Volume 1

Issue 10
November 2007

The newsletter for the thermal management of electronics

FUTURE COOLING

In this issue: Thermal Characteristics of Liquid

.0
Coolants for Liquid Loop Cooling
1 Future Cooling
The continued miniaturization of chips, The best liquid coolant is ionized water
0.8 and the resulting increase in their heat because of its thermophysical proper-
flux, is driving the demand for liquid ties. But, because of its low freezing
ate

cooling. There is a mistaken belief in the point, it can cause problems at low
R

0.6 marketplace that air cooling has reached temperatures. Adding ethylene glycol
fer

8 Thermal Minutes its limit and liquid cooling will solve new or propylene glycol can bring down the
s n

freezing point, but it significantly reduces

Tra

thermal problems. The truth is liquid

cooling typically transfers heat from a the heat transfer characteristics of the
t

0.4
a

space-constrained area to a roomier one water.

e H

where a heat sink or head exchanger

tive

can work. In other cases, it Potassium formate (PF) has been pro-
posed as the best candidate for liquid
a

13 Thermal
0.2Analysis
is used to minimize thermal spreading
Rel

resistance. At the end, a liquid-to-air cooling applications [2]. The PF coolant

e
n t Typ exchanger is needed to release
heat is non-flammable, non-toxic and not
la very corrosive. Inhibitors can be used to
Coo the heat to the ambient air.
r further minimize corrosion.
0.0 W ate
There are te many coolants available
/
r too r ma from when planning for a
choose Mohapatra, et. al. [2] investigated six
te F
Wa ssium specific r cooling liquids that are commercially
18 Thermal Fundamentals t a
Po anol a good / W ate application. In addition to having available today:
th l/W a ter heat transfer rate, factors such
M e as corrosiveness, freezing point, toxicity,
t h ano boiling r
• 50% PG (propylene glycol) solution
ae
2 point, thermal stability and price
L V 64 L V E t
hlf = fla min ar = G / W must r be considered. • PF/water solution
Dh 2 u Dh2 2 E
W ate
PG These tic PAOcan )
r o ma tic fluids ( be divided into two • Silicone
A categories: n
22 Cooling News h a
Alip Dielectricn o
Rdielectric
l 25liquids include
and non-dielectric.
) aromatics, • Aromatic hydrocarbon
ola m XLT C-77
Cosilicate-ester,
her bon ( F
aliphatics, silicones and
y l t • PAO (polyalphaolefin)
Sfluorocarbons.
ca r Some non-dielectric
luro are ethylene glycol, methanol/
Ffluids • Fluorocarbon
water, ethanol/water, calcium chloride
solutions, potassium formate and liquid For each fluid, the authors calculated
23 Who We Are metals [1]. the Prandtl number (Pr), Reynolds
number (Re), Nusselt number (Nu) and

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 
FUTURE COOLING

Liquid Coolant Pr Re Nu
2
h (W/m K) ◄ Table 1. Comparison of Prandtl
50% PG 32.64 545 ---- ----
Number (Pr), Reynolds Number
(Re), Nusselt Number (Nu) and
PF/Water 9.68 1346 5.653 8664
Heat Transfer Coefficient (h) for
Silicone 17.43 1330 6.622 2007 Different Thermal Coolants [2].
Aromatics 12.12 1875 15.05 5329

PAO 75.73 298 ---- ----

Fluorocarbon 16.61 3208 36.09 6339

Liquid Coolant q(W) P(W) Effc=q/P Effc/Eff w

◄ Table 2. Comparison of Heat Dis-
PF/water 26.2 1.69 15.5 0.564
sipation (q), Power (P), Efficiency
(Effc) and Efficiency with Respect
Silicone 6.8 1.14 5.99 0.217 to Water (Effc/Effw) for Different
Aromatics 16.7 1.14 14.66 0.532 Fluids [2].
Fluorocarbon 19.5 2.23 8.73 0.317

heat transfer coefficient (h) inside a specific microchan- Where K, ρ, Cρ and µ are the fluid’s thermal conductivity,
nel. These are shown in Table 1. The calculations were density, specific heat and dynamic viscosity respectively.
derived from equations that were valid for (Re>1000, The Mouromtseff number shows that the higher the ther-
turbulent flow). The Reynolds numbers for the 50% PG mal conductivity, density and specific heat, and the lower
and PAO coolants were less than 1000; thus the Nu and the dynamic viscosity, the higher the rate of heat transfer.
h were not calculated. Table 1 shows that even though
PF/water has the lowest Nusselt number, it has the high- Figure 1 shows the relative heat transfer rates of
est heat transfer coefficient because of its high thermal several liquid coolants with respect to water for an
conductivity. The low thermal conductivity fluorocarbon internal laminar flow using the Mouromtseff number.
has a high heat transfer coefficient because of its high This graph confirms that PF/water provides the highest
Reynolds number (high density and low viscosity). heat transfer rate after water, and that FC-77 provides
the lowest. Figure 2 gives the same comparison between
The question is: At what cost do we need to pump the the liquids for an internal turbulent flow, and Figure 3
liquid to achieve the above values? The pumping power shows the relative performance for a developing internal
was calculated based on the pressure drop and the laminar flow. In all three conditions PF/water and
volumetric flow rates. An efficiency was defined as methanol are the best performers.
Effc = q/P, where q is the heat dissipation and P is the
pumping power required to drive the flow. Table 2 shows
the efficiency of the liquids. The ratio of the efficiency of
Register Now
these liquids with respect to water (Effw) is also shown.
The data shows that PF/water has the best overall
efficiency.

A very useful figure of merit number for single phase

forced convection can be used to determine the heat
transfer rate of a coolant. This is the Mouromtseff number Not already on the Qpedia mailing list?
(MO) [4], and is defined as:
δ
Register now at www.qats.com and get
Κ αρβ Χπ
Μο = your monthly Qpedia right in your inbox.
με

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 
 Relative Heat Tr
0.4

0.2

FUTURE COOLING
0.0
Coolant Type

1.0 Methanol/water is basically a low cost antifreeze that can

Water
Potassium Formate/Water
be used down to - 40 oC, but its toxic and flammable nature
Methanol/Water limits its use to outdoor applications. On the other hand, PF/
Ethanol/Water water is non-toxic, non-flammable and not very corrosive [1].
0.8 EG/Water
PG Water
Relative Heat Transfer Rate

1.0
Aromatic
Allphantic (PAO)
0.6 Coolanol 25R
Syltherm XLT 0.8
Flurocarbon (FC-77)

Relative Heat Transfer Rate

0.4
0.6

0.2
0.4

0.0
Coolant Type 0.2

Figure 1. Relative
WaterHeat Transfer Rates of Different Liquid
Coolants withPotassium
Respect to Formate/Water
Water for a Fully Developed, 0.0
Laminar Internal Flow [4].
Methanol/Water Coolant Type
Ethanol/Water
1.0
EG/Water Water
PG Water Figure 3. Relative Heat Transfer Rate of Liquid Coolants with
Potassium Formate/Water
Aromatic Respect to Water for a Developing Laminar Internal Flow [4].
Methanol/Water
0.8 Allphantic (PAO) Ethanol/Water
In recent years researchers have found interest in
Coolanol 25R EG/Water
Relative Heat Transfer Rate

another type of liquid, nanofluids. These are composed

Syltherm XLT PG Water
of nanometer-size particles with high thermal conductivity,
Flurocarbon (FC-77) Aromatic
0.6 such as copper, gold, silver or aluminum, in a base fluid
Aliphantic (PAO)
such as water, ethylene glycol or propylene glycol [5, 6].
Coolanol 25R
Eastman etSyltherm
al. showedXLTa 40% increase in thermal con-
0.4 ductivity with 0.3 volume(FC-77)
Flurocarbon percent of copper nanoparticles
in ethylene glycol [7]. Devdatta et al. showed the relative
performance of several nanofluids with respect to the
ambient [8]. They used the figure of merit (FOM) for their
0.2 analysis. Because nanofluids are mixtures and not pure
δ
Κ αρβthermal
substances, their Χπ properties can easily be found
Μο = ε
μ
by knowing the percentage of the solid particles in the
0.0 solution. Figure 1 shows an example of how the effective
Coolant Type thermal conductivity of a nanofluid can be found [7]:

Water
Figure 2. Relative Heat Transfer Rate of Liquid Coolants with é K p + (n - 1)K f - (n - 1)f (K f - K p ) ù
Potassium
Respect to Water Formate/H
for a Turbulent 2O Flow [4].
Internal K nf = K f êê (1) ú
ú
Methanol/Water ëê K p + (n - 1)K f + f (K f - K p ) ûú
Ethanol/H2O
EG/Water
PG Water
© AdvancedAromatic é | T: 781.769.2800æwww.qats.com
0.218 ù
Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062
ê d ö
0.754 ç p ÷ ú r nf
0.333 0.4 0.6
CPnf K nf Page 
Flurocarbon (FC-77) Mo = ê1+ (11.285f ç ÷ ú
FUTURE COOLING
δ
Κ αρβ Χπ
Μο =
με
where Knf is the nanofluid thermal conductivity, Kρ is Unlike pure water, EG/water and PG/water enhance the
the particle thermal conductivity, Kf is the fluid thermal heat transfer rate continuously with increasing particle
conductivity, Ф is the particle volume fraction and n is an volume percentages. The advantage of EG/water over
empirical scaling factor that depends on the geometry the PG/water is also obvious.
é K + (n - 1)K f - (n - 1)f (K f - K p ) ù
of the = K f êê p(for example
K nfparticles n=3 for spherical ú particles).
ú
αKβ +δ( n - 1)K + f (K - K )
Devdatta ê Κ ρ pΧπ ûú Another type of coolant is liquid metals. Due to their high
Μοet=al. used
ë the Nusselt number correlations of
f f p

μ ε
Qiang [5], and derived the MO number for the laminar thermal conductivity, there is no need for fine-pitched
and turbulent flow for a nanofluid as: microchannels that can generate very high pressure
drops. In one study, Nanocooler, Inc. used a proprietary
é æd ö
0.218 ù 0.333 0.4 0.6 non-toxic eutectic gallium-indium (GaIn) alloy with a
ú r nf CPnf K nf
Mo = êê1+ é(11.285f 0.754 çç p ÷÷÷ ú ù melting point below -10 oC [9]. The advantage of using
K + (n - 1)Kçèf a-nf(ø÷n - 1)ú f (K f - mnfK p ) ú
- 0.07
K nf = êë K f êê p û ú liquid metals is twofold. Their thermophysical properties
ëê K p + ( n - 1) K f + f ( K f - K p) ûú
are excellent and they can be pumped using magneto-
for laminar
Re £ flow;
2300Re < 2300 (2) fluiddynamic (MFD) pumps with no moving parts. The
thermophysical properties of this liquid metal are given in
Table 3.
é æ d 0.001
ö
0.218 ù 0.333 0.4 0.6
r nf CPnf0.4 K nf0.6
Mo =éê êê1+ (11.2850.6886 æ dçç p ö÷p ÷÷÷ ùú rúúnf 0.9238
f 0.754
ç CPnf-0.07K nf
Mo = ê1ê+ (7.63 f ç
ç è a nf÷÷ ø÷ ú ú m0.5238
êë ë çè a nf ø÷ ú û mnf
nf
û
for turbulent flow; Re > 2300 (3)
Re £ 2300 Property Liquid Metal Water
2300 < Re £ 25,000
Density (kg/m3 ) 6363 998
The subscript nf means nanofluid. Figure 4 shows the
é 0.001 ù
relative performance of wateræwith d ö copper r
0.9238
oxide 0.4
CPnf (CuO)K nf
0.6
Heat Capacity (J/kgK) 365 4181
Mo = êê1+ (7.63 f 0.6886 çç p ÷÷÷ úú nf
nanoparticles for çè a nf ø÷ percentage.
êë a different mixture úû mnf
0.5238
The
Kinematic Viscosity (m 2 /s ) 3.5 x 10- 7 10 -6
figure shows that with laminar flow, increasing the per-
centage of nanoparticles will continuously enhance the Electrical Conductivity (s/m) 3.3 x 10 6 10 -12
nanofluid’s heat transfer. But with turbulent flow the en-
2300 < Re £ 25,000
hancement is only about 5% within the mixture, and after Thermal Conductivity 39 0.61
this the performance degrades. This is because of the (W/mK)
increased viscosity of the nanofluid due to the increased
particle percentage. Equation 3 shows that the viscosity Prandtl Number 0.02 6.62
is raised to a positive number. Thus, the higher the vis-
cosity, the higher the denominator and the lower the MO
number. On the other hand, Equation 2 for laminar flow
shows a negative number at the exponent of the viscosity Table 3. Thermophysical Properties of Liquid Metal in a Nano-
term. cooler, Inc. Liquid Metal Cooling Loop [9].
Figure 5 shows the relative heat transfer performance Ghoshal et al. have shown an increase in the heat trans-
with laminar flow of 60:40 EG/water and 60:40 PG/water fer coefficient inside a channel with constant wall heat
with copper oxide (CuO) nanoparticles in different mixture flux as compared to water [9]. This increase is a factor
percentages. As in Figure 4, the heat transfer rate con- of 60 for the laminar flow, and a factor of about 10 for
tinuously improves by increasing the nanoparticle volume low Reynolds number turbulent flows. They were able
percentage. Figure 5 shows that the EG/water has a to measure a thermal resistance of 0.07 oC-cm2/W for a
slight advantage over the PG/water. copper source exchanger with fin height of 1.5 mm, fin
spacing of 0.5 mm and a base thickness of 0.5 mm.
Figure 6 shows the relative performance with turbulent The pressure rise generated was about 9 Pascals.Figure
flow of 60:40 EG/water and 60:40 PG/water with copper 7 shows a typical liquid metal cooled system by these
oxide (CuO) nanoparticles in different mixture percentages.

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 
 Future Cooling

8 researchers that has been used for a laptop.

Relative Heat Transfer Rate

7
6
5
4
3
2
1
0
Water 5 8 10 13 15
% of CuO Particles in Water
Laminar Turbulent

Figure 4. Relative Heat Transfer Rate of Water with Different

Percentage of Copper Oxide (CuO) for Laminar and Turbulent
Flow [8].
Relative Heat Transfer Rate

2.5
Figure 7. Typical Configuration of a Liquid Metal Cooling Loop [9].
2

1.5 The arena of liquid cooling is expanding fast and the need for
1
reliable coolants is paramount. While this area has seen innova-
tive solutions in the last few years, the marketplace has been
0.5
slow to accept them as viable techniques due to concerns with
0
their relative cost and reliability. But, we expect more advances
60:40 EG/Water + 60:40 PG/Water +
CuO Nanoparticles CuO Nanoparticles in this technology will soon address these issues.
References:
Water 0% 1% 2% 3% 4% 5% 6%
1. Mohapatra, S., An Overview of Liquid Coolants for Electronics Cooling, Electronics Cooling
Magazine, May 2006.
Figure 5. Relative Heat Transfer Rate of EG/Water and
PG/Water with Different Percentage of Copper Oxide (CUO) For 2. Mohapatra, S., and Loitkis, D., Advances in Liquid Coolant Technologies for Electronics Cooling,
Laminar Flow [8]. IEEE Semi-Therm Symposium, March 2005.
3. Bergles, A., Bar-Cohen, A., Krause, A., Advances in Thermal Modeling of Electronic Components
and Systems, Vol.2, ASME Press, New York, NY, 1990.
Relative Heat Transfer Rate

1.2
4. Simons, R., Comparing Heat Transfer Rates of Liquid Coolants Using the Mouromtseff Number,
1 Electronics Cooling Magazine, May 2006.
0.8
5. Qiang, L. and Yimin, X., Convective Heat Transfer and Flow Characteristics of Cu-water Nano-
0.6
fluid, Science in China (series E), 45: 5, 408, 2002.
0.4

0.2
6. Yimin, X., and Qiang, L. Investigation on Convective Heat Transfer and Flow Features of
Nanofluids, ASME Journal of Heat Transfer, 125, 151, 2003.
0
60:40 EG/Water + 60:40 PG/Water +
CuO Nanoparticles CuO Nanoparticles
7. Eastman, J., Choi, S., Li, S., Yu. W., Thomson, L., Anomalously Increased Effective Thermal
Conductivities of Ethylene Glycol-Based Nanofluids Containing Copper Nanoparticles, Applied
Physics Letters, 78:6, 718, 2001.
Water 0% 1% 2% 3% 4% 5% 6%
8. Devdatta, P, Namburu, P., Das, K., Comparison of Heat Transfer Rates of Different Nanofluids
on the Basis of the Mouromtseff Number, Electronics Cooling Magazine, August 2007.
Figure 6. Relative Heat Transfer Rate of EG/Water and 9. Ghosla, U., Grimm, D., Ibrani, S., Johnston, C., Miner, A., High-Performance Liquid Metal Cooling
PG/Water with Different Percentage of Copper Oxide (CuO) for Loops, IEEE Semi-Therm Symposium, 2005.
Turbulent Flow [8].

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 
Page 
THERMAL MINUTES

Human and Technical Factors in

Upfront CFD Success

Figure 1. Upfront CFD Process Improvements Over Traditional CFD Methods. Courtesy of Blue Ridge Numerics, Inc .

The mid-90s brought a sweeping change in the CFD Focus on Human Factors, Then Master the Technical Details
world. Traditional simulation tools built for Ph.D.-level After obtaining an upfront CFD tool, many companies
analysts remained popular, but a new category of “up- expect it to provide immediate results. However, without
front” engineering products gained popularity among a well-conceived implementation plan, even the easiest
multi-tasking design engineers. Today, upfront CFD has of tools will gather dust on a shelf. The biggest factors in
become the simulation workhorse, enabling design engi- determining success are not with technology or software
neers to attack electronics cooling challenges on-the-fly. features. The human factors are key.

When used early and often, upfront CFD can fast be- Four Imperatives to Successfully Roll Out Upfront CFD
come a standard part of the engineering process. Non- 1. Provide adequate training
specialized engineers can quickly down-select through 2. Beware of ‘non-specialist’ specialists
3. Build upfront CFD into your design process
hundreds of design options. As the best directions
4. Perform what-if scenarios
emerge, engineers can add detail and move toward a
more rigorous, quantitative simulation phase akin to
traditional CFD. However, the real business benefits of
1- Provide Adequate Training
It seems obvious, but this is a mistake that many compa-
upfront CFD are achieved in the early design phase –
nies make when rolling out upfront CFD. These tools are
a fact that is often lost on new upfront CFD adopters.
heavily sold on ease-of-use and buyers often equate that
with “no training required.” Imagine if the same thought
process had been applied to a 3D CAD tool. It is

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 
THERMAL MINUTES

important to educate every engineer who may need to

use these tools. And, because upfront CFD is used spo-
radically, it can be extremely productive to provide review
training every 6 to 12 months.

2- Beware the ‘Non-Specialist’ Specialist

Upfront technologies provide the greatest benefits when
implemented throughout the engineering team. Besides
avoiding the obvious risk of expertise walking out the
door, a thorough implementation will enhance an entire
group’s efficiency and innovation.

Many companies will identify one engineer to shoulder

all upfront CFD responsibilities. Others train their entire
team and wait for individuals to show a strong interest.
The result is usually the same: a single member of the
design team is targeted as the group specialist. Unfortu-
nately, this can create the same timing bottlenecks inher-
ent with traditional Ph.D. analysts and severely limit the
impact of an upfront CFD solution.

3- Build Upfront CFD Into Your Design Process

Neglecting to schedule time for upfront CFD activities
in your official process is sure to maintain “business
as usual.” If you typically manage projects using Gantt
charts and gate review systems, be sure to build upfront
CFD milestones into your plans.

Figure 2. Typical Gant Chart [1].

4- Perform What-If Scenarios
Upfront CFD tools are best used to evaluate many design
options at the conceptual stages. When integrated with
parametric CAD tools, upfront CFD allows engineers to
perform numerous hypothetical studies with very little
project definition. Hundreds of potential directions can be
compared to quickly choose the right places to focus the

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 
THERMAL MINUTES

engineering team’s efforts. • If the fluid is a liquid, do you expect any phase change,
such as solidification or cavitation?
That main point of upfront CFD is often missed by its • Do you expect the flow to be laminar or turbulent?
new users. A team waits to employ the upfront design • What moves the fluid: a fan, buoyancy, known-pressure
tool until the late design stages, after most decisions supply line or some combination?
have already been made. The tool is then used for design • Where does heat enter the system?
verification instead of design direction, and provides • Where does it leave?
little impact on the overall project, timeline or innovation. • How does it leave? Via conduction, convection and/or
This is the single biggest issue in successfully rolling out radiation?
upfront CFD technology. Upfront CFD should be used to • Does the system operate in direct sunlight?
do more now and less later on. • Are internal body-to-body radiation effects significant?

Best Practices and Technical Strategies Get a rough sense of what the reasonable results might
for Upfront CFD Success be. Use related test data, experience and hand calcula-
Here are some tips to get the most from your simulation. tions to gauge realistic result ranges.

Start With a Sketch Check Your Assumptions

Before doing anything on the computer (in CAD or in the Every calculation is founded on certain underlying
simulation tool), sketch a conceptual model with the most assumptions and upfront CFD is no different. It is
important geometry and indicate the fluid and heat flow important to know if your assumptions are conservative
paths. or non-conservative (e.g. speculative) in relation to such
critical results as the peak die temperature, total flow
through the system and pressure drop in the system.

Typical Flow Assumptions:

• incompressible or compressible
• laminar or turbulent
• wall conditions (slip or no-slip)

Typical Thermal Assumptions

• perfect thermal contact
• radiation or no radiation
• adiabatic external walls or external leakage

In a heating situation, for example, assuming perfect ther-

mal contact is a non-conservative approach. Any contact
resistance between two parts (which inevitability occurs
in real life) will cause actual temperatures to be higher
than predicted by simulation.

The “no body-to-body radiation” assumption, on the

other hand, is typically conservative. Radiation between
Figure 3. Sample Heat and Air Flow Path Sketch for Power components and the exterior housing is an extra heat
Converter. Courtesy of Blue Ridge Numerics, Inc. flow path along with conduction and convection. Ignoring
this path will typically cause higher temperatures in your
Understand the Critical Physical Effects simulation results. In most forced flow cases where the
• If the fluid is a gas, are compressibility effects signifi- air velocity is more than 1.25 m/s, radiation effects are
cant? minimal compared to conduction and convection, so this

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 
THERMAL MINUTES

can be a time-saving assumption. If radiation effects are If you keep your first models rough and simple, upfront
noticeable (as is usually the case in natural convection or CFD will move much faster. You can also quickly test the
low velocity flow), your simulation has predicted higher relative importance of factors such as variance in mate-
temperatures than what you will see in the lab. This con- rial properties and mesh sensitivity. Later, if you decide
servative assumption thus builds a minor safety margin to run a model with production level details, you will have
into your judgment. a good sense of the potential impact of these factors on
the final results. However, you should always avoid very
“Adiabatic external walls” is another typical conservative fine details that may have no significant influence on the
assumption. In this case, you force all heat to leave the results.
system with the exiting airflow. In real life, heat will also
escape by conduction to the housing, to be carried away General Approach for Internal Forced Air Cooling
by external room air. Ignoring this heat leakage path re- • Keep all components in your model that are significant
sults in higher simulated component temperatures, again to the air flow path, form part of the heat flow path or are
offering a safety margin in the analysis. significant heat generators. Consider both the total heat
dissipated by the components and their heat per unit
KISF (Keep It Simple First) volume power density.
Make your first simulation on a geometrically simple, but
representative model of the system. Start with a relatively • Apply 0 Pa (gauge) pressure to all atmospheric inlets
coarse mesh, but try to include all significant physical ef- and outlets. Do not, however, apply pressure to any
fects. This is mostly important with new designs and early inlets or outlets where you are also specifying a fan,
on in your experience with the analysis tool. One reason velocity or flow rate. Doing so over-specifies the solution
is purely pragmatic: when the result appears inconsistent, and typically causes diverse CFD results among users.
it is important to see why as quickly and clearly as possible.
• Be sure to specify ambient room temperature or
In addition, this simple representative model can be used incoming line temperatures at all inlets.
to understand the general performance of the system,
particularly in areas that require special attention. Where • Assign volumetric heat sources to the significant heat
is the highest temperature? Where is the highest pres- generating components.
sure gradient or choke-point in the flow? If you are using
a fan, roughly where is the fan operating on its perfor- • If heat leakage through housing or case walls is signifi-
mance curve? Later, you can focus more simulation cant, assign a heat transfer coefficient to the exterior
attention on these critical areas. faces of the enclosure. This approach saves analysis
time by avoiding the need to computationally model the
surrounding air. For natural convection, 5 - 10 W/m2K
is a reasonable choice. If the external air is forced
flow, use Nusselt correlations to get an estimate. Then
analyze the simple system with both double and half
of that film coefficient to estimate the best and worst
case effects.
Figure 4. Electronics Assembly Simplification for
Simulation. Courtesy Blue Ridge Numerics, Inc. • Assign appropriate fan curves. Most companies try to
use the same fan part number across multiple designs
Never include full production details, e.g., screws, clips, to save material and development costs. Fans produce
springs, etc. in the model. Instead, bring simulation into more or less air flow depending on the geometry of the
the design process well before that point. Such details systems to which they’re attached. It is worth the time to
typically do not affect the thermal results and only serve verify fan curves provided by the manufacturer in your
to complicate the model and lengthen the convergence time. own lab. You may even need to come up with your own
curve that better matches reality.

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 10
 THERMAL MINUTES

Normal Operating Range

A given fan can deliver only

STATIC PRESSURE

one flow into one pressure

in a given system

Fan
Performance
Curve
Ruben Bons
Sr. Application Engineer, Blue Ridge Numerics, Inc.
AIR FLOW is a graduate of Calvin College (Grand Rapids, MI) and the
University of Tennessee Space Institute (Tullahoma, TN). He
began working with Blue Ridge Numerics simulation tools in
1996 while at Structural Research and Analysis Corporation,
and he joined the BRNI team in 2003 to train and support
Figure 5. Typical Fan Curve Plot [2]. companies implementing upfront CFD. His simulation
experience spans many industries, including aerospace,
Upfront CFD can radically reduce the time and cost of de- defense, consumer products, medical devices, and semicon-
veloping electronic systems when implemented rationally ductor, with particular focus on companies addressing electron-
as a core competency throughout the design cycle. With ics cooling challenges.
a little training and a pragmatic approach, upfront CFD
REFERENCE:
offers major gains in productivity, innovation and first
time quality. 1. http://lifeupfront.com/2007/09/28/inject-upfront-
cae-into-the-engineering-process/)
2. http://www.comairrotron.com/airflow_note.shtml

Jeff Waters
Account Executive, Blue Ridge Numerics, Inc.
is a graduate of Rose-Hulman Institute of Technology.
He began his engineering career at General Motors,
where he introduced a variety of specialty and early CAE
technologies for simulation-based stress, strain, thermal
and flow R&D. In 1999, Jeff joined Blue Ridge Numer-
ics, Inc. to help manufacturing and product development
companies efficiently implement upfront CFD. Read more
Advanced Thermal Solutions, Inc.
about getting the most from upfront CAE on Jeff’s blog: offers custom design solutions for every
www.lifeupfront.com and in his recent whitepaper at www. application. Contact us at 781.769.2800
cfdesign.com/qpedia. for details.

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 11
 Thermal Minutes

Using Flow Visualization for

PCB Thermal Design and Optimization

FLOW DIRECTION ► ◄ F igure 1. Top View of Flow Around Equal Sized

Components, Depicting Reversed Flow in the Wake
of the Component and Rotational Flow on the Outer
Periphery and in Between Components [2].

FLOW DIRECTION ►

▲ F igure 2. Flow Around Two Components of Varying

Heights and the Same Planar Area. The Flow Between
the Components Exhibits Oscillation and Reversal.
the Wake Shows a Near Stagnant Rotating Flow.

PCBs support a multitude of components with varied The problem stems from the intricate topology of the
geometries, electrical functions, power dissipation and PCB. Highly complex flows are observed on PCBs due
thermal performance needs. For a PCB to work prop- to their three dimensional protrusions, i.e., components.
erly, a component’s thermal requirements must be met A typical PCB sees every imaginable flow structure
locally or at the system level. Regardless of the type of [1]. These include laminar, turbulent, separated flow,
housing that surrounds a PCB, its cooling system must reversed flow, pulsating, locally transient and others.
be designed so that diverse components are electrically Figures 1 and 2 show examples of these complex PCB
functional and run at temperatures that help them reach flow structures [2].
their expected life spans.
These flow visualizations were made in a laboratory wa-
Much effort is needed to meet a component’s thermal re- ter tunnel at Advanced Thermal Solutions, Inc. Because
quirements, whether by enhanced fluid flow (liquid or air) the flows are very complex, water flow visualization using
or by adding a cooling solution, e.g., heat sink, onto the colored dyes captures the flow details more accurately
component. Except for a conduction cooled PCB, where than air flow visualization. These flow structures are
a coldplate extracts heat from the board, electronics identical whether in air or water as long as the Reynolds
are typically in contact with some sort of cooling fluid. In number is the same for both conditions.
many cases, the PCB is in direct contact with the coolant.
This creates a very complex problem along with a unique The complexity and structure of fluid flow in PCB topolo-
opportunity. gies creates a unique thermal problem. Despite many

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 13
Thermal Analysis

attempts to improve flow at the inlet of a PCB, the board In Figure 3a, visualization lets the engineer see the fluid
layout diverts the flow to the path of least resistance. This flow pattern on the PCB and identify those areas and
generates regions of stagnation and creates high pres- components where flow exposure is at a minimum. By
sure areas. The components are deprived of cooling air reconfiguring the board (note the location of components
or liquid. Figures 3a and 3b show one such example. 6, 7 and 9 which is rotated 90o), as shown in Figure 3b,
the pressure distribution on the PCB was balanced and
subsequently the flow distribution improved significantly.

Another significant advantage of flow visualization is the

measurement of local velocity. Simple videography (30
FPS) on a scaled distance can provide the approach air
velocity to each component on the PCB. This gives the
designer local knowledge of the flow magnitude and type
for help when choosing the cooling system.

Considering the complexity of fluid flow in a PCB, predict-

ing flow patterns is not reliably achieved with numerical
modeling. In some instances, the result has been signifi-
cant prediction errors in component junction temperature
[3, 4]. Flow visualization has the potential to yield more
insight into a fluid flow or convection cooling problem
than any other single method. Many misconceptions can
usually be cleared up by flow visualization. However, it
3a►
is important to use the technique most suited to a given
problem.

The flow of a fluid may be characterized by one of the fol-

lowing four terms, which must be carefully defined:

Streamline: a curve that is tangent everywhere to the

instantaneous velocity vectors.

Streakline: the locus of all fluid particles that have

passed through a prescribed fixed point during a specific
interval of time. An example of a streakline is a line traced
by the continuous injection at a certain point of dye,
smoke or bubbles.

Pathline: the curve traversed by a particular fluid particle

during a specified time interval, such as that produced by
time-exposure photography of a tracer particle moving in
a flow.

Timeline: a set of fluid particles that form a line at a given

3b► instant, such as would be obtained in a short-time expo-
sure photograph of a line of hydrogen bubbles released
from a wire to trace the velocity profile in a fluid.
Figure 3a. Original Layout, Components.
Figure 3b. Optimized Layout.
Note: Components 6, 7 and 9 are Deprived of the Flow.

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 14
 Thermal Analysis

While a streamline is mathematically the most tractable, Technique Fluid Remarks

the others are easier to generate experimentally. Stream- Smoke entrainment Inexpensive.
lines and timelines are instantaneous entities, while Smoke-wire Air Easy to implement.
streaklines and pathlines are generated over a period of Smoke-tube Air Stabilization of smoke filaments and
time [5]. photography need care.
Results harder to interpret than in dye
Azar, et al., [5] report that the literature abounds in flow
entrainment in liquids due to the greater
visualization methods and their use in a wide range of
extent of diffusion of smoke in air.
situations. Flow visualization methods can be grouped
Dye Entrainment Liquid Inexpensive.
into two categories: those suited to investigate the com-
plexity of streamlines just above a surface, and those Neutral dye (e.g., ink) Relatively easy to implement (injection

suited to characterizing surface heat transfer properties. easier than entrainment, but the latter

When combined, these analyses can provide a detailed more accurate since spurious injection

description of the flow phenomena and their effects on velocities avoided).

heat transfer processes. Good contrast is easy to obtain.

Neutral buoyancy of dye must be
This article is confined to the methods the authors ensured.
consider most applicable for visualizing air flows in Closed-loop systems contaminate in a
electronic systems. These techniques appear in Table short period.
1 and can be reviewed more thoroughly and in greater Interpretation and photography are
detail by reading Garimella [6]. The principle for each straightforward.
technique is described in addition to general implementation Tracers for core flows Relatively inexpensive.
details, with examples of visualizations also given. Dye injection Liquid Recommended wherever practicable,
Unlike shadowgraph and Schlieren, HI (holographic) Tufts Air, especially in liquid-flow facilities.
is especially easier to implement, with less expensive Liquid Excellent contrast can be obtained with a
optical components, and affords high resolution as well good source of illumination.
as transient mapping. PIV (particle image velocimetry) is Wall traces Relatively inexpensive but can be labor-
perhaps the most expensive of the techniques discussed, Oil film/powder Air intensive.
but yields accurate measurements over entire planes in Oil film/ink dot Air, Good for specialized applications where
the flow field. Tufts probes Liquid wall shear stresses and near-wall velocity
Air, fields need to be mapped, e.g. three-
Air Flow Visualization Methods
Sublimation/Evaporation Liquid dimensional corner regions which are
As reported by Azar, et al. [5], air flows can be qualita- Air less accessible to other methods.
tively visualized using smoke entrainment techniques or Optical methods Air,
electronics applications, unlike shadow
quantitatively measured using optical methods. Shadowgraph Liquid
graph and Schlieren; HI (holographic)
Schlieren
Smoke entrainment is the most common visualization especially easier to implement, with less
Interferometry
technique for laminar air flows. But it has somewhat expensive optical components, and
Holographic
limited use in turbulent flows due to its rapid diffusion by affords high resolution as well as transient

turbulent mixing. Smoke can be produced from many mapping. PIV (particle image velocimetry)

sources, but essentially it is made by either smoke-tube is perhaps the most expensive of the

or smoke-wire. techniques discussed, but yields accurate

measurements over entire planes in the
In the smoke-tube method, vaporized oil is used to form a flow field.
visible whitish cloud of small particles as the hot oil vapor
condenses. Consideration must be given to the vortices
shed by the smoke probe itself, since the most visible
small-scale features often arise from the probe’s own
Table 1. Flow Visualization Techniques for Electronic Systems
wake. This effect is important for probe diameter-based
[6].

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 15
Thermal Analysis

Reynolds numbers exceeding about 15. Because it is Type 1964 fog juice, with a boiling temperature of 276ºC,
impractical to reduce the probe diameter beyond a point has also been recommended in the literature for produc-
and still get a reasonable amount of smoke flow, the ing dense smoke that is non-toxic and non-corrosive. An
smoke can be injected upstream of a convergence sec- example of flow visualization obtained over a multi-com-
tion to eliminate wake effects. A key disadvantage to the ponent PCB using the smoke-wire method is shown in
smoke-tube method is that the smoke is produced hot Figure 4.
and rises due to its own buoyancy, thus it doesn’t follow
SO 16-e
the local flow faithfully. To reduce buoyance effects, the
smoke can be cooled in a long length of tubing from the SO 16-m
point of generation before its introduction into the flow.

In the smoke-wire method, smoke is generated as a

sheet by coating a thin wire with oil, stretching it across
the flow, and heating it with a pulse of current. Almost
any wire and power supply can be used in this technique.
The oil should be chosen carefully to have a broad boil-
ing plateau, rather than a single temperature, in order to
generate good smoke. Model train oil is suitable for this
method. TSOP
Wire diameters of about 0.1 mm work well. Stripped AIR FLOW
chromel or alumel thermocouple wires are readily avail-
able choices. Once the wire is coated and allowed to sit,
the oil draws up into regularly spaced droplets on the wire
Figure 4. Experimentally Visualized Flow Fields On a Multi-
in less than a minute. Upon being pulsed with current, the
Component PCB in a 2 m/s Air Flow, Using a Smoke-Wire
droplets produce discrete parallel streaks of white smoke.
Method [5].

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 16
Thermal Analysis

While the smoke-tube method uses large quantities of • Examine the PCB layout at the design stage for
smoke and produces a limited number of streaks that expected thermal performance, e.g., determine flow
are rather thick (10 mm or more in diameter), the stagnation areas.
smoke-wire method can produce very thin smoke
streaks (about 1 mm). • Make component layout recommendations that
provide a thermally optimum board.
Liquid Flow Visualization
•Identify approach air velocities necessary for
Flow visualization is usually easier to perform in water component thermal management and the choice
than air and yields results of better quality. As with smoke of cooling system.
visualization, dye entrainment is successful mostly in
laminar flow. The enhanced mixing in turbulent flow The end result of board level flow visualization is
causes the smoke streaks to diffuse too rapidly to be of PCBs that are thermally optimized and require no re-
value as tracers. Compared to smoke visualization in spin because of thermal constraints. If the board is
air, dye entrainment in liquids is helped by the fact that thermally laid out, heat sinks and other cooling
the mixing between most dyes and water is less intense solutions are often not required.
than between smoke and air [5]. As a result, water-flow
tunnels are frequently used to study air flows by testing
scaled models at lower velocity. This often provides a References:
better description of the flow.
1. Azar, K., Enhanced Cooling of Electronic Components by Flow
A liquid flow model is scaled with a different working fluid Oscillation, Journal of Thermophysics and Heat Transfer, Vol.
6, No. 4, 1992.
than air using dimensional analysis. For a treatment of
the principals of dimensional analysis and similitude that
should be used in applying flow visualization in model 2. Azar, K., and Russell, E., Effect of Component Layout and
experiments, reference can be made to any standard Geometry on the Flow Distribution in Electronic Circuit Packs,
textbook on fluid mechanics. For the flow conditions
Proceedings of the Sixth IEEE Semiconductor Thermal and
Temperature Measurement Symposium (SEMI-THERM VI),
around a model to be completely similar to those of the Phoenix, Arizona, 1990.
prototype, all relevant dimensionless parameters must
have the same corresponding values: the model and
prototype are then said to possess geometric, kinematic, 3. Rodgers, P., An Experimental Assessment of Numerical Predic-
tive Accuracy for Electronic Component Heat Transfer, Ph.D.
dynamic and thermal similarity. Thesis, University of Limerick, Limerick, Ireland, 2000.
To visualize the flow, water-soluble dyes such as food
coloring, potassium permanganate, methylene blue, ink 4. Eveloy, V., Lohan, J., and Rodgers, P., A Benchmark Study of
and fluorescent ink may be injected using hypodermic Computational Fluid Dynamics Predictive Accuracy for Compo-
needles or entrained from holes or slots in the walls of a nent-Printed Circuit Board Heat Transfer, IEEE Transactions on
test section. It is important that the velocity and density
Components, Packaging and Manufacturing Technology (CPMT),
Vol. 23, Number 3, 2000.
of the injected dye equal those of the surrounding fluid to
maintain a stable dye filament and reduce disturbance of
the surrounding flow. 5. Azar, K., and Rodgers, P., Visualization Of Air Flows In Elec-
tronics Systems, Electronics Cooling Magazine, May 2001.
In summary, fluid flow visualization is a powerful and
unique technique for quickly identifying the flow distribu- 6. Garimella, S., Flow Visualization Methods and Their Applica-
tion and approach air velocity to thermally challenging tion in Electronic Systems, Thermal Measurement in Electronic
components in a complex PCB structure. By using this Cooling, Azar, K., Ed., CRC Press LLC, New York, 1997.
technique, one can attain the following:

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 17
 My name is Mitch.
I am an Applications Engineer for a
networking equipment provider.

I still call my mother every Sunday.

I( AM HOT)
AND I’M NOT AFRAID TO ADMIT IT

They say the first step to solving your problem is admitting you
have one. The second is calling us. Now that ADVANCED THERMAL
SOLUTIONS, INC. has teamed up with DIGI-KEY, your electronics
cooling solutions are only a click away.

Through Digi-Key, the global electronics components distributor,

ATS now offers over 380 heat sinks for cooling BGAs and other
hot-running semiconductor devices. The heat sinks are available
for immediate delivery when purchased through Digi-Key’s Web
site and printed catalogs.

Now you can easily get individual ATS heat sinks for prototypes
and testing, or larger volumes for production requirements. And,
with Digi-Key’s online tools, you can check real-time inventory and
order parts 24 hours a day, every day of the year.

Be Cool. Get your catalog today.

For an ATS heat sink catalog call 781.769.2800 or
download it at qats.com. To place an order, check availability
or view orders, visit digikey.com. Or call 1-800-DIGI-KEY.

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 18
 THERMAL FUNDAMENTALS

Dh = do-di

di
do
b
a
ab
Dh = d d Dh = 2
a+b

Hydraulic Diameter:
Definition and Role in Electronics Cooling

When designing an electronic enclosure, internal pres- (1)

4´ A
sure losses must be estimated along the paths of the Dh =
air flow. Friction losses are generated by flow of fluid Pwetted
over the enclosure walls, circuit cards, components, etc. Given the complexities of viscous sublayers, turbulence,
Local losses are the result of fluid dynamic momentum shear stress, surface roughness, etc., the use of the
changes due to sudden expansion or contraction of the hydraulic diameter method introduces some limitations.
enclosure. For one, the use of the hydraulic diameter is only valid if
Both types of losses must be minimized by optimizing the the ratio of the duct’s height to width is less than about 3
design parameters. One of the most important of these is or 4 [1].
hydraulic diameter. Developed for non-circular ducts, the Using the hydraulic diameter method to estimate head
hydraulic diameter is used as an equivalent diameter to loss in laminar flow can also introduce large errors. This
circular ducts when calculating estimated head losses. is due to friction from the action of viscosity throughout
Friction Losses the whole body of the flow. It is independent of surface
By definition, the hydraulic diameter of a non-circular en- roughness and is not associated with the region close
closure is four times the cross sectional area divided by to the boundary walls [2].
the wetted perimeter. This is shown in Equation 1 below. Whether a flow is laminar or turbulent, the hydraulic
The wetted perimeter is the edge of the cross sectional diameter and fluid properties are key parameters in
area that is in direct contact with the flow medium. determining the Reynolds number, as seen from Equa-
tion 2:

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 19
THERMAL FUNDAMENTALS

(4)

(2)

In equation 4, hlf and L are friction head loss and length

of the enclosure respectively. Conversely, for turbulent
In equation 2, ρ, V and μ are density, average velocity flow there is no analytical equation for directly estimating
and viscosity of the fluid respectively. In laminar flow, fluid the head loss. Instead, the value of the friction factor, f
head loss is dependent only on the Reynolds number. is obtained from the Moody diagram using the Reynolds
The higher the Reynolds number, the lower the head number and the ratio of the pipe’s inside surface rough-
loss. This is seen in Equation 3, the laminar friction factor: e
ness and diameter, D. After estimating the friction factor
the friction head loss in turbulent flow can be calculated

64 (3)
fla min ar = using the following equation:
Re
L V2
Substituting the above into the theoretical laminar head hlf = f
loss equation reveals that pressure drop in laminar flow is Dh 2
(5)
inversely related to the square of the hydraulic diameter. As seen from Equation 5, as with laminar flow situations,
This is shown in Equation 4. As discussed, in the case the hydraulic diameter is inversely proportional to the fric-
of laminar flow and for low Reynolds numbers using this tion head loss.
equation may lead to significant errors. This is because
in such flows friction is not associated with the surface An alternative to the Moody diagram for finding the
roughness of the enclosures and the wall boundary layer. friction factor is to use the empirical correlations of
Blasius and Colebrook [3].

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 20
 THERMAL FUNDAMENTALS

Local Losses dependence of local losses on the hydraulic diameter.

In electronic enclosures, local losses are as important as It is not practical for estimating either of the equivalent
friction losses. Local losses occur whenever the dynamic parameters without knowing the individual loss coefficient
momentum of the flow varies due to an abrupt change K values of the chassis elements. Once these loss coef-
in cross sectional area. In telecommunication chassis, ficients are found, it is possible to estimate the system
these losses are seen across the inlet and outlet panels, head loss.
and over the internal mechanical structure of the chassis
including air filters, perforated plates, honeycomb, circuit Loss coefficient values, K are usually provided by the
cards, component blockade, fan tray plenum, etc. manufacturers of chassis components. When not
included, they can be found in any hydraulic resistance
Local losses across an enclosure’s elements are dis- handbook that covers such practical applications as
sipated energy in flows induced by curvature or through sudden expansion and contraction, flow over barriers
recirculation. The local losses over various elements of and orifices, etc. For specific applications where data is
the chassis can be estimated using the following equation neither provided nor available, the K value can be deter-
(6), where K represents the individual loss coefficient of mined experimentally by a simple differential pressure
the chassis elements: measurement across the device.

V2 fLeq V 2 Some typical head loss coefficients in electronic enclo-

hlm = SK =( ) sure applications are listed in Table 1. To clearly demon-
2 Dh 2 (6) strate the effect of the hydraulic diameter, these losses
are re-expressed in the third column.
In equation 6, Dh and Leq are the equivalent hydraulic
diameter and the equivalent length of the enclosure. This Adding the friction and local losses gives the expression
expression is only useful for demonstrating the inherent for the total pressure loss. In the total head loss equation

Chassis Loss Coefficient, K K in terms of Dh Comments

Elements
0.828 13.25 A = Open area
Slotted Plate
A2 P 2Dh2
L = Circuit board
length
4.2 × L 67.2 × L A = Effective free area
Circuit Board
A2 P 2Dh2 of the channel
between two circuit
boards
A1 = Small area
Fan Tray - 1 A 4 PD
0.46 × [ × (1− 1 )]2 0.46 × [ × (1− 1 h1 )]2 A2 = Large area
Sudden
Expansion A1 A2 P1Dh1 P2Dh 2

Fan Tray - A = Small area

0.321 5.14
Sudden
Contraction A2 P 2Dh2

Table 1. Resistance of Common Elements in Electronic Chassis [4].

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 21
Thermal Fundamentals

(7), the summation indicates that all discrete local head

Be Cool.
loss coefficients of the chassis elements must be evalu-
ated and taken into account.
2
V fL
(7)
hl = hlf + hlm = ( + SK )
2 Dh
Get your ATS Heat Sink
Conclusion
Proper use and calculation of the hydraulic diameter
catalog today.
plays a pivital role in estimation of the system air flow
and cooling requirements. Therefore, use of hydraulic
diameter among an enclosure’s key design criteria should
ensure minimum pressure losses, higher air flow through
the area, and more effective cooling of its electronic
components.

References:
1. Fox, R., McDonald, A., Introduction to Fluid Mechanics, Third
Edition, Wiley, 1985.
2. Street, R., Watters, G., Vennard, J., Elementary Fluid Mechanics,
Seventh Edition, Wiley, 1996.
3. Advanced Thermal Solutions, Inc., Hydraulic Resistance and its
Role in Electronic Cooling, Qpedia, September 2007.
4. Ellison, G., Thermal Computations for Electronic Equipment,
Van Nostrand Reinhold, 1984.

For an ATS heat sink catalog call 781.769.2800 or

download it at qats.com. To place an order, check

availability or view existing orders, visit digikey.com,

or call to order 1-800-DIGI-KEY.

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 22
COOLING NEWS

High Performance Cooling for EVENTS

Cavium’s OCTEONTM CN38XX/
CN36XX Multi-Core Processor Materials Research Society - Fall Meeting
ATS PART NUMBER: ATS-691-C2-R0 November 26-30, 2007
Boston, MA
Features & Benefits
ASME Micro/Nanoscale Heat Transfer
International Conference
● Designed to meet or exceed the thermal requirements January 6-9, 2008
of Cavium’s OCTEONTM CN38XX/CN36XX Multi-Core
Processor Tainan, Taiwan

● High Efficiency heat sinks ideal for low or no system Thermal Solutions
airflow environments
January 21-24, 2007
● Includes a 50 x 50 x 10 mm, 6500 RPM, 12V DC Sarasota, FL
blower fan
SMTA - Pan Pacific Microelectronics
● Comes with pre-assembled Chomerics T-766 phase
chang thermal interface material Symposium
January 22-24, 2008
● Mechanical attachment method included, however, Kauai, Hawaii
custom attachment options may be available
SEMI-Therm
March 16-20, 2008
San Jose, CA

IMAPS - International Conference and

Exhibition on Device Packaging
March 17-20, 2008
Scottsdale/Fountain Hills, AZ

IPC Printed Circuits Expo, APEX and the

Designers Summit
April 1-3, 2008
Las Vegas, NV

European Thermal - Sciences Conference

May 18-22, 2008
Eindhoven, the Netherlands

Electronic Components and Technology

Conference
May 27-30, 2008
For further technical information, please contact Advanced Thermal Lake Buena Vista, FL
Solutions, Inc. at 1-781-769-2800 or www.qats.com

Sensors Expo and Conference

June 9-11, 2008
Rosemont, IL

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 23
Who We Are

1
Volume
Issue 3
07
April 20

tronics
of elec
ement
manag
ermal
er fo r the th
wslett
The ne

ith
oling w
vel Co and
ING
COOL
FUTURE

ip -L e
Ch g
In this
issue:
owettin
Electr hannels
Cooling c DES

1 Futu
re
M ic ro V(t 1)
V(t 3)
ELECTRO

- -
-
L WALLS
- +
- - + +
LOY S
CHANNE +
ID
+ +
METALS/AL
+ +
LIQU + + -

Minutes
-
+ + - -
-

mal
-
V(t )

5 Ther ctrodes can

4

ges on ele
n of volta .
d activatio
Th Qh
V(t )
X
el [2]
. Sequence in a microchann
2
MATERIAL

can be
Y
Tc MATERIAL B
ow ett ing on fits, if it
s electr t to any locati ne
Y
MATERIAL
l be
Continuou potentia nted in
Qc
ple impleme
Figure 1. a liquid metal dro it offers
A
stage,
Analysis
h
ssfully
I
d succe
mal transport become ged an
T2

8 Ther
es has packa
T1
vic
ower de
g .
ckagin a system principle
g high-p more pa
Vin
on the
Coolin e, with the ch
ip and works ng the
challeng ints at Electrow
etting controlli are
a major es are tion by
constra
techniqu
t actua id. There
and rel
iability
. Many power of drople sion of the flu ol surface
els en
lev ed wh sired surface ten
ls
ntr and
amenta
Length system develop t the de s to co ponse
or tha anism
mal Fund
Entry
Thermal
red el ch res
Length
conside a lev be many
me switch
ing
ctric
at such cannot
11 Ther
e fast
Entry
100 Combined
(Pr = 0.7) n is ure Th in the ele
dissipatio rat n. on
n tempe ds. tensio nsumpti kes
l metho sion ma
Surface Region
junctio wer co
device
Constant Flux
entiona low po ique in
loped
Heat Fully Deve rface ten
by conv g techn
Region
Entrance etting n of su the
ed ett ,
ain ow tua tio mi sin ing
att ctr ac ing a pro electrow
4.36
20
e of ele
is the us device level ctrowett
3.66
es. In liquid-so
lid
hnique the ry ele fluidic devic
e a ve
NuD 10
One tec nels at provid micro en the nal
Surface
crochan annels r, the tension
betwe
an exter
and mi niscus
Constant
Microch ed by
5
Howeve
ing News
ture
4
Tempera 1
co n). m. surface mo difi the me
diu es is uces
14 Cool
(i.e., sili h
0.5
3
fer me very hig interfac ich red t motio
n.
at trans
0.1
0.05
in is a field, wh induces drople ters
large he
2
ch a ga owet-
electric
-1
0.01
x/D
0.005
id for su of electr ntime
1 = Gz

ving and
veral ce
ReDPr
0.001
price pa ncept angle
The co for mo contact s of se
re drop. vehicle ]. Al- velocitie
pressu pumping nnels [1,2,3 Droplet
vides a

Advanced Thermal Solutions is a leading engineering ting pro the se cha ora tory
hin lab
Page 1
lant wit is still at
the
the coo ncept
We Are
this co .com
w.qats
though 800 ww
16 Who
9.2
T: 781.76
usa |
MA 02062
rwood,
Road No

and manufacturing company supplying complete anced

Thermal Sol
utions,
inc. | 89-27
Access

ht© Adv
Copyrig

thermal and mechanical packaging solutions from

analysis and testing to final production. ATS provides
Getting your Company’s
Message out to Over 21,000
a wide range of air and liquid cooling solutions,
laboratory-quality thermal instrumentation, along with
thermal design consulting services and training. Each Engineers & Industry
article within Qpedia is meticulously researched
and written by ATS’ engineering staff and contributing
Professionals has Never
partners. For more information about Advanced Been Easier.
Thermal Solutions, Inc., please visit www.qats.com Advertise in Qpedia Today!
or call 781-769-2800. Qpedia was launched in 2007 as a technology eMagazine
focused on the thermal management of electronics. It is
ADVANCED THERMAL SOLUTIONS, INC. designed as a resource to help the engineering community
United States solve the most challenging thermal problems.
89-27 Access Road, Norwood, MA 02062
The newsletter is published monthly and distributed at no
T: 781.769.2800 | F: 781.769.9979 | www.qats.com
charge to over 21,000 engineers worldwide. Qpedia is also
available online or for download at www.qats.com/qpedia.
Europe
De Nieuwe Vaart 50 | 1401 GS Bussum | The Netherlands Qpedia’s editorial team includes ATS’ President & CEO
T: +31 (0) 3569 84715 | F: +31 (0) 3569 21294 Kaveh Azar Ph.D., and Bahman Tavassoli Ph.D., the
www.qats-europe.com company’s chief technologist. Both Azar and Tavassoli
are internationally recognized experts in the thermal
management of electronics.

For more information on how you can generate

awareness about your company or products,
contace Ray Santos at 782.769.2800.

© Advanced Thermal Solutions, inc. 2007 | 89-27 Access Road Norwood, MA 02062 | T: 781.769.2800 www.qats.com Page 24