Backup Battery Solutions for Medical Equipment

Challenges

In medical systems, a stable and reliable power supply is essential. To ensure an uninterrupted power supply, a backup battery is used. In the past, larger medical devices used lead-acid batteries to provide uninterrupted power. In fact, they also required very expensive and complex dynamic systems, which made the medical device system large, heavy and expensive. Now with the latest generation of battery charge monitoring electronics, we can safely use lithium-ion (Li-Ion) batteries to accurately determine the available power. This allows medical devices to be smaller and lighter than the previous lead-acid battery technology.

Common alternatives to lead-acid batteries are nickel metal hydride (NiMH) batteries or lithium-ion (Li-Ion) chemistries, both of which offer better energy density. Li-ion batteries offer the highest energy density using a more volatile chemistry that can be hazardous if not handled properly. For patient-critical systems, accurate prediction of remaining charge is critical regardless of the battery chemistry used. With Li-ion batteries, we get the best of both worlds: accurate knowledge of battery charge and the highest energy density.

Previous battery charge measurement electronics reported a gradually increasing error in remaining charge over time. We can only make empirical guesses about how much individual batteries “age” over time. The main reason why the available charge in lithium-ion batteries decreases over time is due to the increasing internal impedance of the electrolyte cathode/cathode materials. Lithium-ion batteries have some well-known properties, such as their strong temperature dependence, impedance changes during discharge, and the fact that high temperatures and rapid overvoltage charging can cause a significant drop in capacity. The internal impedance of a battery doubles after 100 charge/discharge cycles [1], as shown in Figure 1 (a cycle is defined as more than 70% of the charge flowing into or out of the battery). Charging at even 50mV above the maximum cell voltage of 4.2V can reduce the battery’s service life by half [1] (see Figure 2). A battery discharged more than 80% from room temperature to 0 degrees Celsius [1] will have a 5-fold increase in impedance (from N300mOhrn to more than 1.50hm DC impedance), see Figure 3.

Figure 1 Impedance changes with charge/discharge aging

Aging Li-ion batteries with higher impedance will reach the system termination voltage earlier.

Figure 2 Charging voltage affects battery life

Properly charging a lithium-ion battery requires a high-precision charging voltage. Overcharging will shorten the battery life.

Figure 3 Lithium-ion battery impedance is closely related to temperature and depth of discharge (DOD).

The impedance of lithium-ion batteries is closely related to temperature, and the impedance decreases by about 1.5 times for every 10°C increase in temperature.

Impedance is key to the whole equation. In the past, it was very tricky to get battery pack designs to work in production. Typical discharge characteristics at minimum/room/maximum temperatures were required to generate the coefficients used in the discharge estimation polynomial equation. This discharge characteristic can only be estimated by knowing how the impedance of each cell changes. In addition, traditional battery fuel gauge devices need to "reset" the maximum capacity of the battery pack when it is nearly fully discharged. Usually, this is achieved with a specific voltage trip of 7% and a 3% estimated remaining capacity. As an improvement, the Compensated End of Discharge Voltage Value (CEDV) is used to modify the 7% trip voltage and 3% estimated remaining capacity based on the battery load current, which is based on voltage measurements alone.

After accounting for all these uncertainties, designers know that the accuracy of reported capacity can vary by as much as 20%. Because batteries can age unexpectedly over time and because they can buffer the estimated capacity information provided by the fuel gauge and to the user, designers may double the amount of capacity actually required in advance. Of course, a robust medical system will not report the remaining capacity like a laptop: "You have 20 minutes of charge left, you need to plug in now." (This message appears when the battery reaches the estimated 7% remaining voltage.)

Solution

TI's next-generation Impedance Tracking™ algorithm technology solves the inaccuracy of reporting the true remaining battery capacity. The algorithm determines the state of charge of the lithium-ion battery and uses the following parameters as part of the overall battery model to fully predict the discharge behavior:

1. Initially, the total chemical capacity (Qmax) of the battery is the capacity specified in the product specification (for example, the capacity of a 18650 cylindrical lithium-ion battery is 2400mAhr), but the battery fuel gauge will automatically update after the first charge/discharge cycle of the battery.

2. The amount of charge that has flowed into or out of the battery is measured/collected by the "Coulomb Counting" program.

3. The current load current of the system (average load current and peak load current).

4. Since the impedance of a single battery is different under various charging conditions, the internal impedance of the battery will change with temperature, battery aging and discharge conditions while providing current.

5. When the battery is lightly loaded (

An accurate battery capacity estimate can be calculated using the following method:

1. Measure the open circuit voltage of the battery (in a relaxed state)

2. Monitor the battery voltage curve under load (find the battery impedance) and

3. Integrate the current flowing into and out of the battery.

Lithium-ion cells using exactly the same chemistry/anode/cathode materials have very similar relaxed voltage/state of charge curves. Surprisingly, they do not vary depending on the manufacture of the cell. This allows us to determine the maximum capacity of the cell and the remaining capacity of the cell.

For example, if you know that: 1) the 3.6V relaxed voltage correlates well with a 10% state of charge; 2) the battery gas gauge integrates 1000mA of current during charging; 3) the resulting 3.95V open circuit voltage correlates well with a 93% state of charge, the true capacity of the battery is 1206mAh (1000mA/83%). If the battery voltage rises from 3.6V to 3.8V while charging at 1A, the DC impedance is 0.2 Q at 10% state of charge and room temperature. If the minimum voltage the system can tolerate is 3V, then the impedance trace will calculate and report the remaining battery capacity of approximately 7 minutes under a 1A load at 10% state of charge.

In just the past few years, the electronic hardware implementation has evolved. The original chipset consisted of three separate chips: 1) a battery gauge microprocessor; 2) an analog front end (AFE); and 3) a secondary overvoltage protector. The microprocessor integrates the current and runs the battery gauge monitoring algorithm, and also communicates directly with the AFE. The high-voltage tolerant AFE measures the cell voltage with an integrated analog-to-digital converter (ADC), provides overcurrent protection, and performs cell balancing. Both chips can operate independently and safely. The third level of protection comes from an independent secondary voltage protector that triggers a chemical fuse for a permanent fault condition (overvoltage is the most dangerous condition for lithium-ion batteries because it can cause combustion).

The latest generation of Li-ion battery capacity indicators integrates a microcontroller and AFE chip in a plastic package, greatly reducing system-level complexity and board-level space requirements. Communication with the battery fuel gauge is done through the SMBus standard protocol (SMBus is based on the I2C communication protocol). For more information, please visit www.smbus.org. The battery fuel gauge can communicate directly with a compatible battery charger or microcontroller.

Impedance Tracking is essentially a lower-cost implementation of a battery solution that eliminates the need for the battery automatic learn cycle feature. This feature takes hours to implement for all large capacity batteries. Now, all batteries coming off the line are programmed using a tool called the "Golden Image". This file is created during the engineering evaluation phase. The Impedance Tracking algorithm will always be adapted to the state of the battery, so during the first discharge of the battery in the field, Impedance Tracking will accurately learn the true battery pack capacity in the first 40% discharge or charge of the battery. After that, the reported battery capacity will be 99% accurate.

in conclusion

Impedance tracking battery fuel measurement technology can enable medical engineering organizations to design more reliable life support equipment and portable devices with more stable backup batteries than ever before. More importantly, this technology not only provides greatly improved fuel gauge monitoring accuracy, but also eliminates the "reset" step required for the 7% estimated remaining charge (which is not practical in life support medical applications). It does not need to over-engineer battery capacity to meet a specific backup period, and does not need to repeat every battery pack during the production stage, thus providing a lower cost solution.

Understanding and tracking individual cell impedance is key to accurately predicting remaining charge. As mentioned previously, the most important cell aging effects are caused by high temperatures and charging at voltages above their maximum rating, and charging at voltages even 50mV above rated can cut a cell’s useful life in half. Li-ion cell internal impedance increases (ages) during normal use charge/discharge cycles, and impedance increases significantly at low temperatures (without reducing useful life).

The adaptive nature of the impedance tracking algorithm monitors these aging factors by monitoring the cell voltage under relaxed and loaded conditions, and integrating the current during charge/discharge. Because it is continuously monitored, there is no need to "guess" the impedance, so the true battery capacity can be accurately calculated throughout the battery's service life.