The modern dry vacuum pump -- when it works -- is economical ($200 to $400), low-weight (under three pounds) and capable of high-output operation. The dry pump displaced its traditional competition (venturis and oil-lubricated "wet" pumps) in a matter of only a few years during the late 1960s and early 1970s.
The Achilles heel of the dry pump is -- as the FAA said in a 1987 letter to the NTSB -- that "pumps fail catastrophically, without warning, and there is no degradation of performance obvious to the pilot to warn him of imminent failure." Pump failures are not new stuff; what's changed is the magnitude of the media sensationalism surrounding a crash, especially with a senator on board.
Both Airborne (now out of the vacuum pump business, although theirs are the majority of the installed base of dry pumps) and Sigma Tec pumps utilize carbon-graphite rotor construction, with carbon vanes riding loose in the rotor slots. In normal operation, the vanes are thrown against the pump housing by centrifugal force, rising and falling on the elliptical walls and compressing the air trapped in vaned compartments.
Airborne units came in "CW" (clockwise) and "CC" (counterclockwise) models, and for longer life, the proper direction must be observed on installation. In either case -- Sigma or Airborne -- the vanes run dry on aluminum-housing walls. The constant, gradual, wearing away of the graphite is the only lubrication the pump gets. Hence the term "dry pump."
Dry pumps can be used to suck air (vacuum) or blow air (positive pressure), depending on which side you hook the plumbing to. When dry pumps are used to provide pressure (as in de-ice boot systems), in-line filters must be employed to remove carbon dust from the system.
Both Sigma Tec and Airborne pumps have a standard, spline-drive mounting for use on Lycoming or Continental accessory cases. Also, both makers incorporate frangible drive couplings designed to shear in the event of rotor lockup, thus sparing the engine accessory gears of possible damage.
The pump makers differ in their approach to drive-coupling design. Sigma's coupling transmits torque straight to the rotor along a thin quill-shaft (which has since been changed to a speedometer-type cable in the so-called "dash three" pump models). Airborne transmits drive torque to the rotor via a coupling sandwiching shear pins between a nylon torque plate and an upper torque plate, with the rotor spinning on three finger-spools, which "grab into" the rotor.
The Sigma Tec pump drive is "more frangible" than Airborne's nylon torque-plate drive. The Sigma is designed to fail at 100 inch-pounds of torque, whereas it takes 250 inch-pounds to snap an Airborne drive.
That said, even experts can have a tough time pinpointing the primary causes of a given pump failure.
What causes pumps to fail? The use of carbon graphite as a structural as well as a lubricating material certainly seems intrinsic to the problem. Ironically, it's carbon graphite's unique qualities that make current "self-lubed" pump designs possible in the first place.
In our research (which included talking to pump manufacturers as well as mechanics, owners and overhaulers), we identified 10 things that could cause a dry pump to self-destruct:
Additionally, entry into the pump of degreasing solvent (the type sprayed on the engine for routine inspections) can cause failures unless exceptional care is taken. Varsol can enter a pump through its exhaust tube or drive seals. This is something to watch for during any engine spray-down.
Generally, this type of failure occurs shortly after installation of a new pump. Pump manufacturers lay the blame for a large percentage of warranty claims to this cause. (A filter change is usually required for warranty coverage to be in effect after installing a new pump.)
There is concern that particles small enough to pass through filters can mix with the lubricating powder at the rubbing surfaces of the vane and rotor, increasing the wear rate and leading to early pump failure. Even cigarette smoke is bad for the pump -- yours and your engine.
Cooling is often poor, in part because of low humidity at high altitude and because aircraft designers don't always expose the pump to ram air.
The maximum continuous operating speed of Airborne pumps is 4,000 rpm (rotor shaft); for Sigma it's 4,200 rpm. Lycoming pump pads generally turn 1.3 times crankshaft speed. Continental pump pads turn 1.5 to 1.545 times crank speed. This equates to a Continental engine exceeding 2,588 rpm for Airborne or 2,700 rpm for Sigma.
Combine high rpm with high demand (as in a Continental-powered Cessna P210 with de-ice boots flying at 20,000 feet), and you can begin to see why some operators experience so many problems with pumps. Add in a prop over-speed incident to the scenario and you are pushing the pump where it was not designed to operate -- for long.
Attention to instructions can eliminate incorrect installation of pumps. But avoiding occasional engine "kickback" on start-up (or shutdown) is not such an easy matter. If vane/ slot clearances have opened up, one kickback may be all that's needed to jam a rotor and trash a pump.
Of course, pumps respond poorly to having their housings squeezed in a vise, which many installers do while installing fittings. The makers have express warnings not to do this, but installers do it anyway.
Float plane operators (who suffer a relatively high incidence of shock-related avionics and panel problems) have reported replacing vacuum pumps every 50 to 200 hours, on average -- further evidence that shock and vibration can have a destructive effect on pumps.
The FAA has received SDRs describing sticking de-ice boot valves in some aircraft. Ordinarily, pneumatic de-ice boots cycle on and off, alternately inflating and deflating, at the behest of a small timer and solenoid-actuated de-ice boot flow valve.
If either the timer or the valve hangs up in the "inflate" position, however, the vacuum pump can quickly lug and overheat. Until recently, the loss of a vacuum pump in this manner meant not only the loss of boot action, but gyro instruments as well.
According to overhaulers' figures, under the best of circumstances, smaller (211-type) dry pumps are unlikely to operate reliably much over 600 hours, as the vanes will have worn to the point where they are likely to cock and jam.
The pressure regulator valve (note the safety-wired head) is easily visible in late-model Bonanzas. The area at the base of the valve should be checked visually for excess carbon at each preflight. Rapid carbon buildup is indicative of pump distress.
The so-called "boot pumps" (high-capacity Airbornes) are unlikely, in most applications, to last more than 300 to 400 hours. Of course, there are always exceptions.
In short, then, the modern dry pump, by virtue of its design and construction, is acutely sensitive to almost everything in its normal environment: heat, oil, solvents, dirt, water, vibration, mechanical stress and (some would say) the moon and tides.
Even under the best of circumstances -- with a new (gyros only) pump installed by experts, with cleaned lines and new filters -- you still cannot expect much more than five years of normal flying before your pump is a real candidate for failure. The only thing certain is that it will fail. You just can't say when.
What if your pump is on the verge of giving out? Is there a way to tell? Do you have to wait until the DG dies to learn that your vacuum pump has pumped its last breath?
On every preflight (if your cowl is openable), you should get into the habit of visually inspecting your vacuum pump and the pressure relief valve or exhaust tube. Look for oil at the base of the flange (indicating a bad gasket) or tiny bits of stripped nylon in the coupling area (which is open for view -- although you must look closely). Bits of nylon are an indication that the coupling is nearing failure.
In a pressure-type system, give the relief valve (in the firewall area) a visual once-over periodically. Look for carbon buildup (i.e., black soot) indicating possible rotor/vane distress. By checking at set intervals, you can get an idea of the normal soot buildup produced by your pump. Abnormal buildups will then be easy to detect.
In the cockpit, learn to include the suction gauge in your normal visual scan. In a dry-pump system, any rapid fluctuation of the gauge (no matter how intermittent) is a definite warning signal that something is amiss. See that your system is properly adjusted to give the correct readings at 1,500 rpm or above. (Most regulators don't begin to regulate until 1,500 rpm.)
Dual vacuum-pump installations have received a lot of attention, what with Cessna's experiences with the 210 series. Unfortunately, not every small-plane engine has an extra pad available for a standby pump. While the engine manufacturers have developed T-drive adapters for some applications, the drives are not widely "retrofittable." Still, if you want a standby vacuum pump in your system (and you don't want to trade up to a twin), there are ways to do it.
Replacing a vacuum pump need not be a major hassle, if you follow a rational sequence of procedures.
Do not use Teflon tape, pipe dope or unapproved thread lubes. Tighten fittings one-and-a-half turns maximum, using a box wrench. Align fittings for plumbing connections in the aircraft.
In a pressure system that has experienced pump failure, be sure to blow out all lines with compressed air from the panel side, to get remaining bits of carbon. Make sure hoses are connected to the proper fittings (do not swap inlet/outlet hoses by mistake). Use only specified fittings, not pipe fittings.
Airborne specifies a life limit of six years for nylon drive couplings. Factory kits are available for replacing couplings: Kit No. 350 for 211/212 series pumps, No. 352 for 440-series pumps.
In short, don't sit and wait for a failure, and be proactive.
More aircraft repair and prevention articles are available in AVweb's Maintenance Index. And for monthly articles about aircraft maintenance, subscribe to AVweb's sister publication, Light Plane Maintenance.