Banana Bales: Why Your Square Baler Makes Curved, Uneven Bales

By the XES Netting team — bale net-wrap manufacturer · Updated August 26, 2026 · ~15 min read

A man with a New Holland 273 posted the clearest description of this problem anybody has written:

My bales are coming out even twine to twin, but top( or bottom) is tighter than the other so they curve, and if they are picked up from the inside of the banana they fall apart even though they are baled tight.

— ACDII, HayTalk, "Banana bales, how to fix"

Read that first clause again — even twine to twine, but top tighter than bottom. He had already done the single most useful thing you can do with a bad bale, which is to work out which axis it is bending on. That one observation cuts the possible causes roughly in half before you pick up a wrench.

Almost nobody does it. The word "banana bale" gets used for two mechanically different faults — a bale short on one side left-to-right, and a bale short top-to-bottom — and they live in different parts of the machine. Ask about banana bales anywhere and you will get advice for both at once, plus a suggestion to add wedges, which is neither.

Contents

• Read the bale: which way does it bend?
• Count your strokes before you touch a wrench
• Feed rate: the cause behind most of them
• Hay dogs, and the half-slice mystery
• Cross-feed and packer forks: the backwards rule
• Feeder fork timing, with real numbers
• Why wedges are the last resort, not the first
• When it is not the baler at all
• The diagnostic order, in one list
• What banana bales actually cost you
• Frequently asked questions

---

Read the bale: which way does it bend?

Pick up a bad bale and orient it the way it came out of the chamber — strings on top. Now look at it end on.

Bending left to right

The bale is shorter on one side of the strings than the other, so it curves sideways. One side of the chamber is not getting its share of hay. That is a delivery problem: cross-feed, packer forks, feeder fork timing, windrow placement on the pickup, or simply not enough hay to push across.

Bending top to bottom

The strings are even, but the top of the bale is tighter than the bottom, or the reverse, so the bale curls like a wedge stacked front to back. That is a retention problem — the charge is not being held while the plunger works against it. That points at hay dogs.

Get left and right right

Before you follow anybody's left-side or right-side instruction, settle the convention. A member from Western Australia who has posted some of the most careful square-baler explanations anywhere on the internet stated it plainly:

The convention for left and right is to stand at the back of the baler to determine left and right side.

— Coondle, HayTalk, "Claas Markant banana shaped bales / wedges"

Two people describing the same fault from opposite ends of the machine will give exactly opposite advice and both be right. Settle it first.

Count your strokes before you touch a wrench

This is the highest-value thing in this article and it costs nothing. Ride along or open the back window, and count plunger strokes between knotter trips. The number tells you whether the baler is being fed correctly, and almost every shape problem starts there.

count the strokes between tying. Best firm bales will be 12 to 20 strokes per bale. That's what I would do before trying to adjust things in the chamber.

— DB Tracks, Camp Douglas WI · AgTalk thread 1211491

Note the second sentence. He is not offering it as one tip among many — he is putting it explicitly ahead of chamber adjustments. Other operators narrow the window further from their own machines:

count your strokes while baling and if the knotters don't trip between 12-16, your baler is hungry (need more hay in a windrow or to drive faster)

— mstuck21, HayTalk, "Banana bales, how to fix"

My old NH 276 liked 13-15 strokes per bale.

— RockmartGA, HayTalk, "Banana bales, how to fix"

DTN's long-running "Ask the Mechanic" column, working on a John Deere, targets 16 to 18 strokes per bale on that model and makes the same causal claim in the opposite direction: if you are baling too slowly, that alone will cause banana bales, because the feeder fork cannot get enough hay in front of it to push it all the way to the far side of the baler.

Why does stroke count map onto shape at all? Because of how thin each individual charge is. Coondle did the arithmetic:

Twelve flakes in a 36 inch bale means each flake averages 3 inches, doesn't take much difference in the effective feed under one of the two strings to give a banana bale.

— Coondle, HayTalk, "Claas Markant banana shaped bales / wedges"

A quarter-inch bias per flake, compounded over twelve flakes, is three inches of difference across the bale. Nothing has to be badly wrong for the bale to come out visibly bent. Fewer, thicker flakes make the arithmetic worse:

How many slices are you putting is a bale. The fewer the slices the more chance the bale will come out less than perfect.

— mike10, HayTalk, "NH 570 banana shaped bales, need help"

Feed rate: the cause behind most of them

Once you have the stroke count, the picture usually resolves. Hay pushes hay: the material already in the chamber is what carries each new charge across to the far side. Skimpy charges cannot do that job, so the far side stays light and the bale bends.

most often when you hear misshaped bales that equals uneven/too little of hay going in

— mstuck21, HayTalk, "Banana bales, how to fix"

The same member supplied the observation that proves it, and it is the most persuasive single data point in any of these threads. His John Deere 336 made beautiful straw bales and misshapen hay bales — same machine, same day, same settings. The difference was windrow size. When he started building hay windrows the size of his straw windrows, in his words a lot of problems went away, and it took a lot less time to bale as well.

That is a controlled experiment on one machine. It rules out the baler and points squarely at what you drove into it. Others say the same thing about specific models:

Older new hollands are notorious for that. Needs big windrows. Get it full and keep it full.

— F150Mustang, Southern MN · AgTalk thread 1211491

"Get it full and keep it full" is the whole feed-rate doctrine in six words. If your windrows are too thin to hit your stroke target at a sane ground speed, the fix is at the rake, not the baler — merge two windrows into one. Our rake selection guide covers what different rake types do to windrow shape and density, which is exactly the variable at issue here.

The last word on the balancing act belongs to a member who bales by ear:

Balancing crop density vs strokes vs ground speed is an art. I bale with the back window open so I can hear the strokes and ties.

— rjmoses, HayTalk, "Banana bales NH 570"

He also makes a density point worth carrying: a very light bale does not generate enough back pressure in the chamber to build the next one properly. He targets around 50 lb; under about 40 lb the bales stop holding their own shape. Bale weight is a symptom you can measure without tools.

Hay dogs, and the half-slice mystery

Hay dogs — also called hay retainers or bale stops — are the spring-loaded pawls along the top and bottom of the bale chamber. They let the bale move rearward and stop it moving forward, so the plunger has something to compress against. When they stop doing that, the bale creeps forward between strokes and the chamber loses its reference.

They are cheap, they are ignored, and they are the single most common mechanical cause of a top-to-bottom banana.

The hay dog should have no play up or down. If the hay dog can be moved then the springs are weak and need to be replaced.

— mike10, HayTalk, "NH 570 banana shaped bales, need help"

Check the rubber stop under each dog while you are in there — a missing or perished stop changes how far the dog can travel and puts the load in the wrong place. And check every dog, not the two you can reach easily.

Frozen is worse than broken

This is where ACDII's thread earns its keep, because he diagnosed it wrong first and then found the real fault. He started by pulling a bottom dog that was hanging and replacing the spring. The bales did not improve. Then:

Found the problem, it wasnt the bottom broken dog spring, but the top three dogs were frozen in the up position

— ACDII, HayTalk, "Banana bales, how to fix"

They were caked over with old grass and he had never noticed them. The outcome:

Bales are now nice and straight and solid. Upper hay dogs was the issue. They needed to be cut out with a cutting wheel as they were rusted in really bad.

— ACDII, HayTalk, "Banana bales, how to fix"

A broken spring is visible. A dog packed solid with chaff and rust looks completely normal from the ground and does nothing. On a Hesston 4550 another member found a top dog frozen in the down position that took a reciprocating saw and a chisel to remove. Move every dog by hand. Do not look at them.

The honest disagreement

Not everyone agrees hay dogs drive bale shape, and the dissent is worth reading because it tells you what else to check:

I agree on packer fork adjustment but I think non operating hay dogs will not affect bale shape when it applies to banana shaped bales but will adversely cause tying problems.

— Tx Jim, HayTalk, "Claas Markant banana shaped bales / wedges"

Another member answered from direct experience, and the detail he adds — position matters — reconciles both views:

I had two dog springs broken recently. Next to each other on the left, underside of the chamber. While one broken spring isn't going to have much of an affect (or more than one if not side-by-side), I will attest that this somewhat adversely affected my bale shape.

— 8350HiTech, HayTalk, "Claas Markant banana shaped bales / wedges"

One dead dog scattered among working ones is largely absorbed by its neighbours. Two or three dead adjacent dogs leave a real gap in retention along that stretch of chamber, and the bale bows away from it. That is why "frozen in a row" is worse than "one broken."

The half-slice mechanism

If you have ever pulled a bale apart and found what looks like half a slice sandwiched in the middle, this explains it — and it is the same root cause:

the haydogs are not doing their job allowing the hay in the bale to push back into the opening of the feeder. When the next charge of hay comes in, it contacts the exposed bale causing a portion of the hay already in the bale to wrap around making it look like there was only a half slice of hay added in that stroke.

— mike10, HayTalk, "Banana bales NH 570"

There is no missing hay. The bale crept forward into the feeder opening and the next charge folded material that was already there. If you see half slices and bent bales together, you have one fault, not two — and it is in the dogs.

Knotter faults are a separate family of problems that often get blamed on shape and vice versa. If your trouble is missed ties and broken twine rather than geometry, start with square baler knotter troubleshooting instead.

Cross-feed and packer forks: the backwards rule

If the bale is short on one side, you are in the delivery system. On a cross-feed machine such as a New Holland 68, the pickup gathers hay on one side and a set of tines walks it across to the chamber. Wear or misadjustment there starves the far side of the bale.

The adjustment rule is genuinely counter-intuitive and it is the reason people make these bales worse:

If your bale is missing material on the far side (away from the pickup) then you need to move your tines away, towards the pickup side. It sounds backwards but that allows a bigger bite to move more material over.

— WeatherBug, Bottom of the Mitten · AgTalk thread 1211491

The intuition says move the tines toward the light side to push hay there. The mechanism says the opposite: the tine has to start further back to take a bigger bite, and it is bite size, not reach, that determines how much material crosses. Coondle describes the same inversion on packer forks — to put more hay on the left, you adjust so the fork does not reach as far back.

Two more practical notes from that same AgTalk thread. First, on wear: one member found two feed tines on a 273 bent farther than the new ones he had bought; bending them back to match the new profile made the machine work much better. Worn and bent tines drift out of geometry gradually, so nothing ever looks obviously broken.

Second, the complete operating setup for a 273-class machine:

move the back wall behind the cross feed as far forward as it will go, pick up the windrow as far to the right side of the pickup as possible, and make sure you drive slow enough to get at least 14 strokes in a bale, 16 is even better

— Mitchco, SW OH · AgTalk thread 1211491

Three changes, only one of which is a wrench. Where you place the windrow on the pickup is a steering-wheel adjustment available on every pass, and it is free.

On packer-fork machines, one linkage detail gets missed constantly:

Did you move both ends on the packer fork link? Many do not realize the bottom of the link can also be moved to a different hole.

— mike10, HayTalk, "NH 570 banana shaped bales, need help"

If you moved the top of the link, tested, and concluded the adjustment does nothing, you have only used half the available range.

Feeder fork timing, with real numbers

Timing sits between geometry and feed rate, and it is where a machine that is adjusted correctly on paper still makes bent bales. DTN's "Ask the Mechanic" column laid out the check on a John Deere with numbers you can measure, and the framing is worth quoting because it captures why people misdiagnose this: what looks like a tying problem is a feeding problem.

The mechanism DTN describes is that if the feeder forks are slow relative to the plunger, hay placed on the upper left side of the bale falls away before the plunger can drive it into the top left corner — so that side of the bale finishes loose. The procedure:

1. Put the feeder fork pin in the lower hole of the feeder fork.
2. Turn the flywheel in its marked direction until the plunger is on the rearward stroke and the face of the plunger appears in the front feeder fork slot.
3. Roll a little further until the plunger face is in the middle of the front feeder fork slot, and stop there.
4. Measure from the tip of the feeder forks to the left end of the feeder fork slot. That distance should be 9.5 to 11.5 inches.
5. If it is not, loosen the feeder fork drive chain idler and move the feeder fork drive sprocket until it is.

The second half of the same procedure is the "attitude adjustment" — which hole the fork pin sits in, controlling how far the forks penetrate toward the far side. DTN's rule of thumb for that machine: the lighter the hay, the lower the hole, because light, small-stemmed grass needs maximum penetration to reach the far side. Every brand of square baler has an equivalent adjustment even though the name differs.

Timing is not one adjustment, either. As one member pointed out, there is more than one drive to check:

Check all hay dogs for free movement and that they have their springs. Next step likely checking plunger to packer fork timing.

— slowzuki, HayTalk, "NH 570 banana shaped bales, need help"

On machines with rotary feed fingers there is a separate timing spec for those as well, and one member in that thread was explicit that you have to check all of it rather than stopping at the packer-to-plunger relationship. Consult your service manual for the marks; they are model-specific and guessing is how people put a plunger into a knotter.

One safety note lifted straight from DTN, because it is easy to be casual about: after you loosen the chain idler or pull the master link, the feeder fork is in no way balanced and can rotate very quickly, catching a finger between sprocket and chain. Block it.

Also confirm you are actually at rated PTO speed. DTN's example runs 540 rpm producing 82 plunger strokes per minute, which at 16 to 18 strokes a bale works out to roughly four bales per minute. Running slow on the PTO changes every number above. While you are on driveline items, a slipping clutch or a shear bolt that keeps going is a related symptom — see baler slip clutch and shear bolt problems.

Why wedges are the last resort, not the first

Search this problem and somebody will tell you to add wedges to the chamber. Here is why that is the wrong first move, from the member who explained the mechanism most clearly:

Wedges rarely help in changing bale shape because all the wedges do is increase resistance for the hay to push out of the chamber.

— Coondle, HayTalk, "Claas Markant banana shaped bales / wedges"

And the diagnosis it distracts from:

Usually banana shape is caused by uneven packing of hay into the compression chamber.

— Coondle, HayTalk, "Claas Markant banana shaped bales / wedges"

Wedges and chamber tension raise resistance uniformly. A banana bale is a non-uniform problem. Adding uniform resistance to a non-uniform delivery makes both sides denser without changing the ratio between them — and it loads the plunger, the flywheel and the driveline harder for nothing. On the Hesston mentioned earlier, over-tightened tension springs were part of what the owner had to back off before he could see what the machine was really doing.

Twine tension belongs in the same category. Cranking it will not straighten a bent bale, and worn grooves in the twine tensioner pressure plates or the ceramic twine guides will give you inconsistent tension no matter where you set the knob — one member's fix there was to rotate or replace the worn plates. If you are chasing twine behaviour, choosing baler twine covers what knot strength and treatment actually change.

When it is not the baler at all

Two situations where the machine is innocent and adjusting it will only make things worse.

Crop condition

Coondle documented a decade of this on his own farm in Western Australia. In a season with low straw strength and material baled below about 13 percent moisture, he got bales in:

such a banana shape that they fall apart on trying to pick them up even by hand and impossible in a bale bundler.

— Coondle, HayTalk, "Banana bales NH 570"

His machine was verified correct — the dealer and the manufacturer's representatives could not find a fault, and a neighbour facing the same crop conditions returned his machine to the dealer. Brittle, over-dry, low-strength material does not interlock, so the bale has nothing holding its geometry once it leaves the chamber. The fix is baling window, not wrenches. A moisture tester is the cheapest way to rule this in or out before you start dismantling anything.

Rebaled or loose material

Rebaled bales always are lousy

— Teslan, HayTalk, "Hesston 4550 throwing banana bales"

Never diagnose bale shape while running loose, previously baled, or hand-fed material through the machine. The windrow is not uniform and the feed is not representative, so every observation you make is noise. If rebaling is a regular part of your operation rather than a diagnostic accident, rebaling round bales into squares covers doing it deliberately.

The diagnostic order, in one list

Work top to bottom. Do not skip to the bottom because it sounds more like a real repair.

1. Read the bale. Left-to-right bend or top-to-bottom bend? Write it down before you start adjusting things.
2. Count strokes between ties. Target 12 to 20, ideally 14 to 16. Fix this before anything mechanical.
3. Look at the windrow. Is it big enough and uniform enough to hit that stroke count at a sane speed? If not, fix it at the rake.
4. Check where the windrow enters the pickup. On cross-feed machines this is a free adjustment made with the steering wheel.
5. Weigh a bale. Very light bales do not create the back pressure needed to build the next one.
6. Move every hay dog by hand, top and bottom, along the full chamber. No play, no frozen dogs, springs and rubber stops present. Adjacent dead dogs matter more than scattered ones.
7. Inspect cross-feed tines and packer forks for wear and bending against a new part or the manual profile.
8. Check both ends of the packer fork link — the bottom moves too.
9. Verify feeder fork timing and the attitude adjustment to the manual's marks. On a John Deere per DTN, plunger face centred in the front slot, fork tip 9.5 to 11.5 inches from the left end of the slot.
10. Confirm PTO speed is at rating.
11. Only now consider wedges, chamber tension and twine tension.

What banana bales actually cost you

It is easy to treat this as cosmetic. It is not, and the costs compound at every step after the baler.

They fall apart when handled. ACDII's original description is the tell — bales picked up from the inside of the curve came apart even though they were baled tight. A bale that fails in the field is a bale you pick up twice.

They break in a thrower. One member with a bale thrower reported at least one bale exploding per wagon while he was filling it. Every one of those is loose hay in the wagon and lost tonnage.

They will not stack or bundle. Curved bales leave voids in a stack, so loads shift on the road and stacks shed water in the barn. Accumulators and bundlers are built around a rectangular prism; feed them bananas and they jam. This is the constraint that makes bale shape a system problem rather than a baler problem, and it is covered further in small square bale handling systems.

They cost you retail customers. Horse and small-animal buyers judge on appearance before they judge on stem or smell, and a bent, loose bale reads as careless. If you sell into that market — the market that makes tie count and bale dimension worth arguing about in the first place — shape is part of the product.

None of this is expensive to fix. A set of hay dog springs, an afternoon with the manual, and a change to how you rake are the entire repair on most machines. The expensive version is the one where you never diagnose it and rebuild the windrow, the wagon and the customer list around a bent bale. If you are shopping rather than repairing, choosing a small square baler covers what to inspect on a used machine — and hay dogs and feeder geometry belong on that list.

One closing note on scope: we manufacture bale net wrap for round balers. We are not baler technicians, and nothing here replaces your service manual or your dealer. What we can do is collect what experienced operators have worked out across years of these threads and put it in an order that saves you an afternoon.

Frequently asked questions

What causes banana bales in a square baler?

Uneven packing of hay into the chamber. Most often the cause is feed rate — thin or ragged windrows cannot push material to the far side. Mechanical causes are hay dogs that are broken or frozen, worn cross-feed tines or packer forks, and feeder fork timing that is out of spec.

How many strokes per bale should a square baler make?

Twelve to twenty strokes between knotter trips, with 14 to 16 being the range most operators settle on. Under 12 means the baler is overstuffed; over 20 means it is hungry and needs heavier windrows or more ground speed. DTN's guidance on one John Deere model targets 16 to 18.

Will wedges fix banana bales?

Rarely. Wedges increase resistance to hay leaving the chamber, which raises density uniformly. A banana bale is a non-uniform delivery problem, so uniform resistance does not change the ratio between the two sides — it just loads the driveline harder. Fix delivery first.

What are hay dogs on a square baler?

Spring-loaded pawls along the top and bottom of the bale chamber that let the bale move rearward but stop it moving forward, so the plunger has something to compress against. Weak, broken or chaff-frozen dogs let the bale creep forward, producing curved bales and half slices.

Why does my baler make good straw bales but bad hay bales?

Almost always windrow size. Straw windrows are typically larger and more uniform than hay windrows, so the baler gets a full, even charge every stroke. Building hay windrows to the same size as your straw windrows fixes it on many machines — and bales faster too.

The bottom line

Banana bale is one name for two different faults. Before anything else, look at the bale end-on and decide whether it is bending left-to-right or top-to-bottom. Sideways sends you to cross-feed, packer forks and feeder fork timing. Top-to-bottom sends you to hay dogs. Getting that right is the difference between an afternoon and a season.

Then count strokes between ties, because feed rate causes more of these than every mechanical fault combined, and because the count is free. Twelve to twenty, ideally fourteen to sixteen. If you cannot hold that at a reasonable ground speed, the fix is at the rake — build the windrow the baler wants.

Work down the list in order and leave wedges and twine tension at the bottom where they belong. They add resistance; they do not correct delivery. And if the crop is brittle and over-dry, or you are running rebaled material through the machine, stop diagnosing — the baler is not the variable you are looking at.

Related guides

• Square baler knotter troubleshooting
• Choosing a small square baler
• Baler twine buying guide
• 3-tie vs 2-tie small square bales
• Small square bale handling systems
• Baler slip clutch and shear bolt problems
• Rake selection guide
• Hay moisture tester guide
• Rebaling round bales into squares

Hero image: a small square baler working a hay field with the chute feeding stacked bales onto a following wagon — the handling stage where bale shape problems show up first. Photo via Wikimedia Commons, used under its stated license.