1、NA-04-3-3 Plenum Fans in HVAC Equipment: The Good, the Bad, and the Ugly Kim G. Osborn Associate Member ASHRAE ABSTRACT This paper provides an overview of the benejts ofplenum fans and a detailed discussion of some of the pitfalls. Afer covering some of the benejts leading to extensive use of plenum
2、 fans, included is a brief discussion of the major complaint expressed about plenum fans, which is that plenum fans are less ,eficient than housed fans. Finally, the bulk of the paper covers problems that can result from poor design prac- tices, sloppy construction, and careless handling. INTRODUCTI
3、ON Plenum fans have become popular with some segments of the HVAC industry because their use can lead to shorter cabi- nets (read less expensive) and quieter applications. As with all engineering choices, there are trade-offs and pitfalls to be considered. Plenum fans cannot compete with properly ut
4、ilized housed fans for efficiency. Properly utilizing a housed fan means a “draw-through” unit. This configuration puts the fan at the discharge end of the unit, drawing the air through the coil. The fan must be connected directly to the ductwork with two to three equivalent diameters of straight du
5、ct before any turns. In many applications, however, housed fans are not used in “draw-through” configurations or are not properly ducted. In these configurations, the efficiency gap narrows consider- ably. As with any component, you must exercise care in design, construction, and handling of plenum
6、fans. Several examples will be presented, documenting problems that were traced back to problems with frame design, component alignment, and component handling. These issues are important for all fans but some at least can have a uniquely plenum fan twist. For example, any fan can have a problem wit
7、h frame reso- nances, making them impossible to balance properly. Plenum fans, however, with their large, flat inlet plates can transmit these resonance frequencies into the airstream like a loud speaker. Sound and Airflow Data All data presented in this paper were acquired in a labo- ratory set up
8、to acquire airflow data in accordance with AMCA Standard 210 (AMCA 1985) and sound data in accordance with AMCA Standard 300 (AMCA 1995). The laboratory is accredited under AMCA 1 1 1 (AMCA 1989) for performing the Standard 300 testing. It is not accredited for the 210 tests, but enough certified fa
9、ns have been tested here to be confident of the airflow measurements. In any case, most of the data presented are comparative; thus, all that really matters is that the measurement procedures be consistent. Some fan data presented are projected from the test data using methods set out in AMCA standa
10、rd 301 (AMCA 1990). Also, some of the sound data are not presented as specified in AMCA Standard 300. First of all, some data are presented for the frequencies below the 50 Hz one-third octave band, which are not addressed in the standard. Also, much of the data are presented in one-third octaves, r
11、ounded to the nearest tenth dB, rather than summed to full octaves and rounded to the nearest integer, This is done where it better illusfxates the point being made. The testing standards list the measurement error for the test procedure to be zk6 dB for the 63 Hz octave band and it is with poor imp
12、lementation and occasionally with quality control. The latter, of course, is a problem with all varieties of fans. Poor Discharge Opening Location As previously mentioned, using plenum fans can result in shorter cabinets. When used in a draw-through configuration, this is sometimes pushed too far, a
13、nd the discharge is placed ASHRAE Transactions: Symposia 61 9 Frequency Supply Lw- wlscreen Supply Lw- wlo screen 1 O0 Supply Sound Power Vibration Induced Noise 1 7 77.0 77.5 I l - 125 160 200 I l i l l ascharge LW I 82.0 82.0 84.5 85.0 85.9 88.8 Figure 4 Excessive movement on the frame can result
14、in excessive sound at the fan rotutionulfiequency or u multiple thereoj - 250 73.9 74.3 315 14.0 74.1 lengthened to move the opening out of the discharge zone, the next best thing to do is to place something between the fan and opening to break up the pressure pulses. Poor Frame Design A common prob
15、lem with all fans is frames withresonance problems. This can make the fans impossible to balance and leads to excessive vibrations being imparted to the building structure. Plenum fans, however, have a special problem. The inlet cone of a plenum fan is mounted on a square piece of sheet metal (inlet
16、 plate) attached to the frame. The net result can be a large speaker diaphragm. Figure 4 shows one-third octave data for a 27 in. (686 mm) plenum fan in a prototype frame that has resonance problems. The circle marks the location of a vibration-induced tone. The fan rotation speed is 1,823 rpm, whic
17、h gives a primary rota- tional frequency of 30.4 Hz. The noise tone is at the third harmonic frequency (91.2 Hz). Visual observation led to the suspicion that the noise was caused by excessive inlet plate movement. This was borne out by further tests. Figure 5 shows the FFT display of the output fro
18、m two accelerometers. One was located at the base of the inlet plate on the frame, oriented horizontally. This is the dotted curve in Figure 5. The other was located near the top of the inlet plate, also oriented hori- zontally. This is the solid curve in Figure 5. On this run, the fan rotational sp
19、eed was 1,836 rpm, which put the rotational frequency at 30.6 Hz, and there was a substantial spike in the FFT at 91.8 Hz (third harmonic). The plate velocity at this frequency is substantially higher near the top of the inlet plate than at the bottom, which indicates excessive movement ofthe inlet
20、plate. It is this movement that generates the tone. The frame had to be redesigned with additional support cross-brac- ing for the inlet plate to reduce the vibrations. 620 ASHRAE Transactions: Symposia 270 Inlet Plate Movement 21 I i$ i- go /c Eo 18 C , .- ,I ii o O0 *w O 2000 %O o 4000 5000 WO Fre
21、quency -. -u*p- -lower1 270 Inlet Plate Movement 10000 o o) u) .- b E o O - 1.ooc 0.1W Figure 5 This FFT display shows the relative motion between the top and bottom of the inlet plate. Comparison of 49“ Fans . . _ _. - ._ . ._ - Comparison of 49“ Fans 2.500 2.W 2 $ vm e v) P 3 3“ i.oM) b? 500 o 5,0
22、00 io.ooo ispoo mpoo 2900 3opoo SPDO Airow - L /Sec _. . _ .I . .i Figure 6 These airflow curves show the amount of variabiliq possible between fans of the same design from the same manufacturer. Poor Quality Control Quality control issues are not unique to plenum fans or to fans in general, but in
23、the process of troubleshooting air- handler problems, this should be one of the first items to check out. A number of problems can occur, ranging from outright damage to improper alignment of components. Figure 6 shows fan curves for three 49 in. (1,245 mm) SWSI centrifugal plenum fans. These were,
24、in theory, identi- cal fans from the same manufacturer. The best performing fan in terms of airflow was the prototype for the model. This fan was at least partially hand-built, and several earlier wheels had been rejected. The next fan tested, and the poorest performing, was a fan removed from a uni
25、t that had failed an airflow test. As shown, the fan performs substantially below the prototype. As we were using the prototype data for fan selection, this prompted an emergency test program. We took a randomly selected “production“ wheel with inlet cone and placed it in the same frame in which the
26、 prototype wheel had been tested. Although the airflow of this wheel was much better than the one removed from the unit, it still didnt perform up to the level of the prototype. It was also louder than the prototype. Table 3 shows the octave band sound power for the prototype fan and the production
27、fan. The production fan is substantially louder at the blade passage frequency and below. As it had previously been assumed that the prototype sample would be representa- tive of future shipments, the course of action was to publish a revision to the selection database. This episode illustrates the
28、vulnerability of assuming that one particular fan tested is representative of all future specimens. Although this has not been done, it would probably be best to test at least five randomly selected production models for each size of fan to be ASHRAE Transactions: Symposia 621 Table 3. 62 125 250 50
29、0 lk 2k 4k 8k LWA Discharge Inlet Table 4. 98 98 103 102 96 91 87 83 102 93 97 97 93 90 85 83 78 95 Poor Assembly Discharge Inlet tested and develop some method of averaging the performance for catalog and selection program purposes. 104 104 107 103 97 91 88 82 104 99 101 1 O0 93 89 84 82 77 96 Ship
30、ping Damage As with the previous issue, this is not unique to plenum fans but is still worth noting. One of the most vulnerable components of the plenum fan is the inlet cone. During a unit test, it was noted that one of two return fans was substantially louder than the other. After modiing several
31、aspects of the installation, trying to determine what could cause one of two adjacent fans to be noticeably louder, it was decided to remove the fan to see if it performed normally when standing alone. It was still louder than normal. The next course of action was to change various aspects of the fa
32、n assembly looking for the problem, even to the point of changing the fan wheel. Nothing appeared to be wrong with inlet cone, but as the final test, this also was exchanged. The fan performed normally. Table 4 shows the inlet sound for the fan before and after exchanging the inlet cone. As stated,
33、there was no obvious damage to the cone, but careful measurements showed that one side of the cone was compressed about 318 of an inch (10 mm). The damage was probably caused in ship- ping or storage. The cones are typically stored in stacks. This cone was probably near the bottom of a stack, which
34、was dropped a few inches. There are two lessons from this. The obvious one is to treat fan components with care. Particularly, inlet cones are easily damaged. The second lesson is that when faced with a roublesome fan, you may have to look-or measure-very carefully to discern the damaged component.
35、Again, this is not unique to plenum fans, but a common problem with plenum fans is improper positioning of the inlet cone in the opening of the fan wheel. With a housed fan, the position of the inlet cones in the fan wheel is actually set by the width of the scroll, and the assembler only needs to p
36、osition the wheel on the shaft such that the wheel is centered between the cones. With plenum fans, the correct position must be defined, and the assembler positions the wheel according to these measurements. The most common problem is the inlet cone being inserted too far into the collar of the fan
37、 wheel. The primary symptom of this is poor airflow performance. Depending on size, being a quarter inch too far into the wheel can cause a fan to be 5% or more low in airflow. Figure 7 shows two fan curves for the same fan. The curve on the left shows the fan as it was first installed. Being obviou
38、s that the fan was not performing as expected, the wheel was moved along the shaft away from the inlet cone approximately 3/8 in. (1 O mm). The results are shown in the curve on the right side. Here, the increased airflow was approximately 3%-not a great differ- ence, but enough to be the difference
39、 between passing an airflow test and failing it. The proper position for the fan wheel, relative to the inlet cone, is usually provided by the fan manufacturer. These points have been determined though extensive testing to deter- mine the position delivering the maximum efficiency. Often, this posit
40、ion does not deliver the maximum airflow perfor- mance possible. Therefore, with some fan designs, if you need to achieve a small increase in airflow and can stand a small decrease in mechanical efficiency, you can usually move the fan wheel a small amount away from the inlet cone (so the cone has a
41、 shallower penetration into the wheel) and increase the airflow a percent or WO. Tuning Related to Unfortunate Component Spacing As this has only been observed on a few tests, it may not represent a major problem but does lend some interesting insight to plenum fan usage and system design. Figure 8
42、and Table 5 show data for a unit that was tested in the lab. The test was observed by an acoustical consultant sent by the customer. Although the units octave band numbers were within toler- 622 ASHRAE Transactions: Symposia 49“ Plenum Fan: Cone Position 315 60 “ 5.0 s C 4.0 8 3.0 P .- e 8 .$ 2.0 *
43、$.O 0.0 69.7 68.6 -1.1 ZWMI 2m 3oMw 35000 4oow 45000 500 Airflow - CFM . ._ 4 Corr Ps eoor pos L-crr Psgood pos 49“ Plenum Fan: Cone Position 1,600 1,400 n“ t,2W ;j 1,000 pi o0 em .% 600 u 400 200 O P 10,000 12,000 14,000 16,000 18,000 O0 Aimow - USec -_ . Figure 7 All fans are sensitive to the posi
44、tion ofthe inlet cone relative to the fun wheel. Cabinet Test 75 70 65 60 55 50 45 40 120 130 140 150 160 170 180 Frequency 1 - - Lp with perforated diffuser plate -Lp without perforated diffuser plate I Figure8 A standing wave developing inside the cabinet due to component spacing can enhance the b
45、lade puss ofthe fun. An inlet diffuser can sometimes decrease the efect. ance, the consultant felt the sound was too tonal and, indeed, the one-third octave band containing the blade passage frequency was higher than expected, based on the projected level of the bare fan. The airflow was correct, in
46、ferring that the inlet cone was not damaged, so we had to start looking for other causes. The consultant observed that the distance between the fan inlet and the coil up stream of it was approx- imately one-half the wavelength of the blade passage Table 5. Inlet wio Inlet wi 69.8 65.4 65.5 ASHRAE Tr
47、ansactions: Symposia 623 reduces the blade passage frequency sound down to the expected level, it does add some credibility. The conclusion here is that when working on a sound sensitive project with plenum fans and a coil immediately upstream, it may be a good idea to look at the distance between t
48、he fan inlet plate and the coil. If possible, avoid the distance of one-half wavelength of the blade passage frequency. This may be accomplished by changing the physical spacing or by using a fan with more blades, which will increase the blade passage frequency. 32 Poor Return Plenum Design The most
49、 persistent problem we run into with plenum fans has been short return air plenums with the return air coming in at right angles to the fan axis. That is, the return air opening is on the floor, the ceiling, or one of the side walls. This causes the fan to side load; that is, most of the air enters the fan on one side. Side loading increases the sound power, often reduces the airflow, and causes “unstable” airflow. What is observed in these situations is that the airflow rises and falls through a greater range than normally observed, and the static pressure rises wit
