Once upon a time, a client from far-far away sent us a Rexroth A2FM32/61W hydraulic motor for an overhaul. The motor came from a simple open-loop system, in which it was driving a large circular saw. Local mechanics had already disassembled the unit and verified that it was in quite a bad shape, so they joined all the pieces and reached for our "professional" help.
The complete rotary group had to be replaced - the cylinder block, the pistons, the central pin, the retainer plate, and the valve plate, topped up with a seal kit. The motor was assembled, tested, and dispatched to the client.
Not a day passed when the man called saying the motor was "bad". According to him, it would run fine for about a minute, sometimes less, and then stop. Restarting the machine seemed to "fix" the issue, but only for a brief period, which was fun to look at but, of course, impossible to work with.
A bad start-up? Contamination issues? Component failure?.. We had to see it again, so we asked the client to send the motor over. The man was needing the saw so badly, that he opted for a more than 350 km long drive and brought the motor himself!
When the unit was disassembled, the rotary group appeared to be in perfect condition. Apart from some minor scoring of the valve plate and the cylinder block, all the parts looked as if they'd just come from the shelf! Even the client got puzzled because he was expecting to see a major failure.
He was claiming that the motor had stopped rotating completely, and was also positive that the rest of the hydraulic circuit was fine. As no damage was detected, the motor was reassembled, retested, and, once again, showed perfect efficiency. The man was advised to look for a problem elsewhere. Very puzzled, he spent several hours of the 350 km stretch back home thinking about what he could do to troubleshoot the malfunction of his hydraulic system.
No, this story is not over yet. The client called the next day... to confirm that the motor was stopping the same way as before. All he could get from the saw was a couple of minutes of service at best, and then it would stop. The man had the happy idea to check the motor drain line, and when the motor stopped the line presented a huge flow increase, so it had to be the damned motor!
Although the poor fellow was not happy about the hours he'd spent behind the wheel, he had no choice but to bring the motor over once again. The motor went through the same procedure as before - it was disassembled, no wear was detected, it was assembled, tested - and worked OK. At this point, yours truly was asked to "take a look". It would be really ugly for the man to have to go through the same trouble again.
There's nothing I like more than solving puzzles, especially when they have something to do with hydraulics. This case was very interesting. Not only we had to "nail" the problem, but we also needed to find a way to convince the client that the problem was solved "for good" this time. This could only mean one thing - creative testing!
A few words about the previous tests. It was a fixed displacement motor with no loop flushing valve, so only a very basic test was performed after the initial overhaul - the motor was connected to a pressure source, and then a needle valve was used to restrict the return flow to raise the pressure inside the rotary group, with the case drain flow monitored to conclude if the rotary group had excessive leakage.
Such a test is far from perfect, but it is a "fast and dirty" way to check units for excessive leakage. The second time no chances were taken though, and the motor was tested on a "proper" bench that used another hydraulic motor to create the braking torque. That time the case drain flow, work port flow, and the RPM were registered, different speeds and pressures were applied and yet again the motor showed excellent performance - like any decent motor that'd just gotten a new rotary group would. Obviously, the workshop tests were not recreating the real-life working conditions.
I knew that the motor was spinning a heavy wood saw which had tons of inertia. There even was a system of anti-cavitation check valves to compensate for the blade overrun when the DCV would decrease the oil supply. Our braking test bench, however, had zero inertia, so to simulate that condition, I coupled the motor to a 30 KW electric motor, to use the electric motor's massive rotor to simulate the heavy blade.
During the previous tests, the case drain had always been so low, that it had to be evaluated in a very non-scientific (and not safe) "by the eye" manner. This time was no different - and the case drain hose was hanging freely from the side of the motor. I started the pump, and slowly moved the DCV handle to set the hydraulic motor/electric motor assembly in motion. As I was increasing and decreasing the speed of the motor, there was a moment when it suddenly lost all the torque, and the drain line match-thick trickle suddenly turned into a loose fire hose, immediately scattering the curious standers-by with an oil spray. There it was, the malfunction, staring in our oil-covered faces.
I like to say that a successful simulation of a problem in a workshop equals solving the problem. There it was - all the oil passing through the motor into the drain. And then when I stopped it and restarted slowly - it ran fine - just like the man was describing! And then when I decelerated it abruptly, and then accelerated - the case drain boosted up once again... The first and most obvious explanation would be the cylinder block lift, but how? With new original parts? Well.. let's tear it down once again, boys!
All the parts were checked, the valve plate was a perfect match to the cylinder block, the disc springs (or Belleville washers) were there... wait a minute... what's this?!! It became clear at once where the problem was!
The mechanic who'd mounted the motor had mounted only four Bellville washers (the nasty A6VM habit), while this motor required six! As you can see here:
Four washers, even when stacked in an alternating direction, are lower than the guiding washer. The caliper is actually touching the guiding washer and Bellevilles are loose! Even with new parts, the cylinder block was not sufficiently pre-loaded against the valve plate. It was merely touching it, allowing it to pass all the tests, as the steady system pressure would hold the hydrostatically compensated cylinder block compressed against the valve plate.
When the massive saw blade was over-running the oil supply, the motor was working as a pump, creating low pressure in the work port, and since the cylinder block was not sufficiently pre-loaded, the positive case pressure coupled to the low port pressure was creating a force high enough to cause the cylinder block to lift. Once lifted, it would stay "floating" on the escaping oil and wouldn't settle until the oil supply would stop.
The test was a classic cylinder block lift simulation. It was so classic that I even made a video of it (for future didactic purposes), which was then successfully erased from the camera by a colleague of mine (thanks a lot!). The problem was solved by adding the two missing washers. Yes - two measly washers for all that trouble!
By the way, there are different ways you can stack the washers. Rexroth advises stacking them in the following manner:
I know from experience that these motors run fine when all the six washers are stacked in an alternating direction, so if you ever mounted one like this, don't beat yourself about it. Stacking the washers in parallel does create a stiffer spring, though. Belleville washers are the old design, all newer series use coil springs, which are better.
What does this prove? It proves that the pre-loading of a cylinder block is not a mandatory condition for high volumetric efficiency. A well-designed cylinder block is held compressed against the valve plate by the system pressure, so the whole pre-load spring system is only necessary to guarantee that it stays compressed when there is no hydraulic force, which happens at start-up, or, as we've just seen, during the overrun, when a lower than the case pressure develops in work ports.
Of course, there are also vibrations, pressure surges, extreme velocities, temporary low oil viscosity, and many other things contributing to block lift and making the pre-load system necessary. But at a steady speed and pressure, you could completely remove the pre-load springs, and the cylinder block would remain adequately compressed against the valve plate by means of the system pressure alone!
Understanding this phenomenon as well as understanding what a cylinder block lift is and how it can manifest itself is very important for correct troubleshooting. It is also critical to be aware of the fact that standard workshop hydraulic tests are nothing but a good approximation to real-life conditions, therefore some malfunctions are very difficult to recreate in a workshop. In these situations, a good technician should be able to think outside the box and be ready to perform "creative testing" to nail the problem.