German tests on tractor pto power shows interesting results

“My neighbour has the same tractor as me, but he can plough similar land one gear higher.” Sounds familiar?

Experience and common sense say that seemingly-identical tractors won’t necessarily perform the same, though evidence on how much they can differ is thin on the ground.

Which is why Top Agrar magazine took an Eggers pto dynamometer (and its factory operator) out to farms and contractors in Schleswig-Holstein, then compared the pto power of tractors in service with brochure figures and test station results.

The findings are summed up in the two graphs below, which bear some explanation – not least because field power measurement is not an exact science.



Mobile dyno maximum powers for Fendt 716 (left) and Deere 6920. Horizontal bars are individual tractors, shaded areas show ±5% allowable tolerance for pto output, estimated from maker’s claimed engine power. Red vertical line is maximum power recorded in DLG tests.

Fendt TAB


Service hours of tractors tested. Tractor number here corresponds to tractor number on left side of the graph.  There is no obvious connection between max power and service hours.

Tractors and expectations

Fendt and John Deere tractors were sampled for maximum pto power. Models were the same (Fendt Vario 716 and Deere 6920), although types and recorded hours varied.

Of 11 Fendts checked, eight were Vario 716s and three were the 716 TMS version. Of the Deeres, three were two-valve 6920s, three were four-valve models, and three were power-boost 6920S versions.

The latter were measured in non-boosted mode. Graphs show individual tractor results the table has service hours recorded.

How much variation is reasonable?

Differences between components, and the consequent stack-up of tolerances when those bits are assembled into an engine, mean no two motors can be the same. To account for this, makers allow plus or minus 5% production tolerance in power output.

And don’t forget the difference between quoted power and what you might find at the pto. For instance, Fendt says a Vario 716 TMS produces 129kW (173hp) maximum according to ECE-R24.

This is at the engine crankshaft. Allowing for a typical 10%-15% loss between crank and pto stub brings that back to around 116kW (155hp), which drops by another 5% when power drain by ancillaries is included.

Apply the maker’s ±5% tolerance, and you might expect to find maximum pto power in the 105kW-117kW (141hp-157hp) range. In the graphs, the expected power ranges are shown as light shading.


Power can vary between seemingly-identical tractors, and from published values, for a host of reasons. Here are some of them.

Measurement standard. Quoted power varies according to the standard used to measure it, for example whether (and which) engine ancillaries can or must be connected during the test, and how (or if) variables like air temperature and pressure are corrected.

Mobile dynamometer tests are not carried out to equivalent formal standards. For example, during rating to ECE-R24 the engine fan runs at minimum speed, which it may or may not do during an in-service check. So while results can be compared reasonably between tractors checked with the same mobile dyno, drawing parallels with standardised values is risky.

Measurement location. Brochures generally quote engine power at the crankshaft. 10%-15% is lost between crank and pto stub, though the figure can be higher depending on driveline complexity.

Tolerances. It’s not yet possible to make parts completely identical. So fuel injection pumps, injectors, pistons, rings and so on are subtly different, leading to variations in efficiency. This can show up as power differences between apparently-similar tractors.

Maintenance. Tractors in the Top Agrar survey came straight from work and were at different points in their maintenance schedules. Apart from obvious output-sappers like a dirty air filter or out-of-adjustment tappets, too much oil in the transmission can absorb power. And if a viscous-coupled or electronically-controlled fan misbehaves and runs too fast, it sucks away power.

Wear. In general, engine output tends to climb as moving parts bed in, then very slowly drop away as components wear. For example, changes in the injection pump and its drive can affect injection pressure and timing, and thus power. But wear rates depend on use and maintenance, and obviously vary between tractors: in this survey, power and age seem unconnected. Electronics help remove some variation.


Regardless of brand, individual maximum pto power outputs varied by more than 20% – up to 28kW (37.5hp) between tractors of the same type – which could explain why farmer A’s tractor works faster, or pulls an extra furrow, than farmer B’s.

But none of the tractors delivered less power than the makers promised. And in several cases the owner had quite a lot more to play with: Compare the shaded zones (expected power) with measured values in the graphs.

Farmer-tractor outputs were also stacked alongside the model’s maximum power recorded at the DLG station, either as an OECD test or for magazines. Here the story was a little different.

Five of the 11 Fendts came close to (or beat) the Vario 716’s formal figure of 121kW (162hp). By comparison, seven of the nine 6920 Deeres managed the same trick. But please see Watching the watchers and Why the differences? for cautions.


How accurate was the Eggers mobile dyno used in Top Agrar’s tests? Back in 2004 the DLG test station checked a PT 310 MEM against its own calibrated equipment, finding up to 5% error when measuring power.

In the Top Agrar survey, the actual dyno used was not independently calibration-checked. And while ambient temperature and air pressure were corrected automatically during power checks, fuel temperature and quality was not standardised. Naturally the test procedure itself could not be tightly controlled, as it is in a formal test. All of which means that, in this report, differences of 5% or less must be ignored when comparing power values taken from different sources.