Analyzing High-Voltage Surges And Compressors
A number of packaged air conditioning and heat pump units have operated off of our test cord power supply at our shop, and we have also temporarily connected a power supply to equipment on jobsites to perform equipment checks. The statistics vary from year to year, but at least 25% of the claims we receive have either:
These claims are easily satisfied by good service work. Diagnose the problem, repair if applicable, and document the findings and results. As mentioned above, a minor repair can be the simplest and least expensive way to settle some claims, but be sure that the policyholder has witnessed the operational unit and that you have some documentation of this. A signed work order would be excellent.
The following are some suggestions for a checking out the equipment.
EQUIPMENT CHECKCheck or test all starting components. A compressor will not reliably start without all of its starting components being operational.
Every service truck should have a means of testing a capacitor beyond just “seeing if it makes the needle swing” on an analog ohmmeter. Capacitor testers are not very expensive. Start thermistors are very easy to test with an ohmmeter, but usually a defective one has self-destructed and is in several pieces. Start relays are rarely used now, but they can and should be tested. They are notorious for failing intermittently.
If acceptable minimal voltage is not on the terminals while the compressor is trying to start, a perfectly good compressor may not start. Of course, a three-phase motor will not start without proper voltage on all three phases.
Low voltage could be due to a poor connection anywhere in the power supply system and the problem could be intermittent. This could also be due to undersized wiring or an overloaded system. Never apply power to compressor terminals while the terminal cover is removed and where damage or injury could result if one or more of the terminals blew out of the compressor. This can be especially dangerous if the electrical arc ignites the refrigerant oil that is blowing out. While this rarely happens, there can be too much damage or injury to ever ignore the possibility. This is not a warning of only a theoretical possibility. It does happen.
There are mechanical causes for a good compressor not to start. A high-pressure switch resetting prematurely can result in too much head pressure at start- up. A compressor without a hard- start kit (i.e., a properly sized start capacitor and start relay) will not reliably start on a non-bleed thermostatic expansion valve system.
A compressor that draws locked rotor amperage (LRA on the nametag) is usually electrically intact. If there were an open winding, shorted winding, or a winding shorted to ground, the amperage would be something other than LRA. This is one of the reasons this information is on the nametag. This is an indication of a compressor that has failed mechanically, but not electrically — yet. It is at this point that a hermetic analyzer can be quite helpful in determining whether the problem is within the compressor or the system.
Most compressors have an internal thermostat (IT) and/or overload mounted on or planted in the motor winding to prevent it from overheating. It opens a circuit when it senses an overheat condition or too much amperage and then automatically resets and closes again when it has cooled. (See Figure 1.)
This internal thermostat is a safety device and has a limited number of on/off cycles built into it. If it fails with the contact open, the compressor will no longer run due to this circuit being open. If it fails with the contacts closed, the motor will continue to overheat until the insulation on the motor winding breaks down and the motor “burns out.” Either way, the compressor has to be replaced.
LRA will cause the IT to cycle rather rapidly and frequently. This will lead to one or the other of the above. It should now be obvious that a mechanical failure that causes the compressor to operate out of its design envelope long enough will inevitably become an electrical failure.
PHYSICALLY DISMANTLING FOR FAILURE ANALYSISIt is now assumed that the compressor has been brought to your facility for dismantling (also known as a “teardown”) in order to perform a failure analysis. This should be done in an orderly, prescribed, and repeatable manner.
The mechanical failure modes and their symptoms will not be addressed here since they are covered in training programs supplied by a number of manufacturers. But it is important to say that an understanding of the mechanical failure mode(s) is essential to truly determine whether HVS is the root cause of the failure. Only after the compressor is completely dismantled so that all components can be inspected can a decision be safely made as to the cause of failure.
The first step is to check the resistance of the motor windings at the compressor terminals. Be advised that, in the real world, the mathematical sum of the start and run windings is not always exactly the same as the reading between the run and start windings. This seems to be true regardless of the brand or supposed quality of the multimeter used.
If the resistances are “in the ballpark” of the expected resistances, attempt to run the compressor. A prefabricated wiring harness complete with run capacitor, start thermistor, push on terminals, and labeled leads will save time and errors if this is done several times. The same cautions mentioned above concerning not running the compressor with the terminal cover removed still apply.
The supply and discharge connections should be opened and the discharge arranged to catch any oil that may spray out upon start-up and mist out thereafter. A bundle of rags in a large can or bucket will usually suffice. The amperages, voltages, and times should be recorded. If the run is successful, it should be repeated several times and all of the readings recorded. If started and run successfully several times, the compressor is thereby proved not to be damaged by HVS. If the compressor were to be stored for an extended length of time, it would be wise to run it several minutes with a purge of nitrogen only and then sealed. This compressor should remain without internal oxidation for several years.
If the test run was unsuccessful, the next step is to shake the compressor violently to get any particles lying in the sump in suspension in the oil and then pour the oil into a clean container, which can be used for storage if necessary.
Be prepared to label and store this sample as part of the compressor. Some states regard this as a hazardous waste and it is to be handled and/or disposed of properly. If it is not ruled to be a hazardous waste, and it is later determined not to be needed as evidence, mineral-base refrigeration oil can be recycled with engine crankcase waste oil.
Cutting the steel shell of a hermetic compressor should never be done with an oxy-acetylene torch since it will contaminate the interior of the shell. Several types of abrasive wheels can accomplish this rather rapidly if used with a little planning and judgment. Our tool of choice is a 9-in. side-angle grinder, but a smaller side-angle grinder or even a 3-in. air-powered cutting wheel will work. Goggles are essential, and a leather apron is advised. The sparks from the grinder can set clothing or anything within range on fire and will also melt or pit any glass.
The first item to inspect upon opening the hermetic shell is the connector that “plugs” onto the inside of the electrical terminals. If this looks suspicious, the next step may be to check resistance at this point and then perhaps attempt to run the compressor. The compressor may be run without the shell momentarily without overheating, but oil will have to be added to the sump if it is run more than a few seconds. The winding temperature may be monitored with an infrared non-contact thermometer.
If the run is impossible or unsuccessful, the stator will have to be removed from the compressor body for proper inspection. If the compressors have the stator pressed into the cast iron compressor body or roll-crimped into an aluminum body, it will be necessary to remove it from the shell for further analysis.
In all units, the copper-plated steel discharge line connects the compressor body to the shell. To remove the compressor body from the shell may require simply lifting the compressor off the mounting springs, or it may involve having to unbolt, cut, bend, or pry the mounting clips to release the springs or the compressor from the springs. These units are usually considered disposable. In the design there is no thought about the relative difficulty of a teardown.
After removing the compressor from the shell, it may be necessary to cut the body away from the stator. Now the difference between dismantling and tear- down becomes obvious. If the compressor requires a teardown this severe, abandon any thoughts of reassembling it.
Inspection of the stator is useless unless the construction of a motor is understood. On residential-size stators, the run winding is one continuous wire and the start winding is another continuous wire. Larger motors may have a splice, but the circuitry will be same. Small, three-phase, single-voltage motors will have three windings, whereas larger dual voltage motors will have six windings.
Since all parts of a series circuit will have the same current flow, there will normally be relative uniform heating of each winding. Drastic overheating to the point of insulation failure in a non-uniform pattern is abnormal and is indicative of HVS. A coating of thick, sticky soot is a result of an ongoing chemical reaction similar to a fire, except it is continuing without the presence of oxygen.
This reaction in a fluorochlorocarbon atmosphere is only possible after having reached an ignition temperature of 2,800 degrees F. This is possible only with a massive power flow, a.k.a. HVS. (See Figure 2.) Balls of molten copper in the sump or even splatter welded to the internal components are another sign of HVS. Gaps in the motor windings in excess of 3/8 in. or gaps in several places on the same motor are yet another sign. Always look for the burned ends of motor leads or gaps in the windings to be slightly rounded or even balled.
These symptoms have one thing in common. Each exhibits a power flow that is far beyond what would be possible with the normal system. (Circuit breakers, fuses, or overloads should normally prevent this.)
There are some symptoms that initially appear to be caused by HVS but could be from other causes. Metal particles that enter the compressor in the suction gas stream can imbed in the windings and can wear through the winding insulation and/or provide a conducting path between phase-to-phase or from phase to chassis-ground. The prime example is an insulation failure within the slot of the stator. To determine which caused this failure, the source of the metal particles must be either found or eliminated. This is when the contents of the sump and the oil sample can be very important. It there is no such contamination found in either, then this is no longer a possible cause.
HVS damage does not always have to be massive and impressive, but can be rather small, subtle, and very difficult to detect. Remember, the emphasis is to give the claim to the policyholder unless it can be proved otherwise. Be prepared to go over the internal electrical system of the stator in very minute detail. The motor lead assembly can chafe against the compressor body and short out, thereby resembling a HVS insulation failure unless the chafe marks can be found and identified. (See Figure 3.)
The internal thermostat overload may contain the deciding clue, and it may be necessary to carefully grind it open to examine it. The motor windings are in series with the heating element of the internal thermostat. If this element is melted, it is obvious that it would take a very unusual amount of power. Thus, it can be seen that the thermostat can fail open or closed, caused by HVS or by repeated cycling, and this can only be answered by opening it. (See Figure 4.)
Estes is president of W. E. Estes & Son, Inc., Athens, AL. He can be reached via e-mail at firstname.lastname@example.org.
Publication date: 06/03/2002