FDM09-I Bypass Thermal Mass Flow Transmitter Technical White Paper

Venturi × Bypass × Thermal Mass Flow

Flow and liquid temperature monitoring for AI data center liquid cooling, CDUs, manifolds, cold plate loops and semiconductor equipment cooling

1. Why Do Liquid Cooling Systems Need to Detect Subtle Flow Changes?

In AI, high-performance computing and high-power electronic equipment, coolant circulates continuously through CDUs, manifolds, equipment branches and cold plates. For equipment management, flow is more than a single reading: it is a key indicator of whether the liquid circuit is maintaining normal circulation.

As cold plate resistance increases, local pipe blockages develop, valve states change or pump operating conditions shift, main-pipe flow may initially show only a slight decrease or fluctuation. Clearer detection of these changes can provide earlier data to support equipment trend analysis, alarm settings and preventive maintenance.

From Instantaneous Readings to Change Detection Liquid cooling equipment needs to show the current flow and whether that flow is changing over time.

What Customers Need to Know	Conventional Focus on Instantaneous Readings	FDM09-I Design Approach
Is flow gradually deteriorating?	Displays the current flow	Focuses on change detection and trends

Key Concept: Accuracy and Sensitivity Are Different.

Accuracy describes the error between the measured value and the actual flow. Sensitivity concerns whether the sensing signal responds clearly to a small change in input flow. The FDM09-I uses a Venturi and a bypass flow path to place the thermal mass sensing element in operating conditions better suited to detecting flow changes.

2. How Does the FDM09-I Measure Flow? A Visual Overview

FDM09-I Venturi main flow path and bypass thermal mass sensing structure — FDM09-I Venturi Main Flow Path + Bypass Thermal Mass Sensing Structure

1Main Flow Passes through the Venturi Changes in the main-pipe cross section create differences in flow velocity and pressure that can be used for measurement.

2P1–P2 Creates a Differential Pressure This differential pressure drives flow through the bypass.

3A Small Portion of the Liquid Passes through the Thermal Mass Sensing Section The bypass orifice diameter and flow resistance determine the operating flow rate and flow velocity.

4The Heat Transfer Signal Is Converted to Total Main-Pipe Flow Calibration on the actual device establishes the relationship between the signal and main-pipe flow.

The FDM09-I Design Does Not Aim to Maximize Bypass Flow, but to keep the thermal sensor within a suitable, stable operating range that supports the detection of subtle changes over extended periods.

2-1. FDM09-I Mechanical Design: Venturi + Bypass Thermal Mass Sensor

The figure below shows the current FDM09-I mechanical design and how the Venturi main flow path is integrated with the thermal mass sensor. The P1–P2 differential pressure directs a portion of the liquid into the thermal mass sensing section. Calibration then establishes the relationship between the bypass signal and total main-pipe flow.

FDM09-I Venturi pipe mechanical design integrated with a bypass thermal mass sensor — **Venturi Pipe Mechanical Design**
Venturi main flow path + bypass thermal mass sensor; designed to extend the measurement range while maintaining the main-pipe flow field.

FDM09 Three-Cycle Measurement Consistency

Three-Cycle Flow Measurement Accuracy Verification
Three-cycle average measurement accuracy: maximum error of approximately 0.63% F.S., meeting the ±1% F.S. requirement.

Expand to View Three-Cycle Average Measurement Data

Three-Cycle Average Reference Flow (L/min)	Three-Cycle Average Digital Flow (L/min)	Average Error (L/min)	Average CTA
13.96	14.51	0.54	1.892
33.48	33.86	0.38	1.993
45.88	46.42	0.54	2.034
54.02	54.27	0.25	2.054
61.99	61.89	-0.11	2.072
70.43	70.50	0.07	2.088
76.94	77.33	0.39	2.098
83.37	83.94	0.57	2.108
89.88	90.00	0.12	2.116
96.20	96.98	0.78	2.125
102.60	102.98	0.38	2.131
108.55	109.10	0.55	2.138
114.50	115.23	0.73	2.143
120.57	121.27	0.70	2.149
126.48	127.12	0.63	2.154
132.19	133.15	0.95	2.159
138.07	138.96	0.88	2.163
143.91	144.77	0.86	2.168
149.69	150.23	0.54	2.173

3. Selecting the P1–P2 Bypass Orifice: Ø2 / Ø2.5 / Ø3 mm

The following first-stage design estimates are used to guide device verification. Assumptions: water at 25°C, ρ≈997 kg/m³, orifice discharge coefficient Cd≈0.62 and sensor internal diameter Ø3.8 mm. Calibration on the actual product must also account for pipe length, elbows, fittings, the sensor body and inlet/outlet flow resistance.

FDM09-I design estimates for flow velocity at different P1–P2 differential pressures with Ø2, Ø2.5 and Ø3 mm bypass orifices

Orifice Diameter	Bypass Characteristics	Flow Velocity at ΔP=2 kPa	Flow Velocity at ΔP=4 kPa	Design Assessment
Ø2.0 mm	Smaller	0.34 m/s	0.49 m/s	The signal may be weaker at low flow
Ø2.5 mm	Balanced	0.54 m/s	0.76 m/s	Recommended for Initial Verification
Ø3.0 mm	Larger	0.77 m/s	1.10 m/s	Flow velocity is relatively high at higher differential pressure

These values are first-stage design estimates. Final product specifications must be established through testing and calibration on the actual device.

4. Four Liquid Flow Measurement Technologies: A One-Page Comparison

No single flowmeter is the best choice for every application. Selection should consider the fluid, flow range, pressure loss, piping conditions, bubbles, installation space and downstream system integration. The table below provides a qualitative comparison for liquid cooling and equipment cooling applications.

Comparison Item	FDM09-I Bypass Thermal Mass	Electromagnetic	Vortex	Ultrasonic
Operating Principle	Venturi differential pressure + bypass heat transfer	Faraday's law of electromagnetic induction	von Kármán vortex shedding frequency	Ultrasonic propagation / transit-time difference
Fluid Requirements	Calibration is required for thermal properties; independent of electrical conductivity	Generally requires a conductive liquid	Liquids / gases / steam; stable vortex shedding is required	Requires suitable acoustic conditions
Low Flow / Subtle Changes	Design focus: a high-sensitivity operating range	Generally good, depending on pipe diameter and model	Low-flow performance is limited by the Reynolds number and signal strength	Depends on piping, fluid and signal quality
Main-Pipe Pressure Loss	Low to moderate; depends on Venturi and bypass design	Low; the measuring tube typically has no bluff body	Moderate; includes a bluff body (shedder bar)	Low; clamp-on models are particularly non-invasive
Effect of Bubbles	Large quantities of bubbles should be kept out of the sensing section	May affect measurement stability	Single-phase, stable flow conditions should be maintained	Bubbles may cause signal attenuation or measurement errors
Temperature Integration	Liquid temperature measurement can be integrated	Depends on the product configuration	Some products can integrate T/P measurement	Depends on the product configuration
Liquid Cooling Branch Applications	Suitable for monitoring subtle flow changes and liquid temperature	Suitable for general-purpose measurement of conductive coolants	Suitable for medium to high flow velocities and general industrial fluids	Suitable where low pressure loss or clamp-on installation is a priority

5. Which Customers Are Best Suited to Each Measurement Principle?

Operating principles and applications of bypass thermal mass, electromagnetic, vortex and ultrasonic flowmeters

6. Product Selection: Start with the Primary Requirement

Liquid cooling flowmeter selection based on pressure loss, fluid, flow velocity and integrated flow and liquid temperature measurement

Selection should consider more than a single accuracy figure. For liquid cooling equipment, other factors typically include minimum and maximum operating flow, the magnitude of flow changes, allowable pressure loss, pipe diameter, coolant formulation, bubbles, installation space and whether the system requires flow, liquid temperature and totalized flow data.

When Clamp-on, Non-invasive Installation and Low Pressure Loss Are Priorities	Evaluate ultrasonic flowmeters first
When the Coolant Is Conductive and Proven General-Purpose Measurement Is Required	Evaluate electromagnetic flowmeters first
For General Industrial Applications with Medium to High Flow Velocities: Liquids / Gases / Steam	Evaluate vortex flowmeters first
When Subtle Flow Changes in Liquid Cooling Branches Need to Be Detected	Evaluate the FDM09-I first
When Liquid Temperature and Flow Trends Need to Be Integrated	The FDM09-I Can Offer a Distinctive Advantage

7. From Flow Measurement to Liquid Cooling System Data

The FDM09-I Provides Value beyond Flow Measurement by turning flow and liquid temperature into data that a system can use for analysis.

Available Data / Potential Extensions	Customer Value
Real-Time Flow	Monitor the current circulation status of the liquid circuit
Totalized Flow	Record the total quantity passing through over a period of time
Liquid Temperature	Understand the current thermal condition
Low-Flow Alarm	Provide an initial signal for abnormal-condition management
Trend Analysis	Observe changes in cold plate resistance, pipe blockages, pumps or valves
Heat Removal Rate	Can be calculated using the supply/return liquid temperature difference and coolant specific heat
Accumulated Cooling Energy	Integrating the heat removal rate over time can support energy management in kWh

Application Positioning

FDM09-I applications for liquid cooling flow and liquid temperature monitoring

8. From Design Estimates to Product Specifications: FDM09-I Validation Roadmap

This white paper distinguishes design concepts, engineering estimates and production specifications. The current P1–P2 orifice diameters and thermal mass sensor flow velocities are design estimates. Final specifications should be established from calibration results on the actual device.

Step 1 | Water Circuit Baseline Collect synchronized data for main-pipe flow Qmain, P1–P2 ΔP and bypass flow Qbypass.

Step 2 | Orifice DOE Compare low-flow sensitivity, linearity, pressure loss and repeatability for Ø2 / Ø2.5 / Ø3 mm orifices.

Step 3 | Thermal Mass Sensor Signal Establish the thermal mass sensor output response curve against Qbypass to identify the optimum operating range.

Step 4 | Temperature Compensation Establish compensation coefficients at different liquid temperatures to reduce the effects of changes in thermal properties and viscosity.

Step 5 | Fluid K-Factor Establish a K-factor or corresponding calibration for water and different ethylene glycol concentrations.

Step 6 | System Validation Verify full-pipe conditions, bubbles, mounting orientation, straight pipe lengths, long-term stability and pressure loss.

9. Conclusion: What Makes the FDM09-I Different?

The FDM09-I combines a Venturi, a bypass flow path and thermal mass sensing to direct a portion of the liquid from the main pipe into a dedicated thermal sensing section.

The primary design objective is to keep the thermal mass sensing element in operating conditions better suited to detecting subtle flow changes. Flow, liquid temperature, instantaneous readings and totalized readings can then be integrated into liquid cooling equipment systems to support trend monitoring, abnormal-condition alarms and energy management.

“More sensitive, more complete flow and thermal condition data for liquid cooling systems.”

References for the Technology Comparison

Magnetic Flowmeters / Flowmeters: Electromagnetic flowmeters use Faraday's law of electromagnetic induction. They are suitable for conductive liquids, and their measuring tubes can provide low pressure loss.
Vortex Flow Meter Technology / Low Flow Measurement: Vortex flowmeters use von Kármán vortex shedding. Low-flow performance is limited by the Reynolds number and signal strength.
Ultrasonic Flowmeter Operating Instructions: Excessive bubbles in the liquid may cause measurement errors. Installation location and acoustic conditions must be considered.

Disclaimer: The comparisons in this white paper are qualitative descriptions of measurement principles and typical applications. Actual performance, accuracy, measurement range, pressure loss and installation requirements for each brand and model are subject to its official specifications. The FDM09-I P1–P2 orifice diameters and thermal mass sensor flow velocities are currently design estimates. Final specifications must be established through calibration on the actual device.