Viewing page 2 of 2 pages.
Previous 1 2
TESTING
Testing Platform
The Zalman ZM300A-APF was installed in the same system as that used in the
Seasonic PSU review. It is a stable, very quiet, low airflow PC running Windows
98SE, fully updated:
| Case |
Landmark ATX-202 18" tower |
| CPU |
Pentium 4 - 1.6A (overclocked to 2.0 GHz) |
| Motherboard |
Gigabyte GA-81RXP |
| RAM |
512 MB 2100 PC DDRRAM |
| Video Card |
Matrox G400 Max (dual mode, driving two 18" monitors) |
| Hard Drives |
Both drives in cage at bottom of case, behind inflow case fan |
| Seagate Barracuda IV - 40 G |
| Seagate Barracuda IV - 20 G |
| Floppy Drive |
Generic |
| DVD Drive |
Toshiba SD-M1502 |
| CD-Writer |
Creative RW121032E |
| Network Card |
Intel Pro/100VE - Built into motherboard |
| Sound Card |
Creative SB PCI128 - Built into motherboard |
| Fans |
| 1 Panaflo 80mm "L" @5V over stock Intel heatsink |
| 1 Panaflo 80mm "L" @5V over video card / NB heatsink |
|
1 Panaflo 80mm "L" @4V lower front case fan
|
Test Instrumentation and Environmental Conditions
|
CPU temperature
|
Motherboard Monitor 5 reading CPU diode |
|
PSU temperature
|
Veriteq Spectrum 1000 with probe lodged in PSU
heatsink |
|
Fan voltage
|
Heath / Zenith SM-2320 Multimeter across fan terminals |
|
System power
|
Measured with Kill-A-Watt
Power Meter |
|
Noise*
|
Heath AD-1308 Real Time Spectrum Analyzer |
| Room |
27-29° C; ambient noise ~32-34 dBA |
Motherboard
Monitor 5 enables monitoring of temperatures and voltages off motherboard
sensors. With the P4 diode, there is little question of inaccuracy here; it
is usually accurate within 1° C. The Veriteq
Spectrum 1000 will be familiar to anyone who has read other articles on
this site: it is a highly accurate data logger that samples temperatures via
its probe. For all PSU temperature measurements, the probe was lodged next to
the thermistor on the top of one of the heatsinks. The Heath is an ordinary
multimeter that has proven to be fairly accurate.
Kill-A-Watt
Power Meter is a new addition to our test bench that provides accurate power
draw in Watts for all kinds of AC powered devices. In direct comparison, the
readings from the DIY
super simple power meter used before appear to be about 10% too low.
The Heath AD-1308 is a portable half-octave Real Time Spectrum
Analyzer with SPL meter functions. Below 40 dBA, its accuracy is limited to
3 dB increments, down to 23 dBA. Some 15 years old, this LED-based unit has
long since been displaced by digital devices with better interfaces to PCs.
(Shown on page
3 of the Seasonic PSU review.) The "A" weighting was used,
as recommended by numerous acousticians.(It most closely approximates the frequency
response characteristics of human hearing.)
The microphone on the sound meter was positioned about an inch
to the side of the PSU fan exhaust to avoid fan turbulence in the microphone
itself. The dBA obtained here cannot be compared to any other measurements due
to the lack of adherence to a repeatable standard and an uncontrolled reflective
environment. (Sorry, but we will not have access to that University of BC
anechoic chamber till September, at least.)
No effort was made to change acoustics in the lab, which is a small room measuring
12 by 10 feet, with an 8 foot ceiling. The PC sits on the uncarpeted hardwood
floor, under a table on 4 legs that supports the monitors.
NOTE about Room Temperature: Sharp-eyed readers will notice that the
ambient room temperature in some of the tests was upwards of 29° C, compared to 20-23° C in previous reviews. Summer arrived
for a few days, and the lab does not have air conditioning. For comparisons against the data in our other PSU reviews, please remember to factor in this ambient temperature difference.
At Startup
The Zalman starts at a very quiet level. Subjectively, the noise it emits is
a bit smoother and seems quieter than the recently reviewed Seasonic, but the
difference is difficult to measure with the Heath AD-1308 analyzer. There is
a difference in the quality of the noises emitted by the Zalman and the Seasonic:
the latter has a bit more of a soft clicking (bearing noise), which makes it
sound slightly rougher. Aside from fan noise, there is also a small amount of
coil buzzing in the Zalman, not audible when the PSU is running inside a PC.
The Zalman is louder than our virtually inaudible reference power supplies,
which run Panaflo 80 mm fans at 5V or less.
The following table shows measurements taken over 16 minutes with the system
idle, from bootup.
Startup Log A: 28-29° C room ambient
| Time (min) |
0
|
2
|
4
|
6
|
8
|
10
|
12
|
14
|
16
|
| CPU (° C) |
40
|
42
|
44
|
44
|
44
|
44
|
44
|
44
|
44
|
| PSU (° C) |
31
|
33
|
34
|
35
|
36
|
37.5
|
38
|
39
|
39
|
| System power (W) |
78
|
79
|
79
|
81
|
78
|
78
|
78
|
78
|
78
|
| Fan (VDC) |
5.52
|
5.65
|
5.86
|
5.98
|
6.09
|
6.14
|
6.21
|
6.29
|
6.31
|
| Noise dBA |
39
|
39
|
40
|
41
|
41
|
42
|
42
|
43
|
43
|
Note that the PSU temperature continued climbing after the CPU temperature
stabilized at 44° C. In the current room temperature (28-29° C) with
the PC at idle, the PSU heatsink temperature seemed to stabilize at ~39°
C, and the PSU fan voltage at around 6.2-6.3V.
This data conflicts somewhat with Zalman's graph, which shows the fan speeding
up at 42-3° C.
The next morning, with the ambient room temperature several degrees lower at
about 26° C, the startup log was repeated.
Startup Log B: 26° C room ambient
| Time (min) |
0
|
2
|
4
|
6
|
8
|
10
|
12
|
14
|
16
|
| CPU (° C) |
37
|
38
|
39
|
40
|
41
|
42
|
42
|
42
|
42
|
| PSU (° C) |
27.6
|
28.7
|
30
|
30.9
|
32
|
32.2
|
32.5
|
33
|
33
|
| System power (W) |
78
|
78
|
78
|
79
|
78
|
78
|
78
|
78
|
78
|
| Fan (VDC) |
5.5
|
5.65
|
5.81
|
5.80
|
5.87
|
5.82
|
5.89
|
5.83
|
5.91
|
| Noise dBA |
39
|
39
|
40
|
40
|
40
|
40
|
40
|
40
|
41
|
With Hard Use
Photoshop 6.5 and Adobe Framemaker 6.0 were opened with large files in both.
For good measure, an e-mail program and MS Internet Explorer were also opened.
The game Command and Conquer: Red Alert 2 was opened and displayed
on the main monitor. On the secondary monitor, work on this article was done
in Macromedia Dreamweaver 4 while leaving all the other programs on, with occasional
jumps to the game. Here are the results after 30 minutes.
| * |
Start
|
30 minutes
|
| Ambient |
26° C
|
26° C
|
| CPU temperature |
42° C
|
53° C
|
| PSU temperature |
34° C
|
39° C
|
| Fan voltage |
5.8
|
6.4 V
|
| System power |
81W
|
78W to 98W
|
| Noise |
40 dBA (A)
|
43 dBA (A)
|
With CPU Stability Test
CPU Stability Test by Jouni Vuorio
is a useful tool to stress systems. Here are the results over 45 minutes of
CPU stress testing. System power stayed at a constant 124W. The room ambient
as 26° C.
| Time (min) |
0
|
5
|
10
|
15
|
20
|
25
|
30
|
35
|
40
|
45
|
| CPU (° C) |
40
|
55
|
58
|
58
|
58
|
59
|
58
|
59
|
59
|
60
|
| PSU (° C) |
35
|
38.4
|
40.7
|
41.2
|
41.2
|
41.2
|
41.5
|
41.6
|
41.6
|
41.8
|
| Fan (VDC) |
5.84
|
6.18
|
6.87
|
7.02
|
7.07
|
7.1
|
7.14
|
7.15
|
7.14
|
7.19
|
| Noise dBA |
40
|
42
|
44
|
44
|
45
|
45
|
45
|
45
|
45
|
45
|
The measured noise difference between start and finish was about 5 decibels.
This is definitely an approximation, as environmental conditions were not ideal.
The subjective impression jibes with that number -- the fan did get louder,
and this difference was substantially louder than before. The sound level of
the PC at the end of this test is too high to be considered quiet by our standards.
However, running torture programs is not something the usual PC user is likely
to do often, and the noise was substantially less than with most other thermistor-controlled
fan PSUs.
Despite Zalman's charts to the contrary, the fan voltage varies closely with
heatsink temperature even when the latter is well under 40° C. The RPM and
sonic difference effected by a 0.1 to 0.2 Volt changed is subtle, however. Note
that the PSU temperatures did not rise much beyond 41° C, regardless of
continued load on the CPU, and the maximum CPU temperature reached was 60°
C.
Zalman vs Seasonic
The CPU Stability Test result is considerably different from that obtained
with the similar Seasonic SS300-APFC. Referring to the CPU Stability Test on
the Seasonic as reported here,
the following is a comparison of results at the end of the 45 minute CPU stress
test:
| * |
Zalman ZM300A-APF
|
Seasonic SS300-APFC
|
| CPU (° C) |
60
|
65
|
| PSU (° C) |
41.8
|
45.2
|
| Fan (VDC) |
7.19
|
4.84
|
| Noise dBA |
45
|
43
|
| Ambient (° C) |
26
|
23
|
| Power (W) |
124
|
114
|
A few things to note:
- CPU temperatures, fan voltages and ambient temperatures are probably accurate
to within 5%.
- The PSU temperature only tells of the temperature seen by the probe in a
specific spot one heatsink in each PSU. A small repositioning of the probe
or tighter coupling to the heatsink could result in different temperature
readings.
- The noise measurement is most subject to error, due to the limitations of
the measuring SPL meter and variations in background noise.
- CPU, PSU and case temperatures are all affected by the airflow provided
by the PSU fan, as it is a primary hot air exhaust "driver". This
is a basic cooling concept built into the ATX tower case specification.
- Although Zalman claims efficiency of 72-75% and Seasonic claims ~65%, the actual power measured was 10W higher for the Zalman. This suggests the Zalman sample is slightly lower in efficiency than the Seasonic sample.
- These are single samples from production lines that produce thousands of
units. It is not realistic to expect that every Zalman ZM300A-APF and every
Seasonic SS300-APFC will behave exactly like the samples tested here. A 5%
tolerance factor may be generous, but probably not unrealistic.
Keeping these factors in mind, the comparison suggests that
- The Zalman PSU does a better job of keeping itself and the case cooler.
Both CPU and PSU temperatures are significantly lower than with the Seasonic,
suggesting better hot air exhaust and throughput due to its higher fan voltage.
The large heatsinks in the Zalman may also contribute to the lower PSU temperature.
All of this is likely related to the efficiency ratings of the two PSU: 75%
for the Zalman vs. 65% for the Seasonic.
- The ambient temperature was 3° C higher when the Zalman was tested;
one can conjecture that at 23° C, the Zalman would have provided lower
CPU and PSU temperatures and/or lower noise.
- The tradeoff in the Zalman may be a slightly faster rate of fan noise increase
under load. However, the slightly lower noise of the fan in the Zalman may
offset its higher driving voltage.
CONCLUSION
The Zalman ZM300A-APF raises the stakes in the quiet PSU races. Impressively
constructed and designed for excellent cooling while maintaining low noise for
a wide range system loads, the new Zalman takes clear aim at the growing market
for quiet computer power supplies. The shot, while not quite dead center, is
close enough to make the ZM300A-APF a major contender.
PRO
|
CON
|
|
|
- Not as quiet as our reference modded PSUs
|
|
|
|
|
|
|
- Excellent build quality, great heatsinks
|
|
- Universal AC input; Active PFC
|
|
|
|
|
Our thanks to Zalman for the review sample and their support. And thank you,
Silicon
Acoustics, for getting us started with a Zalman PSU.
Zalman products are sold worldwide. Please check the Zalman
website for dealers in your area. Silicon Acoustics sells Zalman products,
of course.
* * * * *
Discuss this article in our Forums.
| Help support this site, buy from one of our affiliate retailers! |
|
