TL;DR — A refuse packer or ejector panel is usually pushed by two cylinders fed from a single flow divider, and no divider splits flow dead evenly. Two percent of a 1,600 mm stroke is 32 mm, which is just under one degree of skew across a panel with pins 2,000 mm apart. The durable fix is not a more accurate divider. It is a circuit that lets both cylinders reach full stroke on every retract, plus a mount that can carry the moment a one-sided load puts on the panel.
The error is geometric, not hydraulic
I have argued about flow divider accuracy more often than about any other part on a refuse body. Suppliers quote two, three, five percent. Operators report a panel that comes out crooked. Both descriptions fit the same mechanism, and pressure has nothing to do with it.
Turner Hydraulics documented the same mechanism on a dump truck: a divider running at 98 percent efficiency extended one cylinder fully and left the other at 98 percent of its stroke. Because the strokes were never equalised, the trailing cylinder fell further behind on each following actuation. Their remedy was a check valve in each cylinder return line, so both units could bottom out and reset the system every cycle.
The arithmetic takes thirty seconds and it explains most crooked panels. Take a 1,600 mm ejection stroke. At 98 percent, the lagging rod stops 32 mm short. Put the two pins 2,000 mm apart and 32 mm becomes 0.9 degrees of skew — invisible on a photo, obvious to a guide roller running 1 mm of clearance.
| Divider efficiency | Shortfall on 1,600 mm stroke | Skew across 2,000 mm pin centres | What it does to the panel |
|---|---|---|---|
| 99% | 16 mm | 0.46° | Absorbed by guide clearance |
| 98% | 32 mm | 0.92° | Panel starts loading one roller |
| 95% | 80 mm | 2.3° | Rod takes bending, gland leaks on one side |
| 90% | 160 mm | 4.6° | Panel binds, cylinder stops before end of stroke |
Those figures are calculated from geometry, not read off a gauge. They are the reason a body that ejects "a bit crooked" later arrives with one rod scored on a single flank.
Why the gap grows instead of self-correcting
A divider is a volumetric device. It splits what it is given, and it has no idea where the panel actually sits. Nothing in the circuit compares the two rod positions, so the only event that can reset the error is both cylinders reaching a hard stop.
On a body that retracts the panel until a limit switch trips, that never happens. The panel stops wherever the switch is, the trailing rod stays 32 mm out, and the next stroke starts from an unequal position. Ten cycles in, the difference is not 32 mm any more. It is whatever the guides will tolerate before friction takes the load.
Full-stroke bottoming is the cheapest correction in hydraulics and the one most often designed out. A return check valve, a sequence that strokes both cylinders to the end, or a mechanical stop all put the reference back where it belongs.
Offset load is the other half
Flow balance only decides whether the panel stays square. Fill pattern decides how much force it takes to keep it there. A rear-loading body filled from one side, or a transfer trailer loaded with dense material on one flank, puts the pressure centroid off centre.
That offset creates a moment the panel has to resist. If the guide rollers and rails are tight, the structure takes it. If clearance has opened to two or three millimetres, the panel rotates until something stops it, and the something is usually one rod in bending. The signature is consistent: a bright polished stripe down one flank of one rod, a wear band flat on one side only, and a gland that weeps on the loaded side first.
Replacing the leaking cylinder in that condition buys a season. The mount and the guide clearance caused it.
What I check before condemning a cylinder
- Rod exposure at rest. Tape-measure from gland to rod end on both cylinders at the same panel position. More than 5 mm difference means the circuit is not equalising, and no seal kit will fix it.
- Does anything bottom out? Trace the retract line. If there is no return check valve and no full-stroke event, the design has no reset and the error accumulates by default.
- Pin centres at rest and at half stroke. If the distance between the two rod-end pins changes as the panel travels, the mounts or the panel are distorting and the cylinders are fighting the structure.
- Guide clearance. Anything above 1 to 1.5 mm lets the panel skew before the rails resist. Check roller wear and rail straightness, not just the cylinders.
- Wear pattern. One-sided polish and a flat wear band point to alignment. Even wear all round with a scored tube points to contamination.
- Fill practice. Ask the operator, not the workshop. A body that only ejects crooked after a particular route is telling you where the load sits.
Four ways to keep a panel square
| Method | How it resets | Corrects per cycle | Where it falls short |
|---|---|---|---|
| Flow divider alone | Nothing — splits by volume only | No | Error accumulates until friction stops it |
| Divider plus return check valves | Both cylinders bottom out on retract | Yes | Needs a full-stroke event in every cycle |
| Mechanical link (torque tube, rack) | Structure forces equal travel | Continuously | Adds mass, needs space, wears at the joints |
| Position feedback with proportional valve | Closed loop trims the fast side | Continuously | Sensors and tuning; overkill below 5° of tolerance |
| One central cylinder on a guided panel | No second cylinder to disagree | Not applicable | Guides must carry the whole offset moment |
When the pair has to be replaced together
If rod exposure differs at rest and the guides are tight, the pair is usually finished. A cylinder that has run skewed for a season has an oval tube and a worn guide bush, and matching it to a new partner puts the whole correction load on the new rod.
Order them as a pair — same bore, same stroke, same pin centres, same port orientation — and state the panel width between pins so the supplier can check the moment the guides will see. Our refuse truck cylinders and compaction station hydraulics are built to ISO 6020 and 6022 geometry and pressure-tested before dispatch. Where a part number is obsolete, custom and OEM replacement works from a drawing or six measurements.
If your ejector runs on one stage over a long body, the tipping side of the same argument is covered on hcictec.com, and the flow supply behind a skewed panel is a power unit question we cover at trailerhydraulic.com.
What to send with an enquiry
Bore, stroke, retracted length, pin centres, port orientation, working pressure, single or double acting, and the panel width between pins. Add whether the circuit already bottoms both cylinders on retract, because that answer changes the seal and bush recommendation more than the bore does.
One sample is the minimum order. First-article lead time is 45 to 60 days, and 50 pieces or more run 30 to 45 days, confirmed against the quotation. Quotations come back inside 48 hours, and every cylinder carries an 18-month warranty from dispatch with a full maintenance kit available.
Frequently asked questions
Why does my ejector panel come out crooked when both cylinders are new?
Because two new cylinders on one flow divider are still two independent volumes. A divider at 98 percent leaves the trailing rod 32 mm short on a 1,600 mm stroke, and without a full-stroke event that offset carries into the next cycle. Check rod exposure at rest before you look at the cylinders themselves.
Will a more accurate flow divider solve it?
It buys time, not a fix. A divider at 99.5 percent still leaves a few millimetres per stroke, and the error is cumulative as long as nothing resets it. Return check valves that let both cylinders bottom out cost far less than a high-accuracy divider and remove the drift entirely.
Can I replace just the leaking cylinder?
Only if rod exposure matches at rest and the guides are tight. Once a cylinder has run skewed, its tube is oval and its bush is worn, so a new partner takes the whole correction load on one rod. On a paired ejection circuit we quote them as a pair.
Does offset load damage the cylinder or the body?
Both, in that order. The panel rotates until the guides stop it, and the rod bends to follow. The cylinder fails first because the gland is the tightest clearance in the chain. Fixing the guide clearance and the fill pattern is what stops the replacement failing the same way.