Find more articles on: Variable Frequency Drives
Deciding where to put your drives requires more than just casual consideration.
If your VFDs are failing prematurely, it may be because they're located in your motor control centers (MCCs). MCCs optimize space by putting motor controls and branch-circuit protection in a common enclosure. They allow you to consolidate wiring, simplify installations, and keep the electrical controls out of the manufacturing environment. However, MCCs weren't designed to house electronics. Several potential problems can arise when VFDs and other heat-producing, solid-state devices are located in MCCs.
Why not MCCs? Voltage reflection, which can destroy a motor, may occur when the motor cable is longer than or equal to the critical length; lead lengths as short as 25 ft are sufficient for such an event (Figure). Thus, voltage reflection is a function of the lead length of the power cable between the drive and motor.
We can predict voltage reflection using the following equation:
Eq. 1: Lcritical=Vcable × tr
where, Lcritical is the critical cable length in feet, Vcable is the pulse speed from the drive to the motor in feet per microsecond (ft/µsec), and tr is the rise time of the output pulses from the VFD under consideration in microseconds (µsec). Vcable is often estimated as 500 ft/µsec. The Vcable, or “propagation factor,” is the speed at which the pulse travels from the drive to the motor in ft/µsec. An independent study of 17 different drive manufacturers suggests rise time of the reflected pulse may be as short as 0.117 µsec. Plugging these numbers (0.117 µsec and 500 ft/sec) into Equation 1 gives 29.25 ft as the critical distance for voltage reflection. In other words, motor cable lengths must not exceed 29 ft in most installations if VFDs are located in the MCCs. You can alleviate this situation with several solutions like using output dv/dt filters or a definite-purpose inverter-fed motor. If 230V power is available, you can also use a 230V VFD.
The first two solutions add cost and complexity. For example, dv/dt filters are large and hard to install in MCCs. How would wiring a 230/460V motor for 230V solve anything? If voltage reflection does occur, the voltage peak will be less than 1,000V, and the motor insulation guarantee level will be 1,000V. On 460V drives, 650V (DC bus voltage)×2=1,300V. On 230V drives, 325 (DC bus voltage)×2=650V — well below the insulation rating of 1,000V. You can avoid all of these drawbacks by putting VFDs close to their motors, not in MCCs.
Thermal considerations. Putting a VFD in an MCC requires a special enclosure. Any savings you might expect from reduced wiring costs will disappear in the face of special engineering charges and cooling requirements. Solid-state switching devices typically generate losses in the range of 4% to 6%, requiring heat dissipation within the MCC.Calculate the VFD heat loss using the following two equations:
Eq. 2: Heat loss=hp×746W/hp×[(1-VFD efficiency)÷(motor efficiency)]
Eq. 3: BTU load=heat loss×3.41
For example, a 15-hp drive with an efficiency of 96% and connected to a 90% efficient motor will yield these results:
Heat loss=15 hp×746W/hp×[(1-0.96)÷0.9]×3.41
Heat loss=11,190W×[(.04)÷0.9]
Heat loss=497.3W lost
BTU load=497.3W×3.41
BTU load=1,696 BTUs/hr
Since VFD losses are nearly proportional to horsepower, you can estimate losses at 33.3W or 113 BTUs/hr per horsepower. For example, the losses on a 30-hp drive would be about 1,000W. This is like having a 1,000W heater in the enclosure. So the challenge is to get the heat out of the MCC bucket without compromising the VFD heat sink design or restricting VFD cooling air.
What about front ventilation to prevent a heat-stacking effect from multiple VFDs in a common vertical enclosure? This is easy to do in a NEMA 1 enclosure, but not so easy in a NEMA 12 enclosure. If the heat sinks are mounted on the front door of the enclosure, load conductors must route to the door. Prudent control panel design limits the number of conductors traveling across a hinge; an unwritten rule instructs you to never run power conductors across a hinge. You can flip the drive over to shove the heat sink fins through the door's gasketed opening, but this solution puts the drive in an awkward position for service technicians.
Many specifications call for input line reactors and/or output filters, which are heat-producing devices that require large mounting areas. Most VFD manufacturers have pre-engineered designs for placing filters inside the drive's enclosure. Twelve-pulse drives, often a means for eliminating harmonics, don't fit well into MCC enclosures because their large phase-shifting transformers produce too much heat. It's much easier to retrofit harmonic mitigation devices to freestanding or wall-mounted VFDs than it is to mount and connect these devices to drives inside MCCs.
Project coordination. Putting VFDs in MCCs can complicate project coordination. For example, if a blower motor changes from 15 hp to 20 hp, it's easy to change from a 15-hp starter to a 15-hp unit — both starters typically fit into the same size MCC bucket. However, a 20-hp drive will most likely be larger than a 15-hp drive. This may require a bigger MCC bucket and will require more cooling air through the MCC.If the MCCs aren't yet released, you'll have a time delay for re-engineering the MCC line-up, but if not, you'll have expensive field changes.
Other concerns. Motor control centers have “wireways” through which to route all control and power wires. However, such wiring should run in separate wireways — often a vendor requirement. Installing a VFD in an MCC may void the performance part of the VFD warranty, because such environments can give off unnecessarily high levels of RFI and EMI.The start/stop and speed controls typically come from equipment not mounted in MCCs. Mounting the VFDs in the MCC complicates the control wire routing and adds to installation costs. Also, placing the VFDs in the MCCs cancels some of the wiring cost savings provided by the new communications interfaces.
Serviceability of the drive may also become an issue. Placing the drive in an MCC will typically make the drive components harder to get to, and therefore harder to service or replace. If you need to upgrade to a new drive, it's much easier to replace a wall-mounted unit than a unit buried in an MCC bucket. Instead of buying a replacement VFD based on criteria such as ease of programmability or communications capability, your overriding concern should be to find a VFD that fits into the existing MCC enclosure.
Some people promote the “quick change-out” benefit of mounting drives in MCCs. However, most drives over 20 hp require more than the 7 in. to 9 in. of mounting depth typically available in MCCs with vertical bus. If you remove the vertical bus to accommodate deeper drives, you lose the flexibility of moving plug-in units to create more space within the enclosure. In other words, many higher horsepower VFDs in MCC enclosures are simply common AC bus feed drives. They aren't mounted in quick change-out buckets — they're actually floor-standing enclosures with a common AC feed.
Installing VFDs in MCCs is generally a bad idea, but it makes sense in certain situations, such as when you have dedicated equipment with small drives that can't be mounted near the motors. But consider such an installation carefully because other options may more readily give you what you're looking for. For example, you can move the VFDs to the controls portion of the specification, especially on projects with new building controls. The controls supplier will then be responsible for purchasing and coordinating the VFD systems. Don't settle on any particular layout until you understand all the design ramifications. Carefully look at your application and support from your vendors. If your project is daunting, hire an integrator who isn't beholden to any manufacturer. Then decide what to buy based upon what works best and is supported by the most reliable service plan.
Olson is manager of engineered drives at ABB, Inc., Automation Technology Products Division, Drives and Power Electronics, New Berlin, Wis.