• Home
  • blog
  • Gigabyte GV-N66256DP Fanless Graphics Card

Gigabyte GV-N66256DP Fanless Graphics Card

Not many manufacturers offer a good performance AGP card with fanless cooling. The upgrade market still consists mostly of AGP-based systems, and the GeForce 6600 represents close to the high water mark for AGP performance. This is why Gigabyte’s fanless GeForce 6600 is still worth a close look with our new graphics card testing methodology.

April 8, 2006 by Devon

Gigabyte GV-N66256DP
Fanless GeForce 6600 for AGP
Market Price

The GeForce 6600 is old news, based on yesterday’s technology.
The same goes for AGP. Even fanless cards are no longer very newsworthy, as
most major manufacturers now offer one or two fanless models. However, not many
manufacturers offer a good performance AGP card with fanless cooling.

This is why Gigabyte’s fanless GeForce 6600 is still worth looking at, even
after the market has moved on. The upgrade market still consists mostly of AGP-based
systems, and the GeForce 6600 represents close to the high water mark for AGP

Gigabyte was one of the first manufacturers to recognize a demand for fanless
graphics cards. For a time, their heatpipe-cooled cards were the only game in
town if you wanted a fanless graphics card that wasn’t an entry-level performer.
The GV-N66256DP is part of Gigabyte’s original lineup of heatpipe-cooled cards.
While a 6600 may be considered entry-level these days, it’s still a capable
performer as an upgrade to an older system.

As usual, the retail box is several times the size of the card.

Gigabyte did not develop the fanless cooler on their own. A sticker on the
box reads “Cooled by Zalman”. Zalman is well known as a noise-conscious
manufacturer of CPU heatsinks, so it’s no surprise that Gigabyte decided to
use their expertise. The result of this collaboration is what Gigabyte calls
“SilentPipe” technology.

The heatsink is provided by a company well known for designing low-noise

As can be expected from an entry level gamer card, the accessories are fairly basic:

  • A brief instruction manual, with basic installation instructions and an
    overview of nVidia’s drivers. Troubleshooting tips are minimal.
  • A driver CD
  • Bundled game: Thief – Deadly Shadows
  • DVI-A to D-Sub adapter to allow the DVI port to be used with an analogue
  • A TV breakout box with component and S-Video output

Standard accessories.


(from Gigabyte’s
web site
Chipset NVIDIA GeForce 6600 + HSI
Memory 256 MB
Memory Bus 128 bit
Memory Type DDR 16M x 16
Bus Type AGP
Bus Speed 8x
D-Sub Yes
TV-Out Yes
DVI Port Yes (DVI-I)
Multi View Yes

By now, the performance of the GeForce 6600 should be no surprise to anybody.
Avid gamers will no doubt turn up their noses and shell out for the latest and
greatest, but performance should be fine for casual gamers.

A more likely market for the GV-N66256DP is home theater,
not gaming. Low noise is essential for home theater components, so
a fanless graphics card makes good sense. And, because 3D graphics are
rarely used in a home theater, 3D performance is not relevant.

An external power source is needed.

Not mentioned in the specifications is the presence of an
external power connector. This connector indicates that the AGP slot may not be able to supply enough power to operate the
card. The PCI Express version of the same card does not have a power
connector. Even though the card exceed the power delivery possible via
AGP, it is within the capabilities of the PCI Express bus. The maximum power
available through PCI Express is 75 watts, so the GV-N66256DP can safely be
said to draw less than this. It’s actually a lot less than this.


The gold heatsink is quite distinctive, especially in contrast
against the royal blue PCB. There are actually three separate heatsinks: A small
one for the GPU chip itself, a larger one for the rest of the board, and another
large one on the back side of the board that allows the heat to be spread across
both sides of the card. The two larger heatsinks are connected by a single heatpipe.

There’s actually two separate heatsinks visible here. The one on the lower
right cools the GPU.

A heatpipe transfers heat to the back of the card.

To understand how the cooling system is supposed to work, you need to picture
how the card will be positioned in a typical case once it is installed. In a
standard ATX tower style case, the AGP card is the topmost expansion card in the system.
The card is installed “bottom up” with most of the main components
hanging underneath the card.

This arrangement is very poor for passive cooling, since most of the heat generated
by the card gets trapped underneath it. The solution is to somehow transfer
the heat to the above the card, where the airflow from the CPU and system fans
can carry it away. In the case of the Gigabyte card, this is
achieved by using a single heatpipe that wraps its way over the top edge of
the card and feeds the heat to the “topside” heatsink.

A wave of fins designed to take advantage of system airflow.

There is a small gap between the heatsink and the card so that air can flow underneath
it as well as across the main fins. As the photo above shows, the fins are arranged
in the shape of a wave, with the crest of the wave near the tail end of the
card. Here, they are more densely packed than they are closer to the tail of
the wave.

This design is not merely cosmetic. It is probably meant to take advantage of typical
system airflow in the case. The airflow normally flows to the back of the case, forced by a back panel exhaust fan and the PSU fan. The crest of the waved fins will
probably see the strongest airflow, and the tail of the wave the weakest. This may explain
why the fins are farther apart at the tail of the wave; with less airflow, they
need to be more widely spaced to cool at optimum efficiency.

The rear-mounted heatsink is a good way of getting around the way the
awkward thermal conditions for expansion cards in vertical cases. This approach
has been adopted by almost all of the more recent cards on the market. However,
there’s one disadvantage: In moving the heatsink to the rear of the
card, the card is thicker than allowed by the ATX tower form factor. Under
ordinary circumstances this doesn’t matter. Conflicts with other cards are all
but impossible since the graphics card is invariably the topmost card in the
system. However, some motherboards have big northbridge heatsinks or other components
that do not allow enough space for a heatsink above the AGP slot. The problem
is most common on micro-ATX boards where the components are often more tightly packed.

This exotic looking heatsink faces downwards in a vertical ATX system.

Although the bulk of the heat is likely to be dissipated by the rear-mounted
heatsink, there is also a sizable heatsink on the face of the card. Not much
airflow can be expected across this side of the card, and the design of the
heatsink reflects this. In comparison to the rear heatsink, the fins are thinner
and more widely spaced. Several of them are also hooked, presumably to increase
the surface area without having the fins infringe on the expansion slot below.
These fins should benefit from whatever air manages to curl its way out from
underneath the card.


This sturdy modified LX-6A19 (D8000) case from Cool Cases became our test system housing.

For this review, we designed a
graphics card testing methodology that we hope to use in future reviews. It is an in-system
test, designed to determine whether the card is capable of being adequately
cooled in a low-noise system. By adequately cooled, we mean cooled well enough so that no misbehavior related to thermal overload is exhibited. Thermal misbehavior in a graphics card can show up in a variety of ways, including…

  • Sudden shutdown without warning.
  • Jaggies and other visual artifacts on the screen.
  • Motion slowing and/or screen freezing.

Any of these misbehaviors are annoying at best and dangerous at worst — dangerous to the health and lifespan of the graphics card.

The test system was built around a midrange Pentium 4 Northwood processor
as an example of a mid-powered system that is fairly easy to keep quiet.
The system is used for our heatsink reviews, so we are already intimately
familiar with the thermal characteristics of the processor and motherboard.

Test Platform

  • Intel
    The Thermal Design Power of this P4-2.8 (533
    MHz bus) is 68.4 or 69.7W depending on the version. As the CPU is a demo model
    without normal markings, it’s not clear which version it is, so we’ll round
    the number off to ~69W. The Maximum Power, as calculated by
    & CPUMSR
    , is 79W.
  • AOpen
    AX4GE Max
    motherboard – Intel 845GE Chipset; built-in VGA. The on-die
    CPU thermal diode monitoring system reads 2°C too high, so all readings are
    compensated up by this amount.
  • Scythe Shogun heatsink, cooled by a Nexus 120mm fan undervolted
    to 7V.
  • OCZ
    PC-4000, 512 MB
  • Seagate Barracuda IV 40G 1-platter drive, sitting on foam in the
    bottom of the case
  • Antec Neo HE 430
    ATX12V 2.01 compliant power supply
  • Modified LX-6A19 (D8000) case from Cool Cases, outlined in detail below
  • Nexus 120mm fan controlled by a variable voltage fan controller

Measurement and Analysis Tools

  • CPUBurn
    processor stress software
    graphics demo to stress the GPU
  • 3DMark05
    benchmark software, used as an alternate means of stressing the GPU
  • SpeedFan
    version 4.25
    software to show CPU temperature
  • Seasonic Power Angel AC power meter, used to measure the power consumption
    of the system
  • A custom-built variable DC power supply that allows us to dial in exactly
    what voltage is powering the system fan

System airflow is quite good, allowing the CPU and system fans to run at close
to inaudible speeds without compromising system cooling. The intake is about
the size of a 120mm fan. The only restriction is an air filter. A much more
restrictive cover for the filter was removed because it impeded the airflow
too much.

The one and only intake…

…and the same view, with the bezel removed.

There are two points of exhaust: The 120mm case fan, which will be run at a
number of different speeds, and the 80mm fan in the Neo HE power supply. The
case fan can be expected to exhaust the bulk of the heat, since the
Neo HE is unlikely to ramp up
even at the maximum power draw that the test
system is capable of. The amount of airflow through the system can be controlled
by adjusting the speed of the case fan, thereby giving us a way of controlling
how difficult the thermal environment inside the case is.

Only two possible points of exhaust: The orange case fan and the fan in
the power supply.

The airflow in our test rig is typical of an ATX case. Air flows in through
the intake near the bottom of the front panel, and is pulled up to the top rear
corner. Most of this air will bypass the expansion cards altogether, but a small
amount will be pulled across the rear of the card as it is pulled between the
CPU heatsink and the case fan. The bulk of the air will be pushed through the
CPU heatsink.

The air will flow from the lower right to the upper left, drawing a small
amount of air across the VGA card.

Thermal testing consisted of running CPUBurn and RTHDRIBL simultaneously to
generate as much heat as possible. An initial test was run with the system fan
running at 12 volts, and then the fan was progressively slowed down to make
the thermal environment more difficult.

Because no thermal monitoring was available on the GV-N66256DP, CPU temperature
was used to determine when the temperature had stabilized. Once the temperature
was determined to be stable, the stress software was left running for at least
another 20 minutes while we watched the screen carefully for visual
artifacts that might indicate overheating. The last test, with the system fan
running at 5 volts, was left running for more than an hour.


Ambient conditions at the time of testing were 22°C, and 121V / 60 Hz.

Initial testing was done with the integrated graphics on the
system motherboard to establish baseline levels for system power and CPU temperature.
Unfortunately, neither RTHDRIBL nor 3DMark05 would run on the integrated graphics
chip, so only the only load on the system was CPUBurn. Proper stress on the
integrated graphics might raise the results by another 5~10W, but the impact
on the rest of the system temperatures would probably be minimal.

VGA Test Bed: Baseline Results (no external VGA card)
System State
System Fan Voltage
System Power Consumption (AC)
CPU Temperature

Once the baseline testing was complete, the Gigabyte GV-N66256DP was installed, and the
real testing began.

VGA Test: Gigabyte GV-N66256DP
System State
System Fan Voltage
System Power Consumption (AC)
CPU Temperature
CPUBurn Only
155W (peak)
140~150W (typical)

CPU temperature changed very little compared to the original
baseline results and power consumption changed very little with the speed of
the system fan. For this reason, only a few of the more interesting results
are shown in the table above; the rest are redundant.

Even with the system fan running at its slowest speed (5 volts), no signs of
visual artifacts were ever present, so the card passed our test. The card
has since been installed in a permanent system where it has been running World
of Warcraft
and Lineage II: The Chaotic Chronicle flawlessly for the last two

An estimate of the power consumption of the graphics card can be obtained by comparing
the total system power with and without the card. The
following procedure was used to derive the estimates in the table below:

  1. The system AC power consumption was measured at various loads, with and without the Gigabyte card.
  2. We used efficiency test results from our Antec Neo HE 430 review to estimate the approximate DC output power being delivered at the various AC input power measured.
  3. The power consumption of the graphics card at idle was assumed to be the difference
    in power demand between the two systems when both were running CPUBurn.
  4. The power consumption of the graphics card under load was assumed to be the
    difference between the system with the card running CPUBurn and RTHDRIBL simultaneously,
    and the baseline system running CPUBurn only. This ensured that any load on
    the CPU from RTHDRIBL did not skew the results, since the CPU was running
    at full load in both cases.
Video Card Power Consumption: Gigabyte GV-N66256DP
GPU State
Estimated DC Power Consumption
Total increase in System Power (AC)

As a final practical test of the card’s capabilities, we ran five consecutive
iterations of the 3DMark05 benchmark. Power consumption averaged about 10~15
watts lower than the full CPUBurn + RTHDRIBL test. This is a much more realistic
test of the kind of load the card will be under in actual use; gaming is a dynamic load, and very few games place the system under the kind of
continuous load that CPUBurn and RTHDRIBL achieve. As with the RTHDRIBL test, no
visual artifacts were detected.


The GV-N66256DP is still a worthwhile investment
as a no-noise upgrade to an older AGP-based system. Our efforts at overheating
the card all went for naught, and it performed flawlessly even with the most minimal airflow cooling in our test bed. The Zalman-designed heatsinks appear to work perfectly well.

This achievement is all the more impressive because the amount
of power consumed by the card is not trivial. 32 watts may be relatively little for a graphics
card these days, but keep in mind that a 32 watt CPU would need careful airflow
design and large heatsink to be cooled passively. The fact that Gigabyte
managed to overcome the cooling limitations of the ATX form factor is impressive.

Many thanks to Gigabyte
for the GV-N66256DP sample.

SPCR Articles of Related Interest:
Chaintech AA6800GT + Arctic Cooling NV Silencer 5
Arctic Cooling ATI Silencer 2 VGA Cooler
Arctic Cooling VGA Silencer

Zalman ZM-80 VGA Heatpipe Cooler

* * *

Discuss this article in the SPCR Forums.

Leave a Comment

Your email address will not be published. Required fields are marked *