The following episode is from my early "troubleshooting" years, the times when I thought I knew it all...
I went to investigate a warranty claim of a
hydrostatic transmission that had recently been rebuilt. It consisted
of a Hydromatik A4VG56 closed circuit pump with automotive (DA)
control, which drove two (one for each axis) A6VM motors. The machine
was a small (relatively) aircraft-puller at a remote south-region
airport.
The pump and two motors had entered the workshop
with the usual "not enough force" complaint. No noticeable damage of
the rotary groups had been detected, so some lapping was applied, the
motors got new snap rings, the seals had been replaced, all the
components had been tested and sent to the client. However, when the
machine was put back in service, it presented the same symptom as
before. As I already knew the mechanics who recommissioned the machine,
I was sure it was not the case of "unhealthy" start-up procedure, so it
was obvious I had to see the machine in operation.
A few words about this type of machines. Although
relatively small in size, these plane pushing/pulling machines are very
heavy. Their whole structure is made of a very thick steel sheets to
give them extra weight. The main reason for them to be so bulky, is to
provide the necessary amount of tire traction to move large planes.
The symptom was apparent loss of torque in one direction
(forward) when the oil temperature rose. As the mechanics had lots of
machines with hydrostatic transmissions to look after, they already
knew that when a machine lost "force" when hot, it meant due
overhauling, that's why they'd sent the complete transmission (the pump
and both of the motors) to the workshop. The machine seemed to operate
fine with cold oil. The "loss of force" was tested on a small ramp the
machine had to climb to enter the airfield. Although the inclination of
the ramp was quite small, the overweight bustard could climb it only
when the oil was cold.
Nobody could tell me what the normal operating temperature
for this machine was, as, as usual, the machine had no temperature
gauge and no history of temperature readings. So we took the machine
for a small spin, and I saw that the loss of force was starting to
happen when the oil temperature hit 80 C. Equilibrium temperature
was around 85 C (with ambient temperature of 25 C). Although a
little high, it's a common temperature level for this type of
machinery, and the Portuguese hot climate is one of the reasons for it.
Apparently, the machine had done already around 10000 hours of service
with the same type of oil (VG 46, high Viscosity Index), and did not
present any problems. (If the machine were mine, I would install an
additional radiator to lower the maximum oil temperature to at least 70
C, though).
Anyhow, the loss of torque was there and needed
looking into. First thing you want to measure when troubleshooting a
hydrostatic transmission is, of course, the charge pressure. Frankly
speaking, I was sure I was going to see the pressure plunge, but
it did not. OK then, next thing - high pressure. It appeared, that the
loss of traction was being caused by insufficient high pressure, which
did not pass 350 bar with hot oil, and was not enough to make the
machine climb the stupid ramp. But what was causing the loss of high
pressure? Suspecting control issues, I installed a metric thread
measuring fitting to one servo cylinder, took the pressure reading,
then did the same thing to the other, the control pressure coming from
the DA (automotive) valve, with the motor accelerating at 2200 rpms,
was around 25 bars on both sides, which was more then enough to
maintain maximum pressure of 420 bars or more. The reason I was
measuring one side at a time was very simple - I only had one fitting
with the metric thread in my toolbox at the time. Apparently
insufficient servo pressure was not the problem....
Then I suspected relief valves malfunction, so I
dismounted both of the valves, inspected them, and after finding
nothing wrong, I interchanged them from one port to the other, to see
if the problem changed sides. Nothing changed...
One of the motors used in the transmission was of a
variable displacement type, so it also fell under suspicion, however
(luckily) it was very easy to interchange the pressure hoses of the pump, which was the next thing I did.
If the problem were in the pump, it
would change direction, if the problem were motor related, it would
remain in the forward direction. Changing the hoses inverted the problem,
so it was definitely the pump. At that time I opted for dismounting the
pump from the machine and bringing it back to the shop for
investigation.
Opening the pump revealed no signs of particle wear
or any kind of damage to the rotary group, which proved that all start
up procedures had been done by the book, so the pump was reassembled
and mounted on the test bench. All I needed to do was to invent a way
of simulating the problem on my test bench. I already knew for a fact,
that at normal test stand operating temperature of 50 C, the pump was
operating OK, so, to heat things up, I connected the charge pump
suction line to the drain line, turning the pump casing into an oil
tank, and shut off loop flushing. No need to say that after a minute of
intense "pressure making", the casing temperature hit 90 C. I knew that
the cut-off valve had been "cold" adjusted to 420 bars, pressure
limiters to 450 bars, but when the oil temperature hit 90 C, the
highest pressure I could get from one side was 350 bars, which meant I
was staring at the problem's ugly face! (Or it was staring
at mine...)
The best thing that can happen when you troubleshoot
a problematic component, is you succeeding in simulating the problem
you saw on the machine on your test bench. I was rubbing my hands, as I
was pretty sure I was about to nail the problem. This time I had all
the fittings of the world, and no yoga positions were required to
install a fitting and take a pressure reading. Staring at both servo
pressure gauges at the same time I turned the pump controlling signal
on...
There was no need to raise system pressure or conduct any
further tests, as I already new what the problem was only by looking at
both servo pressures at the same time and,
yes, I was cursing aloud at myself. As the servo pressure rose in one
servo cylinder, reaching the 25 bars, it also rose in the opposite
servo cylinder, reaching 10 bars, thus reducing the effective delta P
in the servo-system to 15 bars, which was just about enough to maintain
maximum system pressure of 350 bars. In these pumps the
swashplate mechanics is designed in a way that system pressure
forces the swashplate to zero angle. The higher the system pressure,
the higher the servo pressure must be to maintain the desired
displacement.
Further investigation showed that this phenomenon
was happening due to excessive wear of the sleeve of the pump's control
directional valve, combined with the existence of tank line orifice.
When the directional valve spool moved to the problematic side, the
internal leakage from the pressure gallery to the tank gallery,
aggravated by temperature rise, was enough to raise pressure inside the
tank gallery to 10 bars due to the existence of the small damping
orifice.
The problem was instantly solved by taking the orifice out
and the pump was immediately recommissioned, as the machine was needed
badly. The client was advised to replace the directional valve, but so
far the machine has been in everyday service for three years already (that I know for sure) and nobody replaced the valve yet...
The biggest lesson I drew from this troubleshooting episode was to ALWAYS check both servo pressures at the same time when troubleshooting a closed circuit pump, and always bring along enough extra fittings in my test accessories toolbox.
