Liquid immersion cooling technology with natural
`convection in data center
`
`Morito Matsuoka and Kazuhiro Matsuda
`Cybernedia Center, Osaka University
`/EEC Research Institute
`1-32 Machikaneyama, Toyonaka
`Osaka, 560-0043 Japan
`
`Hideo Kubo
`Electric & Mechanical Eng. Development Div.
`Fujitsu Limited
`Nakahara-ku, Kawasaki
`Kanagawa, 211-8588, Japan
`
`operation from the aspect of power consumption of IT
`equipment. The PUE in the world has been decreasing since
`2006, and reaches around 1.1 at minimum [3]. When the
`PUE=1.0 is demonstrated, the power consumption decreases
`around 9% from the PUE=1.1 level. From the aspect of
`OPEX, the 9% is significant.
`
`
`
`Abstract— We proposed a
`immersion cooling
`liquid
`technology with natural convection for high power server boards
`used in data centers with high cooling-efficiency. The cooling
`performance was evaluated by CFD simulation and actual
`experiments. As
`the refrigerants, several non-conductive,
`thermally and chemically stable fluids were applied, including
`silicone oil, soybean oil, and perfluorocarbon structured liquids.
`The CPU temperature in the refrigerant monotonically decreases
`with the Rayleigh numbers of the refrigerant. The smoother
`refrigerant is better for cooling the high power CPU. The
`changes in any CPU task and any slot-removal give limited
`cooling effects to other slots and other CPUs. These are quite
`useful features for stable operation of servers in data center. As a
`result, this proposed technology with natural convection exhibits
`promising potential for low energy and space-saving board
`cooling which demonstrates a Power Usage Efficiency (PUE)
`below 1.04.
`
`Keywords—liquid immersion, natural convection, data center,
`cloud computing, power usage effectiveness, computational fluid
`dynamics
`
`I. INTRODUCTION
`Data center is an important business infrastructure that
`supports various cloud services. In addition, it will act a
`central role for promoting several future Internet of Things
`(IoT) businesses [1].
`However, the electric power cost for the data center is
`increasing around 8 times in this 10 years. DOE reported that
`the power consumption of the data center worldwide reaches
`around 2% of the total power consumption in the world [2].
`The power consumption originates from the 3 sectors in it. It
`includes servers, air conditioner and power supplies. In order
`to suppress the total power consumption, we need to increase
`each power-efficiency of each equipment and also increase the
`total management efficiency of the operation [3].
`Nowadays, server power is dramatically increasing in the
`data center including high power computer infrastructure
`(HPCI) and high-end General-purpose computing on graphics
`processing units (GPGPU) system. For such high power
`data center, air-cooling technology is not effective enough to
`handle the heat generated.
`PUE (Power Usage Effectiveness) [4] is an effectiveness of
`
`
`978-1-5090-4026-1/17/$31.00 ©2017 IEEE
`
`
`
`
`Fig.1. Example of total power consumption of data centers with air cooling
`and liquid immersion cooling. Liquid Immersion Cooling system can reduce
`total power consumption.
`
`
`In this study, we proposed and demonstrated a sophisticated
`cooling
`technology for such high power consumption
`server board
`in data centers. Also, we are aiming
`to
`demonstrate the next challenging and super high level stage,
`PUE =1.0x (x<4). In order to demonstrate the high efficient
`data center for high heat density servers, the liquid immersion
`technology was applied in this study. Already, several trials of
`the liquid immersion technology were performed, including
`immersion with pumping to provide circulation [5,6,7] and 2
`phase immersion [8]. In those systems, low PUE was
`demonstrated for high heat density servers, as shown in Fig. 1.
`Also, in a conventional data center with air cooling system, lot
`of space is required for effective cooling, in which air
`conditioners and server racks are located. On the other hand,
`the required space for the immersion cooling technology can
`be ideally suppressed below around 1/3 of that of an air
`cooling system, as shown in Fig.2.
`On the other hand, our technology is the liquid immersion
`technology with natural convection, without any pumping or
`any fan in the bathtub for the refrigerant circulation. This
`system enables high efficient cooling technology for high
`density server with PUE= 1.0x, which is expected to be lower
`
`Authorized licensed use limited to: Alexander Karl. Downloaded on June 17,2021 at 17:21:20 UTC from IEEE Xplore. Restrictions apply.
`
`Immersion Systems LLC – Ex. 1017
`PGR 2021-00104 (U.S. 10,820,446 B2)
`1 of 7
`
`

`

`cooling
`immersion
`demonstrated
`already
`the
`than
`technologies. Because, for operation of this system, electric
`power required for circulating the refrigerant is unnecessary,
`and only electric power for circulating the cold water having a
`low specific gravity inside the cold plates and heat exchange
`of the water is required.
`Also, for the conventional liquid immersion technology,
`floor loading level sometimes exceeds 1,000kg/m2, which is
`higher than the standard of conventional building. Therefore,
`low floor loading level is also one of our goal feature.
`
`In
`this paper, computational fluid dynamics (CFD)
`simulation was tried at first for making basic design of the
`liquid cooling immersion system with natural convection, and
`next the actual experiment was demonstrated.
`
`
` By using CFD, we achieved the natural convection without
`any pump or any fan inside the bathtub. For the simulation,
`Flow Designer Ver. 2017 of Advanced Knowledge Laboratory,
`Inc. was used [9]. Figure 5 shows a CFD model we
`constructed. Upper one is a top view of the bathtub and lower
`ones are side view and front view of the bathtub. The light blue
`part represents cold plates placed in the bathtub. Dark blue part
`on top of the side and front views are filled by air.
`For the actual experiment, 6 server units were placed in
`the bathtub, each server unit consists 4 servers, and eventually
`48 CPUs (Intel Zeon processor which power reaches around
`
`
`Fig.2. Comparison of space of data center with conventional air cooling
`technology and with immersion technology.
`
`
`
`
`II. EXPERIMENTAL PROCEDURES
`The boards were immersed into the bathtub filled by the
`refrigerant without any pump or any fan for the refrigerant
`circulation in the bathtub. The models of the liquid immersion
`cooling system with natural convection and a snap shot of an
`actual experimental bathtub are shown in Figs 3 and 4.
`For the CFD simulation, the 2 CPUs with 140 W power at
`maximum and 16 memory boards with 10 W power at
`maximum were set on a mother board. The 24 or 48 mother
`boards (48 or 96 CPUs) were immersed into the bathtub with
`600mm X 600mm X 870mm. The 9,000,000 meshes are
`applied. The size of each mesh is small enough for the
`simulation, because the spacing between the board is smaller
`than the mesh size. The total power of the system reached a
`maximum of 14kW. Four cooling plates with 560mm X
`560mm were implemented in the bathtub, three of them were
`placed parallel to the wall and one to the bottom, and cool
`water with temperature between 15 to 35 degrees C was
`flowed inside the plates. The heat exchanger for making the
`cold water is located outdoor. As mentioned above, the
`electric power required for circulating the refrigerant is
`unnecessary, and only electric power for circulating the cold
`water having a low specific gravity inside the cold plates and
`heat exchange of the water is required.
`
`
`Fig.3. Experimental apparatus for liquid immersion cooling with natural
`convection.
`
`
`
`Fig.4. Snap shot of liquid immersion system with natural convection.
`
`
`
`
`
`Fig.5. CFD model for liquid immersion with natural convection.
`
`
`
`
`Authorized licensed use limited to: Alexander Karl. Downloaded on June 17,2021 at 17:21:20 UTC from IEEE Xplore. Restrictions apply.
`
`Immersion Systems LLC – Ex. 1017
`PGR 2021-00104 (U.S. 10,820,446 B2)
`2 of 7
`
`

`

`140 W at maximum task) are immersed into the bathtub. The
`CPU power of 100 W was applied to each CPU at maximum
`task. The total size of the system actually reaches around 1/3
`of the conventional rack structure for air cooling, as already
`shown in Fig.2.
` As the refrigerants, we applied several types of non-
`conductive, thermally and chemically stable fluids, including
`silicone oil, soybean oil, and perfluorocarbon structured
`refrigerant (Fluorinert) [10].
`
`
`III. RESULTS AND REPARE DISCUSSIONS
`
`A. CFD Simulation –Determination of Optimum Refiregirant-
`Typical cooling water flows inside the cooling plates
`analyzed by CFD is shown in Fig.6. The water flow in 3 side
`cooling plates and one bottom cooling plate are clearly seen.
`The cooling water is flowed from the heat exchange system
`located outside. The temperature and the flow rate were
`changed from 15 to 35 degrees C and from 0.005 to 1 m/sec,
`respectively. A typical refrigerant circulation analyzed by CFD
`is shown in Fig.7. Each line represents the flow of the
`refrigerant. As clearly seen in this figure, no pumps and no fans
`to generate convection are contained in the bathtub, and the
`refrigerant is circulating around the board with natural
`convection. This convection is caused by an upward flow due
`to the heat generated only by the CPU. In other words, the
`driving force of the natural convection is the only heat of the
`CPUs.
`
`circulates only based on the natural convection and the heat is
`removed from the CPU surface, effectively.
`
`
`
`
`Fig.7. Typical flows of the refrigerant in the bathtub with natural convection.
`
`
`Fig.8. Typical temperature distribution inside the liquid immersion tub with
`natural convection. In this case, the refrigerant is 30% ethylene glycol.
`
`
`
`
`Fig.6. Typical cooling water flow in the cooling plates.
`
`temperature
`typical snapshot of
`Figure 8 shows a
`distribution. As clearly seen in the image, CPUs are heated
`and the refrigerant heated by the CPU is circulating from the
`bottom to the top of the bathtub. Figure 9 also shows a snap
`shot of flow rate distribution of the refrigerant viewed from
`the front and side. The space above the middle line is the air
`layer, and the refrigerant is filled up to the line. As clearly
`seen in the image, an upward flow is generated by the heat of
`CPU. The flow rate is about 0.1 m/sec at 140 W. Figure 10
`also shows typical temperature distributions inside the cooling
`plates. As shown in the figure, temperature on the top of the
`cooling plate is higher than that on the bottom of the cooling
`plate. This
`temperature distribution corresponds
`to
`the
`refrigerant circulation and heat transfer from the refrigerant to
`the cooling plate. These results indicate that the refrigerant
`
`Fig.9. Typical flow rate distribution inside the liquid immersion tub with
`natural convection. In this case, the refrigerant is 30% ethylene glycol.
`
`
`
`
`
`
`Authorized licensed use limited to: Alexander Karl. Downloaded on June 17,2021 at 17:21:20 UTC from IEEE Xplore. Restrictions apply.
`
`Immersion Systems LLC – Ex. 1017
`PGR 2021-00104 (U.S. 10,820,446 B2)
`3 of 7
`
`

`

`numbers of the refrigerants, a sort of the index of the viscosity,
`and CPU surface temperature. Also, Fig.13 represents a
`relationship between the Rayleigh numbers and flow rate of the
`refrigerant on the CPU surface. Here, the Rayleigh numbers are
`estimated at 25 degrees C due to the model’s geometry. As
`seen in these figures, the CPU temperature monotonically
`decreases and the flow rate increases with the Rayleigh
`numbers. These results indicate that the smooth refrigerant, like
`fluorinate, is better for cooling the high power density CPU. In
`this model, we assumed that the heat generation in the CPU is
`uniform. Therefore, the temperature monotonically decreases
`from inside the CPU and to the CPU surface. Since, in the real
`CPU, heat generates only at the junction point, this model is
`not accurate, but as we will look at the experimental results
`later,
`the
`simulation
`results qualitatively explain
`the
`experimental results well.
`The CPU temperature monotonically decreases until a
`certain water flow level. As long as the flow rate of the water
`inside the cooling plate is around 0.1 m/sec, the heat generated
`from the CPU is removed well at any temperature. The larger
`the Rayleigh number is, also the lower the water temperature
`inside the cooling plate is, the better the cooling performance
`is. When the CPU power reaches 140W, in case of FC3283, in
`order to keep the CPU surface temperature below 50 degrees
`C, the water temperature in the cooling plate should be reduced
`below 30 degrees C. Also, when using Si50, it is necessary to
`keep the cooling plate water temperature below about 20
`degrees C.
`It was also found that when the heat sink is attached to the
`CPU surface, even if the Rayleigh number of the refrigerant
`changes, its cooling efficiency is not greatly affected. This
`indicates that the heatsink with fins spaced by 6 mm (set in the
`CFD model) was not affected by the natural convection of the
`refrigerant.
`
`B. CFD Simulation -Impact of Slot Removal-
`In this section, we investigate the cooling effect on other
`slots when some slot is inserted or removed. Figure 14 shows
`a typical convection pattern when one slot is removed from
`the bathtub. As shown in the figure, it has little influence.
`Interesting thing is that the flow of refrigerant is nearly zero
`without refrigerant flowing backward (in this case downward
`flow) in the empty space from which the slot is removed. This
`indicates that the convection occurs only in a certain portion
`of the slot, so the change in convection in the portion without
`slots does not affect other slots. From this, it seems that as a
`kind of pipe model, circulation paths for the refrigerant with
`the same cross section is made, so that the flow velocity of the
`circulation flow path becomes constant. Therefore,
`the
`refrigerant does not flow backward to the empty slot. This has
`very advantageous features in terms of maintenance. This
`means that any board can be inserted and removed even
`during operation without change in the cooling parameters.
`
`C. CFD Simulation -Impact of Task Allocation-
`In the same way, the influence of the cooling effect on
`other CPUs and slots when task is given to a CPU in a certain
`
`Fig.10. Typical CFD image of water flow inside the cooling plates.
`
`
`
`
`
`Fig.11. Typical front views of temperature distribution for several refrigerant in
`liquid immersion tub with natural convection.
`
`
`Next, we analyze the thermal characteristics for several
`refrigerants. Figure 11 shows some patterns of temperature
`distribution for several refrigerants. From the left side, the
`pattern corresponds to fluorinate FC3283, and fluorinate FC43
`[10], and silicon oil 20, also silicon oil 50, and last, on the right
`side, soybean oil. As clearly seen in the images, the
`temperature of the CPU reached around 50 degrees C in every
`case. Also, when the fluorinate was used as a refrigerant, the
`temperature of the refrigerant was kept lowest. When silicon
`oil is used, the refrigerant temperature reached around 50
`degrees C. This suggests that fluorinate FC3283 and FC43 of
`smoother refrigerant are better than silicon 20, 50 and soybean
`oil of muddy or sticky refrigerants. These results indicate that
`the cooling effect for natural convection strongly depends on
`the viscosity of the refrigerant.
`The more detailed simulation results are shown in Figs.12
`and 13. Fig.12 represents a relationship between Rayleigh
`
`Authorized licensed use limited to: Alexander Karl. Downloaded on June 17,2021 at 17:21:20 UTC from IEEE Xplore. Restrictions apply.
`
`Immersion Systems LLC – Ex. 1017
`PGR 2021-00104 (U.S. 10,820,446 B2)
`4 of 7
`
`

`

`portion. The impact of the CPU task is shown in Figs.14 and
`15. As shown in the figure, the effect of the task on the other
`CPUs and slots is limited even when some CPU power
`changes dynamically. The impact of the CPU power to other
`CPU temperature are summarized in Figs.16 and 17. As can
`be seen from these figures, the higher the Rayleigh number,
`the lower the CPU surface temperature is as mentioned above.
`For any of the refrigerants, only the surface temperature of the
`CPU whose electric power has been changed is changing and
`the CPU surface temperature above or below that the electric
`power is not changed hardly changes. In other words, The
`influence on the other CPU due to the CPU power of a certain
`portion is limited. Since the convection which depends on the
`heat of the CPU occurs, the convection does not affect other
`CPUs and other slots. This has very advantageous features in
`terms of operation.
`
`D. Experiments
`Next, we performed the actual experiment. The simulation
`results predict well the experimental results qualitatively. The
`natural convection enables the effective cooling of high power
`density board. As described before, the 24 server boards are
`immersed into the refrigerant bath. No pump or no fan is
`placed inside this bathtub. From the CFD results, among the
`refrigerants we applied, we concluded that FC3283 is the best
`refrigerant with the best cooling performance in case of the
`immersion technology. Therefore, hereafter, we use the
`FC3283 as a refrigerant in an actual experiment.
`The cooling performance of immersion technology using
`natural convection is analyzed by the actual experiment. The
`lower the temperature of flowing water inside the cooling
`plate, and also the higher the flow rate, the higher the cooling
`efficiency of CPU. In other words, it is understood from these
`results that natural convection technology can be used
`sufficiently within a practical range.
`
`
`
`Fig.13. Relationship between the Rayleigh numbers of the refrigerant and
`flow rate on the CPU surface. Here, the CPU power is 140 W, and water flow
`rate is 0.1 m/sec.
`
`
`
`
`
`Fig.14. Typical convection pattern when task is not applied to a part of
`CPUs(b) and some slot is removed from the bathtub(c).
`
`
`
`
`
`Fig.12. Relationship between the Rayleigh numbers of the refrigerant and
`temperature of the CPU surface. Here, the CPU power is 140 W, and water
`flow rate is 0.1 m/sec.
`
`
`
`Fig.15. Typical CPU temperature distribution when over-task is applied only
`to CPUs of one slot. The heat of the CPU’s do not affect the other CPU’s
`temperature.
`
`
`
`
`Authorized licensed use limited to: Alexander Karl. Downloaded on June 17,2021 at 17:21:20 UTC from IEEE Xplore. Restrictions apply.
`
`Immersion Systems LLC – Ex. 1017
`PGR 2021-00104 (U.S. 10,820,446 B2)
`5 of 7
`
`

`

`
`
`
`Fig.16. Relationship between Rayleigh number of the refrigerant and CPU
`surface temperature. This graph shows the change in each CPU’s surface
`temperature when the upper CPU’s power changes from 140 W to 0 W. Here,
`circle dots correspond to the upper CPU’s surface temperature and triangle
`dots correspond to the lower CPU’s surface temperature. Only upper CPU’s
`temperature changes.
`
`
`
`
`Fig.17. Relationship between Rayleigh number of the refrigerant and CPU
`surface temperature. This graph shows the change in each CPU’s surface
`temperature when the lower CPU’s power changes from 140 W to 0 W. Here,
`circle dots correspond to the upper CPU’s surface temperature and triangle
`dots correspond to the lower CPU’s surface temperature. Only lower CPU’s
`temperature changes.
`
`
`
`
`Figure 18 shows the change in CPU junction temperature
`with respect to the CPU power. Unlike simulation, the
`temperature is not the CPU surface temperature but the
`junction temperature, because we cannot measure the surface
`temperature in this real environment. The junction temperature
`data comes out from each CPU itself. As clearly seen here, the
`CPU junction temperature rises linearly with the CPU power.
`In this system, the upper limit of CPU junction temperature is
`97 degrees C. As seen here, it can be said that up to
`120W CPU power can be applied. These results show that
`this natural convection method has practically enough
`performance, considering that the maximum CPU power of
`HPCI is almost around 100 W at maximum.
`
`
`
`
`Fig.18. Relationship between CPU power applied and CPU junction
`temperature (Tj) for liquid (FC3283) immersion with natural convection.
`
`
`
`
`Likewise, the effect of heat sink on the CPU is shown in
`Fig.19. The solid line represents the case with CPU heat sink
`(corresponding to the left side of the pictures) and dashed one
`represents the case without CPU heat sink (corresponding to
`the right side of this pictures). As clearly seen in this figure,
`by attaching a heat sink, the cooling effect with heatsink is
`about twice of that without heatsink. In other words, if there is
`no CPU heat sink, the upper limit of CPU power is only about
`60W, that is, tasks of about 50% can be applied. The CFD
`results shows that the heatsink effect is not remarkably high.
`This is because, the geometry of the heatsink model in the
`CFD is not perfect.
`the actual
`the PUE, based on
`We also evaluated
`experiment with natural convention. The PUE was calculated
`with the power consumption of the servers, and pumping
`motors and heat exchange system implemented outside the
`system. For operation of this system, electric power required
`for circulating the refrigerant is unnecessary, and only electric
`power for circulation of the cooling water with a low specific
`gravity and heat exchange of the water flowing through the
`cold plate is required. Eventually, the PUE was demonstrated
`confirmed up to around 1.04. This value will be lower by
`improving the heat exchange system. Also, the floor load of
`this system was suppressed below 1,000 kG/m2.
`These results indicate that the proposed technology
`exhibits promising potential for practical cooling technology
`in an actual data center.
`Although the simulation results by CFD need to further
`adjust the absolute value, the tendency is in good agreement
`with the experiment result.
`
`
`
`
`
`Authorized licensed use limited to: Alexander Karl. Downloaded on June 17,2021 at 17:21:20 UTC from IEEE Xplore. Restrictions apply.
`
`Immersion Systems LLC – Ex. 1017
`PGR 2021-00104 (U.S. 10,820,446 B2)
`6 of 7
`
`

`

`system, PUE below 1.04 was demonstrated. Also, the floor
`load of this system was suppressed below 1,000 kG/m2.
` Consequently, this proposed technology with natural
`convection exhibits promising potential for low energy and
`space-saving board cooling in data center.
`This system can be applied to large-scale clouds including
`HPCI and supercomputer systems, because it scales just by
`connecting the liquid tank sideways. In addition, since the
`refrigerant used in this study dries at the time of picking up the
`board, the maintenance property is much better than in the
`case of using the oil.
`On the other hand, since the cost of the refrigerant is still
`high at this moment, the initial investment is necessary and the
`evaporation of the refrigerant must also be supplemented. To
`solve the economic aspect may be a future challenge.
`
`
`ACKNOWLEDGMENT
`This study was supported by the development and
`demonstration projects for CO2 emission reduction of the
`Ministry of Environment
`in Japan. CFD analysis was
`supported by Advanced Knowledge Laboratory, Inc. The
`perfluorocarbon structured refrigerants (Fluorinert) were
`provided by 3M company. The heat exchange system for
`cooling water was developed by Takasago Thermal
`Engineering Co., Ltd. We sincerely thank for their helpful
`comments and discussions. We also thank Mr. Fujimaki and
`Mr. Yamamoto of Fujitsu Limited for their fruitful discussions.
`
`
`
`REFERENCES
`
`[1] http://www.businessinsider.com/internet-of-things-cloud-computing-
`2016-10
`[2] https://energy.gov/eere/buildings/data-centers-and-servers
`[3] Kazumasa Kitada, Yutaka Nakamura, Kazuhiro Matsuda, and Morito
`Matsuoka, “Dynamic power simulation utilizing computational fluid
`dynamics and machine learning for proposing ideal task allocation in a
`data center”, Internationa Conf. on Cloud Computing, 2016.
`[4] https://www.thegreengrid.org/en/newsroom/news-releases/statement-
`green-grid-regarding-pueashrae-standard-904/
`[5] Fujitsu. 2016. Totally Submerging a Server in Liquid Reduces Power
`Consumption up to 30% - Data Center Innovation by a Novel Cooling
`Technology. http://journal.jp.fujitsu.com/en/2016/08/15/01.
`[6] Erich Strohmaier, Jack Dongarra, Horst Simon, and Martin Meuer.
`2017. The GREEN 500 LISTS June 2017.
`https://www.top500.org/green500/lists/2017/06.
`[7] David Prucnal. 2015. Doing more with less: Cooling computers with oil
`pays off. The Next Wave. Vol.20 No.2:21-29.
`[8] http://www.3m.com/3M/en_US/novec/products/engineered-
`fluids/immersion-cooling/
`[9] http://www.akl.co.jp/en/
`[10] http://www.3m.com/3M/en_US/company-us/all-3m-products/~/3M-
`Fluorinert-Electronic-Liquid-FC-3283?N=5002385+3294001628&r
`
`
`Fig.19. Relationships between CPU power applied and CPU junction
`temperature with heatsink and without heatsink.
`
`
`
`
`
`IV. SUMMARY
`We proposed an immersion cooling technology with
`natural convection for high power servers used in data centers.
`The cooling performance was evaluated by CFD simulation
`and actual experiments. Although the simulation results by
`CFD need to further adjust the absolute value, the tendency is
`in good agreement with the experiment results. The smoother
`refrigerant is better for cooling the high power CPUs. Among
`the refrigerants tried in this study, a perfluorocarbon structured
`refrigerant (Fluorinert) proved to be the most suitable for
`immersion cooling with natural convection.
`The change in any CPU task and any slot-removal gives
`limited cooling effect to other slots and other CPUs. In other
`words, the change in the parameters of a certain portion,
`including CPU and slot, is limited only to that portion itself,
`the
`influence on other places
`is very small. These
`characteristics are extremely excellent from the aspect of
`stable operation of the data center.
`For operation of this system, electric power required for
`circulating the refrigerant is unnecessary, and only electric
`power for circulating the cold water inside the cooling plates
`and heat exchange of the water is required. Eventually, in this
`
`
`
`Authorized licensed use limited to: Alexander Karl. Downloaded on June 17,2021 at 17:21:20 UTC from IEEE Xplore. Restrictions apply.
`
`Immersion Systems LLC – Ex. 1017
`PGR 2021-00104 (U.S. 10,820,446 B2)
`7 of 7
`
`