheel Orientation

Each hydrostatically driven wheel requires a planetary gear hub, a hydraulic drive motor to turn the planetary, and associated hydraulic hoses/plumbing. On every known fruit harvester in existence, these items are on the inside of the wheel, that is, on the side of the wheel that does not face the outside of the machine. Also on the inside of the wheel on these fruit harvesters are the components and structural framework required to raise and lower the wheel. The space that is taken up by these items (specifically, the drive motor, associated hydraulic hoses/plumbing, and the components and structural framework required to raise and lower the wheel) forces the wheel outward from the picking tunnel. When the width of this space is in the neighborhood of one foot, the minimum side-to-side distance between the tires on opposing sides of a harvester with a five foot wide picking tunnel and 3 or so inch wide tunnel walls would be 7.5 feet. This wheel spacing presents a problem when the goal is to transport the harvester on a trailer with a width no greater than 8.5 feet.

To address the wheel spacing issue, a decision was made to flip the wheels 180 degrees; that is, to position wheels so that the drive motors and wheel raising/lowering components are outside the wheels. This enables the tires to be placed within a couple inches of the picking tunnel sidewalls, and also provides better service access to hydraulic drive motors and wheel raising/lowering components. The downside of this is that to change tires requires (A) wheel access through the picking tunnel walls, or (B) that the machine be raised up high enough (and then blocked in place) to provide wheel access below the picking tunnel and the attached cluster collection platform.  Of these two design options, raising the machine for tire changes was felt to be the most attractive.

Machine Raising/Lowering

Incorporating the ability to independently raise and lower each wheel was one of the initial design specifications for the hazelnut combine for three primary reasons. First, it enables the machine to be leveled on sloping terrain. Such leveling not only provides stability, but it also helps ensure that the picking tunnel remains parallel to the vertical axis of the plant, which in turn, provides more uniform shaking. Second, it enables the machine to be raised for the harvest of taller plants. Without the ability to raise the machine in the field, the harvestable height of plants is largely restricted by over-the-road transport height limitations. Third, the machine can be raised to keep the collection platform frame from plowing through material surrounding the base of plants. This is needed in fields where plants have been planted into soil mounds, or where mulch has been added around the base of plants.

For our machine, we selected a maximum lift height of 30 inches. Cylinders with a 30 inch stroke are readily available/stocked and hence less expensive.  Also, 30 inches is about the maximum lift height found on commercially-available fruit harvesters with the ability to independently raise/lower wheels (note that greater the lift height, the greater the likelihood of a roll when wheels are improperly raised on steep terrains). We have also added an automatic self-leveling system into the design – a feature that is easier and less costly to incorporate into a 3-wheeled (versus a 4-wheeled) machine,

Tire Selection

With a maximum lift height of 30 inches, it was decided to shoot for a tire with a diameter not much larger than 30 inches to better facilitate removal of the wheels from under the picking tunnel when the machine is fully raised.  Additionally, the larger the tire diameter, the less available room around the tires for other components when the machine is in the down position,

Tire selection was also tied to a design goal of minimizing soil compaction and other damage to the orchard floor. This goal is largely obtained by (1) minimizing overall machine weight, (2) using flotation tires (i.e., tires designed to provide a larger footprint for weight distribution), and (3) incorporating a tire tread design that balances traction needs with the need to minimize surface damage and/or deformation. Under the assumption that the self-propelled combine (as it was being planned) could weigh as much as 14,000 lbs when loaded, we searched for flotation tires with a load rating of at least 5000 lbs, a tread more common to turf tires (TRA code R-3), and, as previously noted, a maximum diameter not too much greater than 30 inches.

After scouring agricultural tire data, it became apparent that common flotation tires with diameters under 30 inches have, at best, load ratings approaching 3000 lbs. A BKT tire representative was contacted for tire selection assistance and they suggested their RIDEMAX FL 693 M which is a 10-ply radial tire designed for agricultural trailers and tank trucks used mainly for on-the-road transport. The tire has an overall diameter of 31.5 inches, a static loaded radius of 13.5 inches, a section width of 15.7 inches, and a load rating at 5 mph of 5020 lbs at 35 psi and 9640 lbs at 87 psi. At an operating speed of 45 mph the tire has a load rating of 2550 lbs at 35 psi and 4888 lbs at 87 psi. Tire size is listed as 400/45 R 17.5. This tire has a hybrid tread that is somewhat characteristic of a turf tire, but with a bit more traction and self-cleaning capabilities. A set of three RIDEMAX FL 693 M tires were purchased, and special rims were fabricated to match the planetary gear hubs being used. Once the tires have been installed on the completed harvester, a local tire specialist will determine (via visual observation) the proper inflation pressure for each tire in order to maximize traction and minimize soil compaction and other damage to the orchard floor.

Powertrain

The decision to move to a self-propelled machine brought with it the need for a powertrain; in this case, an engine, variable-speed hydrostatic pumps, wheel drive motors, and planetary gear hubs. On-line searches and visits to salvage yards were conducted in an attempt to locate a used machine that contained all these elements. On-and-off searching for a couple months (primarily focused on self-propelled windrowers and swathers) failed to turn up a reasonably priced machine. On a suggestion from Tim Osterhaus, a visit was made to Scott Hacker Equipment in Lexington, Illinois -- a company specializing in used corn detasselers. There we purchased a self-propelled unit produced by Louks Manufacturing Company of Gilman, Iowa (now defunct). From this unit we salvaged a Ford 300 Industrial In-Line 6 engine, two Danfoss Series 40 M46 variable-speed axial piston pumps, and four Auburn Model 6 Power Wheel® planetary gear drives with Eaton hydraulic motors. At an engine speed of 2000 rpm, each M46 piston pump has a maximum output of 24 gallons per minute. On the Louks detasseler, one of these pumps was fully dedicated to propulsion (i.e., the four wheel-drive motors), and the other pump could be switched from providing propulsion (i.e., it was used to double the ground speed of the detasseler) to powering the detasseling motors.

Shakers

Between 2018 and 2022, four different mechanical harvesters were used by UW-Madison researchers to remove hazelnut clusters. Based on an evaluation of their performance, it was felt that a rotary shaker system would be the best shaker system to install in a prototype hazelnut combine, this since rotary shakers have the ability to reach into a plant and make contact with almost all branches regardless of branch orientation. This capability is of increased importance given the fact that the machine will be used to harvest plants that will differ in height, width, shape and branch structure because of differences in genetics and/or the manner in which they have been managed (i.e., grown naturally as bushes, maintained as single-trunk trees, or pruned to something in between these two extremes).

From a design perspective, rotary shakers are relatively complex, with designs optimized through years of use and experimentation. In full realization of this, Littau Harvester of Stayton, Oregon was approached about providing one of their vertical axis, freewheeling rotary shaker systems for the hazelnut combine. For the past couple years, Littau has successfully harvested hazelnuts near their manufacturing facility with their over-the-row harvesters, proving the ability of their shakers to remove hazelnut clusters.

Littau recommended two of their shaker systems: the HHDX unit with an overall height of 88 inches and the TF-14 system with an overall height of 116 inches. These heights include the framework that enables the entire shaker assembly to be pushed outward to avoid snapping a shaking rod when its end makes direct contact with a trunk or large branch. The shorter HHDX shaker was selected, and both shaker frames were attached such that each could be independently raised/lowered with its own 24-inch stroke hydraulic cylinder. With young, bushy plants, it is advantageous to run the shaker as close to the collection platform as possible. As plants grow, it’s advantageous to move the shaker up with the plant canopy, and thus away from the base of thicker shoots/trunks where there are fewer nuts, and contact with the shoots/trunks are more likely to dampen shaking and damage shaking rods.

The diameter of a HHDX shaker (sans support frame) is 61.5 inches, and in Littau machines, the left and right shakers are set up with a minimum and maximum tip-to-tip gap of 1 inch and 20 inches, respectively. This equates to a maximum “shaker system width” of 20 + 61.5*2 = 143 inches.

Machine Size and Relative Location of Major Components

With the major components (i.e., engine, main pumps, shakers, tires, thresher, drum-in-drum sorter) identified, the overall layout of the machine and machine sizing was undertaken.

Factoring into machine sizing were over-the-highway transportation requirements.  In almost all jurisdictions, an oversize load is a load that exceeds a width of 8.5 feet, and/or a height of 13.5 feet, and/or a length (for a single trailer) of 53 feet, and/or a weight of 80,000 lbs. Oversize loads generally require special permitting, lighting and/or signage depending on what is being transported, what limit is being exceeded and the route being traveled.  With respect to width, one additional caveat (albeit an important one) is that pilot car services are generally required when the load’s width exceeds 12 feet.

Based on these transportation requirements, the maximum height of the combine (in its transport position) was fixed at 10.5 feet so that the 13.5 foot height would not be exceeded when the combine is transported on a flat-bed trailer with a maximum bed height of 3 feet.  Additionally, the maximum width was set at 142 inches to avoid exceeding the 12 foot transport width that could trigger the need for a pilot car in some areas. Although the 142 inches is less than the Littau HHDX maximum shaker system width of 143 inches, this is not an issue since shakers are not being forced outward during transport, and if need be, they could be locked in a position that holds them closer to the center of the machine during transport.

For symmetry purposes, it was decided to place the 60-inch wide tunnel in the center of the 142 inch wide frame.  This always means that the distance from the outside of the picking tunnel to the outside of the machine is 41 inches on each side. It also guarantees that when the shakers are swung to the Littau maximum outward position, they will extend beyond the machine’s frame the same distance on each side.

To accommodate the shakers, their frames, their lifting/lowering mechanisms, and their need to swing, 78 inches of machine length were dedicated to the shakers.  This portion of the combine’s length is herein referred to as the shaker section.  Because of the two shakers, there is a left side shaking section and a right side shaking section.

The next design decision was to establish the distance between the front of the machine and the shaker sections.  Establishing this length was a compromise between two competing goals: the need to keep it short to minimize overall machine length for easier transport and maneuverability, and the need for sufficient distance between the front of the machine and the shakers so that plants are well into the tunnel before they contact the shakers (if the distance is too short, shaking will cause more material than desired to fall onto the ground in front of the machine instead of onto the collection platform).  In the end, the distance from the front of the shaker sections to the front of the machine was fixed at a somewhat arbitrary value of 64 inches.

The areas in front of the shakers were reserved for the single front tire and the operator compartment.  Because it is somewhat standard to place operator compartments on the left side or center of self-propelled machines, it was decided that the left front section would house the operator’s compartment, and that the entire right front section would house the single front tire and its steering cylinder and linkage.

From first inception of a three-wheeled combine, it was envisioned that the storage bin – the component with the largest fluctuation in mass - should be placed between the wheels on the side of the machine with the two wheels, this so the fluctuations in bin content would have the least overall impact on weight distribution to the three wheels.  The bin was thus located immediately behind the right shaker section.

Because of the space taken up by the storage bin on the right side of the machine, it forced (because of length needs) the engine with its attached pumps to the left side. This in turn meant that the rotary air-lock, threshing unit, and drum-in-drum sorter (which are all stacked one above the other to facilitate material flow) had to be placed rearward of the storage bin on the right side.

With the engine and main pumps on the left side, it was logical to also place the fuel tank, oil reservoir, oil cooler and oil filters on the left side.  Additionally, it made sense to locate the vacuum fan on the left side, nearest to the main pumps since it has the hydraulic motor with the highest required flow.

After completing a rough layout with components in their approximate and relative location to each other, an overall machine length of 22 feet was selected.  This equates to a length distance from the rear of the machine to the rear of the shaker sections of 10 ft 2 inches.  Although a shorter overall length is achievable via crowding various components, it was felt that for a prototype, the longer length left more flexibility for future changes/adjustments.

Tire Location and Steering

A couple competing design goals were to (A) keep the wheels in as far as possible so that the machine could be driven on a standard flat bed trailer which has a width of 8.5 feet, and (B) keep the wheels out far enough for stability and to make room for vacuum lines and other items between the wheels and the tunnel.  After some study it was decided to center each rear wheel 48 inches from the middle of the machine, and to center the single front wheel 51 inches from the middle of the machine.  Although this means that the wheels will overhang a standard flat bed trailer, this was not felt to be a problem given that individuals wheels are 16 inches wide.

The left rear wheel (i.e., the single rear wheel on the left side of the machine) was designed to be adjustable front to back from a distance of between 63 and 81 inches from the rear of the machine.  The right rear wheel was selected to be placed 63 inches from the rear of the machine.  Without a complete machine design, the weight distribution to the wheels is a large unknown, so front-to-back wheel locations were very much a guess.  The right front wheel was located 228 inches from the rear of the machine, resulting in an overall wheel base of 228 inches – 63 inches = 165 inches = 13 ft 9 inches.  Like the rear wheels, the relative front-to-back location of the front wheel was somewhat random.  Given as it is the only front wheel, the mindset was to keep it as forward as possible to reduce its vertical load (i.e., to transfer more load to the rear tires). 

In addition to an unknown weight distribution, making the front-to-back location of the left wheel adjustable was done because of concern that placing the left wheel too far back on the left side of the machine could result in greater unwanted vibration of the left front of the machine, and/or greater unwanted vertical deflection of the left front of the machine relative to the right side (a situation which can cause the collection platform on opposite sides of the machine to interfere with each other).  The right rear wheel was not made adjustable front-to-back because moving it forward to align with the left wheel reduces the volume of the collection bin. Also, keeping it in a fixed position simplifies the design and lowers the cost and weight of the machine.  The impact of offsetting the rear wheels on the overall steerability/maneuverability of the 3-wheeled machine is unknown, and thus is something that can be evaluated with this experimental machine.

Locating the right front wheel 3 inches further from the middle of the machine than the rear wheels was required so that the wheel would not interfere with the tunnel frame during steering. Note here that it was decided to have the pivot point for the front wheel directly over the top center of the wheel.  By moving this pivot point slightly to the outside of the tire, it would have been possible to move the front tire closer to the middle of the machine (from 51 inches to closer to the 48 inches of the rear tires) and still not have the tire interfere with tunnel framing.  However, since the impact of moving the pivot point (from being centered over the tire to a point outside the center) on overall steering forces and overall maneuverabilit was unknown, it was decided to keep the pivot point centered over the tire.

One of the design goals was to provide a fairly tight turning radius.  More specifically, the steering was designed so that when both rear wheels where located the same distance from the rear of the machine (i.e., each were 63 inches from the rear of the machine), the center for the machine’s turning radius would be located at a point 63 inches forward of the rear of the machine and 6 feet to the left of the machine for a left turn, and at a point 63 inches forward of the rear of the machine and 6 feet to the right of the machine for a right turn.  To achieve these turning centers requires that the front wheel rotate 53.3 degrees when turning left, and rotate 82.7 degrees when turning right.  Thus, to go from a far left to a far right steering position requires the wheel to turn 136 degrees.

For this project, it was decided to use the hydraulic steering motor and cylinder used on the corn detassler. This was partly due to the feeling that the steering motor-cylinder combination were likely sized to provide a reasonable “rate of steering”. The hydraulic cylinder used for steering on the corn detassler was a tie-rod cylinder with a 3 inch bore and 8 inch stroke. 

The mechanism for raising and lowering the front wheel involves telescoping a 4.5x4.5 inch square tube inside a 5x5 square tube.  Attached to the top of the 5x5 square tube is a round vertical shaft.  The 5x5 tube with the attached shaft is herein referred to as the steering column, and it must be rotated to turn the front wheel. Although this steering column could be rotated by directly attaching the hydraulic steering cylinder to the end of a steering arm affixed to the steering column, such a design results in widely varying forces on the steering column as the cylinder is extended and retracted. To avoid the shortcoming associated with direct attachment of the cylinder to the steering arm, the cylinder was attached to a special pivoting control arm, which in turn was attached to the steering arm with a special connecting link.

Cluster Collection System

The term “collection platform” is herein used to refer to each of the two assemblies (one on each side of the tunnel) at the base of the machine upon which the hazelnut clusters fall, are collected, and then conveyed away for further processing

The most critical element of each collection platform is a set of collector plates.. Each plate rotates independently toward the rear of the machine as it is pushed backward by a passing plant. As the plant passes, the plate snaps back into its original position, thereby closing up the area in front of the plant and minimizing loss of crop to the orchard floor. In their non-rotated (i.e., closed) position, these plates are angled upward toward the center of the picking tunnel. This slope, as well as the opening and closing action of surrounding plates, and the wiping action of passing branches, moves the crop off a plate and to the side of the picking tunnel where it is conveyed to storage.

Two fundamentally different methods are used to produce the rotational action of a collector plate. The traditional method is to attach a pivot pin to the collector plate. This pivot pin passes through a sleeve that is welded into the frame in such a way as to achieve the initial desired collector plate angle and the plate’s plane of rotation. A tension spring is used to snap the collector plate back into position after it is no longer in contact with a plant. A more recent method of obtaining the rotational action of a collector plate involves the use of a molded mount that consists of a rubber block adhered between two metal plates. One metal plate is rigidly attached to the collector plate and the other is rigidly attached to the frame. The rubber block acts both as the plate’s center of rotation and as a torsional spring. The plan was to use molded mounts in the hazelnut combine, however the mount is expensive and in laboratory tests conducted with the mount, plate rotation appeared limited to 45 degrees. The less rotation a plate has, the longer the plate must be to properly perform its job.

Conveyance of soft fruits to storage is accomplished with either flat (generally ribbed) belts or bucket conveyers. Bucket conveyors have an advantage over ribbed belts in that they can move more product per conveyor length, and they can move product up steeper slopes. Conversely, they generally have more moving parts and a higher initial cost. Bucket conveyors are essential where an easily damaged product must be moved vertically upward for processing, storage, and/or transport to the opposite side of the machine.

As hazel plants age, a greater number of old and broken branches are knocked off by shakers. During harvesting research, machines that utilized bucket conveyors to move material from the collector plates to the rear of the machine and then straight upward, would frequently experience a build up of sticks at the horizontal to vertical transition of the conveyors. Plugging with sticks was also a problem in machines with belt conveyors where the conveyors passed through narrow openings or where the conveyors had a combination of a low side and a steeper slope.

To simplify design, lower overall machine costs, and combat problems with sticks, a decision was made to employ screw conveyers to move material from the collector plates to the rear of the machine. Screw conveyors are routinely used to move cereal grains, shelled corn, soybeans, coffee beans and other more resilient crops. A single rotating auger is the only moving part.

A hand-cranked conveyor featuring a 9-inch diameter auger was assembled to test the ability of a screw conveyor to move hazelnut clusters loaded with sticks. Welded to a 4.5 inch diameter opening in the end and side of the conveyor was a tube through which clusters were vacuumed out of the conveyor. The end of the conveyor was left open. For testing, the conveyor was loaded with a mix of hazelnuts and sticks. Vacuum was generated with a Hasatsan H2200 hazelnut harvester. The entire system worked extremely well. Sticks were either broken up by the auger or they rode along the top of the auger until they fell off the end of the conveyor.

The Hasatsan H2200 uses a 20 inch diameter fan spinning in a 28 inch diameter housing at a typical operating speed of 3000 rpm. The current plan is to incorporate a similar fan into the prototype hazel combine to draw hazelnut clusters from the rear end of the screw conveyers into a vacuum chamber at the top of the machine. Located inside this chamber will be a 7-in diameter or smaller auger that drops clusters into an air-lock that then feeds the clusters into the threshing unit.

The final design of each collection platform consists of 27 collector plates over a twelve foot length.  These plates are mounted to a twelve foot long “collector deck” which is hinged to the “main collection platform weldment” such that it can be locked at different angles, therefore enabling a variety of collector plate tilts.

The main collector plate weldment is a 12-foot section consisting of the housing for a 7-inch diameter auger, rectangular and a round metal tubes that run parallel to the length of the machine (and protect the bottom and inside of the platform from damage), and a series of gusset plates that (1) form a rigid structural assembly between/with the tubes and auger housing, and (2) hold the shape of the auger housing.

The main collector plate weldment contains holes that are used to bolt the entire collection platform to the bottom of the machine such that the entire assembly can be quickly removed, and if need be, replaced with an alternative collection system.  This was a feature incorporated because of the experimental nature of the machine.

Gathering Heads

The title “gathering head” was given to the weldment bolted to the front of each collection platform.  These heads serve several functions. As designed they (1) protect the auger motor from dirt/damage, (2) hold the auger motor in place, (3) protect the front of the collection platform from damage/dirt, (4) enable quick maintenance/access to the auger motor, (5) pick up downed plants and channel them into the tunnel, (6) feed low branches up into the shakers, (7) handle forces applied by plants and the ground, (8) facilitate routing of auger motor hydraulic hoses, (9) provide for tire rotation (right side) and cab movement (left side), (10) bolt-on to facilitate easy replacement/redesign, and (11) contain a blunt front edge to minimize shearing of plants.

Design of the gathering heads was one of the more difficult tasks largely because of their need for proper structural support, and the fact the wheel on the right, and the cab on the left restrict placing support framing in their respective regions. Also impacting this support problem was the decision to have a more gradual entrance to the tunnel sides to help minimize clusters getting stripped off before entering the tunnel. Specifically, instead of a sharp, 90 degree angle between (1) the sheet metal on the front of the machine, and (2) the sheet metal lining the tunnel, a surface was added at a 45 degree angle to both of these surfaces. The intersection between this 45 degree surface and the front surface was located at a point 11 inches from the tunnel surface.  Likewise, the intersection between the 45 degree angle and the tunnel surface was located 11 inches from the front of the machine.

Bin

Early in the design process, a measurable amount of time was dedicated to the design of a dump bin.  Work to add a dump bin to the machine was undertaken because of the trend in the farming industry to the bulk handling of agricultural produce via semi-trailers and other large containers. This trend is no different in the Oregon hazelnut industry. As of January 2023, 42 self-propelled Monchiero nut harvesters were being used in the Willamette Valley of Oregon to collect hazelnuts. The Monchiero harvesters sweep nuts off the orchard floor, separate the nuts from most of the dirt collected with the nuts, and then convey the nuts into a dump bin at the rear of the machine. This bin is emptied by first raising up the bin, then backing the harvester up to the receiving vessel (or driving the receiving vessel under the raised bin), and then dumping the bin. One shortcoming of the Monchiero units is that it can not dump in taller trucks/trailers. This shortcoming resulted in the special development of a 215 cubic foot, low-profile dump cart by Hillco Technologies, Inc. of Nezperce, ID. This dump cart, with a maximum dump height of 14 feet, is used to transfer nuts from the Monchiero unit to taller trucks/ trailers.

The dump bin designed for the hazelnut combine utilizes a cart with three sets of rollers set in special tracks that guided the cart as it was moved up and down by a single, 3.5 in dia bore x 60-inch stroke, double-acting lift cylinder.  Resting inside, and hinged to the cart was a hopper that via a unique 4-bar linkage system (connected to a single, 10-inch stroke, double-acting tilt cylinder), was capable of rotating the bin through 140 degrees at any height. When fully raised and tilted, the leading edge of the hopper is approximately 8.5 feet above ground with the machine in a fully lowered position, and at a maximum height of 11 feet above the ground when the machine is fully raised.  With side dumping, the machine can be driven alongside the receiving vessel, or the receiving vessel can be driven alongside (parallel to) the combine.

In the end, incorporation of this dump bin into the combine was scrapped.  This was due to four primary reasons.  First, to be able to dump a bin outward at any height means that there can be no fixed, continuous structural framing members outside and above the bin, nor in the space occupied by the bin.  While such restrictions on structural framing can be overcome, they do come with a higher material cost, and generally an increase in machine weight.  Second, the mechanisms required to lift and tilt the hopper take space, and hence reduce potential bin capacity.  Third, a moveable bin increases the complexity of the connection bringing nuts into the bin (note that nuts are blown into the bin through a hole near the very top of the bin side). Fourth, a dump bin located immediately behind the shaker section removes/blocks the most convenient route for hydraulic hoses running from the right rear to the front right of the machine.

In short, adding the dump bin added a level of complexity to the design, and due to the fact that it was not a critical need in this experimental/test machine, the KISS design principal won out.  However, as a result of scrapping the dump bin (in favor of a stationary bin), two other features were added to the machine.  First, the lower portion of the stationary bin was designed so that it could be quickly removed and replaced with an alternate configuration – a configuration that could support, for example, an unloading system that utilizes the capabilities of the machine to move nuts with air.  Second, the switch helped drive the development of the main frame as a series of bolted-together, smaller subassemblies. With this subassembly system, a dump bin could be reincorporated into the design by simply substituting the frame subassembly containing the stationary bin, with a new frame subassembly featuring a dump bin.

It is fully anticipated that at some point, hazelnuts grown in the Upper Midwest will be handled at the farm level in the same way as cereal grains, shelled corn and soybeans; that is, they will be moved via semi-trailers and large gravity wagons to local agricultural marketing cooperatives or special receiving stations where they will be dried, temporarily stored and then moved to a processing facility.  When this occurs, a dump bin will be a highly desirable feature on combines.

Hydraulic System

The combine will be powered by three hydraulic pumps. One Danfoss Series 40 M46 pump operating in a closed loop system will be fully dedicated to turning the three drive wheels, the second Danfoss Series 40 M46 pump operating in a closed loop system can be switched between (A) powering the material transport fan motor during harvest, and (B) powering the drive wheels for an increase in ground speed during non-harvest activities. The third pump, with an output of 30 gallons per minute and operating in an open loop system, will be used to power the two shaker motors, three screw conveyer motors, the airlock motor, the thresher motor, the winnowing fan motor, the steering motor, and the drum-in-drum sorter motor. The third pump will also be used to power all hydraulic cylinders. This includes the three cylinders used to raise/lower each of the three wheels, shaker raise/lower cylinders, a cab lift cylinder, the two cylinders used to change the rotatable front cross beam (discussed later) between transport and harvest modes, and if someday needed, bin dump and tilts cylinders.

Movement of all cylinders will be electronically controlled from the operator cab, and all motors powered by the third pump (except the steering motor) will be switched on and off from the cab. The speed of pumps 1 and 2 will also be controlled from the cab.

Providing major assistance in the design of the hydraulic system and in the securing of hydraulic components has been Adam Goudreau, a Field Sales Representative for Berendsen Fluid Power.

With respect to hydraulic motors, Censsa 20.3 cc/rev hydraulic motors have been obtained for each wheel and the material transport fans. Other hydraulic motors planned for use are a Char-Lynn T Series 244 cc for the collection platform augers and the cross auger in the vacuum chamber, a Char- Lynn T Series 102 cc for the airlock, a Char- Lynn T Series 80 cc for the thresher, a Char- Lynn T Series 370 cc for the drum-in-drum sorter, and a Danfoss Group 1 gear motor for the winnowing fan. 

All hydraulic valves have been ordered.  An tank oil filter assembly and tank filters, tank clean-out cover, the oil cooler and cooler fan, a suction filter head and suction filter, a filler breather, temp/float switch, pressure gages, and level gauges have been obtained along with spools of hydraulic hose and numerous hydraulic fittings.

Operator Compartment

As previous noted, the 3-wheeled machine has been laid out with the single front wheel (which is both driven and steerable) located on the right side of the machine, and with the operator compartment located in the front left of the machine (i.e., directly across from the front wheel). 

Many over-the-row harvesters place the operator atop the machine, directly above the plant row.  While this location provides an ideal line-of-sight for the harvest of shorter crops like grapes and blueberries, it can be a problematic location when harvesting taller plants.  During some of the hazelnut harvesting trials conducted at UW-Madison, plants were so tall that the operator’s view forward from atop the machine was completely blocked, forcing the driver to rely on a person walking alongside the harvester for steering adjustments.  To this end, to properly align an over-the-row harvester for closely spaced plants, the machine must have auto-steer and/or the operator must be able to see the base of the bush or tree as it enters the picking tunnel.  Suffice it to say that the latter is difficult unless the operator is positioned below the plant canopy.  Two other reasons for not placing an operator compartment with a roll-over-protection structure (ROPS) atop a harvester are (1) it restricts picking tunnel height if it can not be lowered for over-the-road transportation, and (2) it requires up-and-down climbing.

The ideal vertical location for an operator compartment for nut harvest is believed to be as close to the ground as possible such that there is an unrestricted forward view under the plant canopy when harvesting single-truck trees.  Multi-shoot plants with branches that extend or sprawl outward from the plant base are likely to partially or wholly block the forward view, but should not interfere with the operator’s side view of plants entering the picking tunnel.

To enable the operator compartment to be as close to the ground as possible, a hydraulic cylinder on the prototype combine will be dedicated to raising and lowering the compartment.  When fully extended, this cylinder will allow the compartment to continually ride on a pair of spring-loaded caster wheels, even when the rest of the machine is raised and lowered.  In addition to providing a more cushioned ride, this feature helps dampen the compartment from machine vibrations, and ensures that the mass of the compartment and operator do not negatively impact harvester stability.  This same cylinder can be retracted to move the compartment upward, to give the operator a view over the top of smaller plants.  In such raised positions, the compartment will not be as well isolated from machine vibrations.

Since: (1) space limitations require the operator compartment and front wheel to be on opposite sides of the machine, and (2) stability needs require the bin to be on the same side as the front wheel, it follows that the operator compartment and bin will be on opposite sides of the machine.  This means that the operator can not directly observe bin alignment with receiving vessels during unloading.  For this reason, an LCD display attached to four different cameras will be mounted in the operator compartment.  One of these cameras will be used to monitor the bin discharge area.  It is not yet known where the other three cameras will mounted.  Options include a back-up camera, a camera to monitor nut level in the bin, a camera monitoring the rear end of the screw conveyers (i.e., point where clusters are sucked out of the conveyors), a camera monitoring the drum-in-drum sorter and aspirating chute, and a camera pointed forward atop the machine.

Frame

By far and away the most difficult and critical harvester design element is that of the frame. No other element controls the overall cost of the machine more than the frame. If elements of the frame are over-designed, or a portion of the frame geometry is not efficiently structured, the frame will have excess mass. Frame mass is directly proportional to its cost. In general, when you double the mass of a frame, you double its costs. Additionally, the more mass a frame possesses, the larger and heavier must be the tires and the greater the power required to propel the machine (which in turn can drive the need for a heavier and more expensive powertrain). Soil compaction and damage to the soil surface also increase with frame mass, and heavier machines are generally more difficult and expensive to transport.

As planned, the hazelnut harvester has six major elements that add to frame complexity:

1. Large picking tunnel. The only way to tie one side of an over-the-row harvester to the other side is via framing members that cross over the top of the picking tunnel. The higher the top of the picking tunnel, the further this cross framing is from the ground, and the greater the total bending moment induced in the cross framing by forces working to change the spacing between the separate halves of the machine.
2. A more open front structure. Most over-the-row harvesters have a fully enclosed picking tunnel. The top of this tunnel will bend tall plants before they reach the shakers, and also dampen the vibration of the plants as they move through the shakers. A primary design objective is to eliminate this contact as much as possible until plants have moved past the shakers. This requires that the top of the tunnel and any framing across the front of the machine be removed or moved upward so as not to contact plants. For the hazelnut combine, a rotatable front cross frame was included in the design.  This “high clearance” beam ties together the two sides of the machine at a point just in front of the shaker sections.  It is hinged at four locations and lowered for transport with two hydraulic cylinders that are activated via a switch in the operator compartment.
3. Three-wheels. An over-the-row harvester with three wheels has two wheels on one side of the row and one on the other. This asymmetric machine support induces front-to-back torsion in the overall frame due to the mass in the corner of the machine without wheel support. To help minimize this torsion, the operator compartment (which occupies the space that would normally be occupied by a fourth wheel) will be lowered to the ground where it will be supported by its own caster wheel during harvest.
4. Machine lift system. When each wheel on the combine is completely lowered, the harvester will be raised 30 inches. This increase in height adds to the total bending moment that can be potentially induced in the framing system crossing over the top of the picking tunnel. Additionally, the bending moment induced in the vertical support attached to a wheel, will increase linearly with an increase in the distance the wheel is lowered (i.e. the machine is raised).
5. Large rotary shakers. No fixed framing members can cross through the area occupied by rotary shakers that are free to move-side-to-side and that are designed to be raised and lowered while in use. This means that all structural framing must be located under, outside, and above the shakers.
6. Side dump bin (if/when included). As previously noted, to utilize a side dump bin, there can be no fixed framing member outside or above the bin as these members would interfere with the ability to raise and tilt the bin, respectively. The inability to utilize the area outside and above the bin, complicates frame design from the standpoint that the bin is right behind the right shaker, and the only place to place framing is above, below and on the very outside of the shaker.

The frame must be designed to support all unit operations, the powertrain, and the hydraulic systems. This equipment must be located and supported in such a way that it’s accessible for service/maintenance. Also, room must be provided for the safe routing of numerous hydraulic hoses and electrical lines.

The overall design approach for the combine frame was to use more of a space-frame concept with light gage steel angles for major axial-force members, and small diameter rods for X-bracing.  In many cases, the function of X-bracing was replaced with thin steel sheeting that also serves as machine covering, tunnel lining, tunneling flooring, and bin walls.  The most commonly used angle was a hot-rolled section with 2.0 inch legs and a thickness of 0.125 inches (2x2x1/8 angle).  The design strength in axial compression (c Pn) for this angle (with respect to the least radius of gyration) is 14.5 kips for an effective length (KL) of zero, 12.1 kips for a KL of 2 ft, 7.0 kips for a KL of 4 ft, and 3.3 kips for a KL of 6 feet. 

Pound for pound, tubes are typically a better structural option than angles for members subjected to bending and/or compressive loads.  However, it is frequently not possible to paint the inside of tubes, and this can be problematic when/where water gets into tubes, resulting in rusting (a major concern with very thin-walled tubes) and/or unsightly rust staining.  Water inside tubes is common when holes are drilled into the tubes to attach components.  Bolting thru a tube is generally not as ideal as bolting to an angle.  Bolting thru a tube typically requires a longer bolt and also means that you are drilling thru two metal surfaces (instead of one with an angle).  This shortcoming of tubes can be avoided by welding tabs to the outside of the tubes to which items are bolted (a move that tends to unnecessarily complicate design).

As initially roughed out, crude calculations indicated that 2x2x1/8 angle should be more than sufficient for most components.  It a couple cases, some thicker angles and angles with larger or smaller legs were selected.  Much component sizing and the overall layout off components was based on experience and a general feel for structural behavior.  A detailed structural analysis was not performed, largely due to uncertainty of how the machine would be loaded during normal use, non-normal use (e.g., getting pulled out off the mud), strapping/lifting during transport, jacking during repair/assembly, etc.

The initial plan was to build the main frame by bolting together three separate weldments: a left side frame, a right side frame, and a center frame.  The decision to break the frame into such sub-assemblies was due to concerns regarding fabrication of a single frame, specifically, the ability to accurately fixture a large assembly, move/orient it, and even paint it.  Additionally, the only place to weld up such a large frame would be at the West Madison Agricultural Research Station which is not an ideal location as it is several miles from the Agricultural  Engineering shop on the UW-Madison campus where virtually all part fabrication occurs.

It soon became apparent that separating the frame into these three weldments did little to alleviate the major fabrication issues in that the left and right side frame weldments with a 22 ft length, 9 ft 8 inch height, and 41 inch width, would still be too large to easily fixture, to maneuver during fabrication, and to have painted by Kuhn-North America.  It was subsequently decided to break each side weldment into four separate 41 inch wide sub-weldments: (1) a front frame 9 ft 8 inches tall and 5 ft 4 inches long, (2) an upper shaker section frame 14 inches tall and 78 inches long, (3) a lower shaker section frame also 14 inches tall and 78 inches long, and (4) a rear frame 9 ft 8 inches tall and 10 ft 2 inches long.  Outside of their overall size, the left and right rear frame weldments are quite different, as are the left and front frame weldments.  Conversely the left and right upper shaker section frames are mirror images of each other, as are the left and right lower shaker section frames.

The only shortcoming of breaking each side frame into four separate weldments is that it requires that hot-rolled angles intended to run the entire length of the sidewall (or a good portion of it) must be cut at locations between where sub-weldment join, and they must be effectively reconnected without a measureable sacrifice in function.  This concern basically disappeared when it was realized that a relatively small, simple, and structurally efficient method of connecting the ends of two angles could be made with a single bolt.  To connect the ends of two 2x2x1/8 angles, it was decided to use 1 inch diameter bolts.  A 2x2x1/8 inch angle has a cross-sectional area of 0.49 sq inches (1.65 lbs/ft at 0.284 lbs per cubic inch), and a 1 inch bolt with coarse threads has a thread root diameter of 0.86 inches and hence a minimum cross section area of 0.58 sq inches. To facilitate use of the bolt, the machined end of a short piece of structural tubing with a 1.75 inch O.D. and 1 inch I.D. is set flush with the end of each angle, and then the tube is welded to both legs of the angle.  Each tube is expected to be at least 3 inches long in order to provide plenty of weld to transfer load from the tubes into the angle (for an 1/8 inch deep weld, a 3 inch length provides 2 x 3 inch x 1/8 inch = 0.75 square inches of force transfer area).

One problem with this connection is that with the tube welded into the angle it’s not possible to use a wrench or socket to turn a 1 inch hex nut onto a 1-inch bolt, as the distance across the corners of a 1 inch hex nut and bolt is 1.732 inches (max), just under the diameter of the tube.  In order to tighten the nut on the bolt, the nut will be held in place with an open-end wrench, and the head of the bolt will be turned after plug-welding a 5/8-inch nut to the head of the bolt.  Before plug welding, it is beneficial to run a 5/8-inch drill thru the nut to enlarge the hole (this also removes the threads and all the zinc plating on those threads – something not wanted during welding).

One attribute or aspect of going to smaller sub-assemblies/weldments is that it generally results in a greater number of short framing members.  This is not necessarily a bad thing given that fabrication of sheet metal components in the Agricultural Engineering shop is restricted by the 6 foot shears, the 6 foot press brake, and the 54 x 60 inch size of the CNC plasma cutting table.

Considerable thought went into the design of each frame weldment, with many decisions supported by structural calculations, and almost all decisions documented.  This material will not be shared here, as they can be confusing without illustrations.  Instead, a couple main points are articulated in the following paragraphs.

The primary objective in design of the frames in the shaker section was to provide for as much up-and-down, as well as side-to-side (swinging) movement of the shakers.  To accomplish this, the upper and lower shaker sections frames form an opening height of 7 ft 4 inches and opening width of 78 inches.  This is enough to allow for the tips of the shakers to rotate outside of the 142 foot width of the machine.

From a structural perspective, the right front frame needs to hold the front wheel and its steering mechanism in place, and it must transfer loads imposed by the front wheel to the points where the front frame is connected to (1) right upper shaker section frames, (2) the right lower shaker section frame, and (3) the right end of the rotatable front cross frame. All forces acting on the front wheel must get transferred through the previously described steering column. This steering column connects to the front frame at two locations: at its top, and near its bottom.  At its top, the steering column slides into a 40 inch cross beam.  This cross beam must handle all vertical load that is transmitted between the front wheel and the machine (i.e., the weight on the front tire), and it must handle horizontal forces that are similar in magnitude to the horizontal forces acting on the tire.  The lower connection point does not transmit vertical load, but must handle higher horizontal forces than those acting on the cross beam.  The horizontal force acting on the tire is equal and opposite to the force applied to the tire by the soil.  This force is generally the traction force induced by the wheel drive motor.  In situations where the ground under the front wheel is soft, the rear wheels could help push the front wheel through the soil in which case the horizontal force would differ from the normal traction force.  Also, high horizontal forces can be induced when the machine gets stuck and is improperly pulled out of a stuck situation

Each rear wheel gets moved up and down by a “rear wheel” tube that is centered 17.5 inches from the center of the tire.  This tube is not subjected to any vertical force as that is resisted by the hydraulic cylinder that raises and lowers the wheel.  However, the tube must resist torsional forces induced by traction forces, and it must resist bending forces due to traction forces and the 17.5 inch offset.  Each rear wheel tube is held in place with an upper and a lower rear wheel bracket (RWB).  These wheel brackets must allow for vertical movement of the wheel tube, and the upper RWB must also hold the lift cylinder.  Design objectives were to keep these brackets lightweight, to minimize play/slop between the tube and bracket but yet provide relatively free sliding, minimize component wear, facilitate assembly, minimize space requirements, provide for front-to-rear adjustability of the left wheel, and avoid greasy surfaces.

In the final design, each upper and lower RWB consists of four separate weldments that are bolted together to form a rigid sleeve around the tube.  A rather simple assembly is facilitated by the manner in which these four weldments are bolt together.  More specifically, the rear wheel tube assembly can be slid or dropped into the machine frame while the frame rests on the ground, and then the front, back, and inside RWB weldments can be bolted in place around the wheel tube.

As previously noted, the rotatable front cross frame ties the two sides of the machine together at a location just in front of the shaker sections.  This cross frame is a fairly complex weldment, designed to house several hydraulic hoses and or wires.

At the rear of the machine, the two sides of the combine are connected with a center frame.  The top of the center frame lies in the horizontal plane that defines the top of the machine, and the back of the center frame lies in the vertical plane that defines the very back of the machine.  The center frame has a width of 60 inches (i.e., the width of the tunnel), a height of 2.0 feet, and a length of 70 inches.  Design objectives for the center frame were to keep it as simple and lightweight as possible, and to use it as an easily accessible housing for hydraulic controls – a housing that protects them (or the hoses connected to them) from UV radiation.  The frame must also house the cross auger whose housing needs to function as (1) a vacuum chamber, and (2) a space where the velocity material sucked up thru tubes from the collection platform is effectively stopped.

Detailed Working Drawings

Individual part drawings are being accumulated as fabrication progresses.  Once the build is complete, the combine will be detailed using Onshape, a computer-aided design (CAD) program that is delivered over the Internet and makes extensive use of cloud computing.