Benchmark Comparisons

To test the new HTPC case and especially the performance change when switching to the Zalman CPU and GPU coolers, we installed our standard ATX test bed with the Zalman provided ZM460-APS power supply. Measurements were made at 22.5 degrees Celcius ambient, and averaged over the course of 5 minutes after running for over 30 in the respective mode. Two instances of Folding@Home , rthdribl, and a comprehensive hard-drive loader are run simultaneously for the "load" test.

ATX Test Bed
DFI LanParty UT 915P-T12
Pentium 4 530 Prescott 3.0ghz
OCZ 512MB DDR2 x 2
Thermaltake Golden Orb II
Seagate 120gb SATA Hard Drive
Chaintech GeForce 6600GT
MSI DVD-CD/R/RW Combo drive
Zalman ZM460-APS PSU

Our standard test bed has been a solid set of hardware for reviewing for several reviews now, but one of the first things that we noticed when using the ZM460-APS power supply was how much less heat it put out than the MadDog unit, which we had been using before. Before getting into detail about the other differences, here's a chart showing the performance of the case first with the Thermaltake Golden Orb and stock 6600GT cooler, then with Zalman equipment in its place. The 80mm case fans were both hooked up to 12v connections, the video card into a 5v line, the CPU fan for the Golden Orb II directly into the motherboard, and the CNPS-9500 into the included Zalman FanMate II

Without much of a surprise, the VF700-Cu comes out looking the most impressive here - we suspected the stock cooler wasn't the greatest and our guess couldn't have been any more accurate. The CNPS-9500 is a strong performer too, but only marginally better than our already stellar Golden Orb II. In its defense, the CNPS-9500 was also lighter, easier to install, and as we'll see shortly, even quieter too. Another conclusion to make about these results includes the fact that motherboard MOSFETs really do benefit the most from downward facing fans, even if cross-blowing fans like the one in the CNPS-9500 does a fantastic job of keeping the CPU cool. Also, probably because of an overall cooler internal temperature, the passively cooled chipset also sees a rather decent benefit from the Zalman products.

Even more impressive is the noise change It should be noted that our new sound meter doesn't even register below 28dBa, and the ambient noise floor of the lab is pretty much a steady 29 dBa on the money. Seeing 30 on the meter from 6" in front was pretty hard to believe, but subjectively, it was completely understandable. With the lid closed, turning on the system was barely perceivable from a normal distance (3-6ft) away. With noise being such a concern in an HTPC environment, it's hard not to love this case; even with the standard ATX test equipment, these are some record low scores. Only the video card fan was really keeping it from being even subjectively lower.


Final Words

Zalman's display of their quality of engineering, design, and manufacturing have all been top-notch for each and every product that we've looked at here. With a closer look at the ZM460-APS in the near future, we will surely only add to this concensus, as even with our short amount of work with the unit here, we've grown very fond of the product.

Without a doubt though, the HD-160 is a fine HTPC case. The layout is smart, simple, and elegant. The construction leaves nothing to be desired and should stand up to many years of home entertainment. Its thermal and noise performance were record-breaking, and they only became better when we used the CNPS-9500 and VF700-Cu. If only a perfect, stealthy optical drive system could've been implemented, and a VFD where the digits were more visible from actual living room distances, this could've been the perfect HTPC case that we were looking for when we started. However, the new Zalman case is closer than many by a long shot, and for one who doesn't want to settle for something that "just works", the HD-160 is a great choice, even for the relatively high asking price of around $250. For those who really want to go all out, we'd readily recommend all three products that we looked at in detail; each stands out in its own right.

Zalman Components Installed
Comments Locked

48 Comments

View All Comments

  • Stele - Monday, April 3, 2006 - link

    Good post, well said.

    As for using BJTs in audio power amps, I agree with you. It's a good thing... properly selected and implemented, they're generally known to give better sound than FETs (especially in the form of substantially lower distortion and noise figure), as much as triodes/vacuum tubes have excellent distortion and noise characteristics. I used an all-discrete, all-BJT design for an audio power amp competition several years ago; trickier to get right but ultimately satisfying sound quality compared to some designs that took the short cut and dropped in power-amp ICs (mostly from ST, IIRC). :P
  • Stele - Sunday, April 2, 2006 - link

    Hi topher42,

    quote:

    What are you calling integrated?
    Most power fets are one transistor and nothing else in the package.
    Not much integration.


    Now how did that slip by me... yes, you're perfectly right. Was trying to be careful not to make that mistake but in the end still tripped on that common misuse of the term. Thank you for pointing it out, I stand corrected :)

    20kW amps??? That's powerful alright, no surprise about being a headache to keep cool :)

    I mentioned TO-3's along with the other packages to illustrate my point that discrete MOSFETs don't always have to come in 3-legged form, though you're right that they're a big too big for motherboard mobos.

    quote:

    The EE's I deal with that don't know a fet from an asic are usually the digital
    designers designers that never get closer to a part than their VHDL code.


    I see your point, and it kind of explains certain things nowadays... oh well, gone were the hands-on approach of the good old days :P


    Hi Clauzii

    As with topher42 you're right that a discrete MOSFET is not an 'IC', srictly-speaking. I really must be more careful not to mix them up!

    However, a MOSFET is not really a "constructional technique in a transistor"; rather, it is a transistor (the most common type of FET, of which there are several kinds) that is made using a particular constructional technique (metal/polycrystalline on gate oxide on silicon etc as opposed to bipolar junctions on BJTs) and that works using a particular method of operation (field effect, as opposed to BJTs). :)


    Hi AnnonymousCoward,

    quote:

    Most basic, yes, but not the most common.


    Granted not the most common in the objective sense, and you may not need to work with them much unless you're involved in, say, power circuitry design. After all, you did mention that you normally use BJTs more. :)

    However, it is really common enough in terms of availability and application that it's about universally known, like the 'regular' BJT, especially in (but not limited to) the semiconductor community/industry. Besides, major semiconductor companies sport hundreds to more than a thousand of such devices in their product lists and they're used in an enormous number of electronic devices, from handphones to motherboards to power amplifiers (audio and electrical).

    quote:

    So, in power-controlling applications, when are discrete MOSFETs used and when are relays used?


    MOSFETs are used in many (dare I say most?) modern power supply applications in electronic devices these days, especially in switch-mode (as opposed to linear) power supplies. Briefly, the MOSFETs in such power supplies are used as digital 'switches' that are rapidly turned on and off a hundred or more times a second to provide power - the amount of power required is varied by changing their duty cycle (how long they stay ON or OFF). This is a whole field by itself, so no-one can blame you if you're not thoroughly familiar with it (I don't pretend to be either! :P).

    Relays are used as mechanical switches that isolate two or more circuits yet allow one to control the other(s). A common (but certainly not only use is to allow a low-voltage low-current circuit to control/switch a high-voltage high-current circuit, e.g. when you use a battery-powered light-sensing circuit to turn your room light on.

    Opto-couplers and FETs (in the form of solid-state relay ICs) can do the job too, but relays are cheap, readily available, fairly reliable, easy to troubleshoot and generally can handle more current and voltage than a semiconductor counterpart. Of course, relays are bulky, noisy, can introduce damaging surges in the energising circuit when switching on and off (due to the inductance of the coil), and cannot switch as fast as a semicon equivalent, but that's the way it is :)
  • Clauzii - Sunday, April 2, 2006 - link

    To repeat myself:

    MOS-FET is a constructional technique in a transistor, be it a small one inside a IC or a single "gate" in a device.

    An IC is more than one component in the same housing. The MOS-FETs used on mobos are NOT ICs, but Field Effect Transistors made using Metal On Silicon process.
  • Clauzii - Sunday, April 2, 2006 - link

    PS: When do DT check up on their f...... Forum System - it´s flawed like h... :)
  • Clauzii - Sunday, April 2, 2006 - link

    To repeat myself:

    <b>MOS-FET is a constructional technique in a transistor, be it a small one inside a IC or a single "gate" in a device. </b>

    An IC is more than one component in the same housing. The MOS-FETs used on mobos are NOT ICs, but Field Effect Transistors made using Metal On Silicon process.
  • topher42 - Sunday, April 2, 2006 - link

    stele.

    "especially since a MOSFET is an IC."

    What are you calling integrated?

    Most power fets are one transistor
    and nothing else in the package.

    Not much integration.

    Most of the 6 pin devices I have used
    were for low current noise differential
    inputs were you wanted identical devices
    and temperature tracking.

    And I have used the big packages when
    designing and building 20 kilowatt
    amplifiers. Pain to cool....

    I was pointing out that the bios would
    have a problem controlling the cpu voltage
    without power fets. to3's would be a
    little big for that app.

    The EE's I deal with that don't know a
    fet from an asic are usually the digital
    designers designers that never get closer
    to a part than their VHDL code.
  • Clauzii - Saturday, April 1, 2006 - link

    Nice with a guy that knows what he´s talking about :)
    Also, almost all amplifiers for professional use (including old estimated Hafler P500, Crown, Ampeq etc.) are build whith MOS-FETs in single TO-3 (metal) etc. housings.
    MOS-FET is a constructional technique in a transistor, be it a small one inside a IC or a single "gate" in a device.
  • nullpointerus - Saturday, April 1, 2006 - link

    I'd just like to put my vote in for "geeky electrical thingies."

    Seriously, does this matter to 99% of the readers?
  • SonicIce - Friday, March 31, 2006 - link

    It's official! Joshua Buss of Anandtech does bad case wiring jobs!
  • JoshuaBuss - Friday, March 31, 2006 - link

    Believe it or not, I don't do 'great' wiring jobs inside cases for testing as a matter of choice. Typically, users don't spend enough time routing cables neatly enough to stop all airflow blocking problems, thus I feel it's more accurate to actually leave a little cable clutter behind. Besides, if I spend much more time on it than the average user would, I'm not providing a good sense of how easy it is to do the cable management - I'd simply be showing off. The best way to give people a sense of how the case is laid out is to spend roughly the same amount of time in each case trying to clean them up a bit and see how far I get. With this case, I was able to get the majority of the mess out of the way with almost no effort, which is a testamanet to the ample space inside the HD-160.

Log in

Don't have an account? Sign up now