Battery management IC solutions for small battery packs are proliferating at a feverish rate with the Internet of Things. Solutions vary in functionality and integration level. However, especially in larger battery packs for industrial and automotive use, I find that many battery management controller solutions out there are optimally designed for smaller packs, and some designers try to fit these into larger pack battery systems with mediocre or even sometimes disastrous results.
There are different challenges in the larger battery stacks, and these must be taken into account if a robust system design is desired. Enormous amounts of energy can be stored in large pack cells in series, sometimes in the area of tens of kilowatt-hours. Cell balancing is certainly more of a challenge in larger battery stacks since every cell in the stack must possess the same state of charge. This is a growing market.
Your choice as a designer must meet your system needs of physical size, performance (accuracy, reliability, power efficiency), and cost, depending upon the final system targeted deployment. Many times, if not most times, compromises need to be made and tradeoffs between specifications must be made. These are difficult decisions, and if you choose the best product for your design (this takes a great deal of research but is well worth your time to do this at the outset) then your design efforts will be rewarded with a smoother road and quicker time-to-market.
The following analysis will help you make those difficult design decisions. I will be having many more of these analyses on Planet Analog going forward.
The automotive industry, as well as the industrial market, needs to lower costs of automobile electronics, EVs, UPS, energy storage systems, e-bikes, e-scooters, etc. Forty-eight volt battery management systems (BMSs), as well as HV BMSs up to and exceeding 1,000 V, have growing needs, and we are about to see a whole new emphasis on lower-cost efforts coupled with a challenge to maintain and improve reliability and performance. The following is one of the first offerings that has caught my attention in this effort.
During my recent visit to Freescale in Arizona, where the Analog Power Business resides along with its nearby fab, I delved into the design details regarding battery management solution design for automotive and industrial needs for this product and in general as well.
Freescale solutions for battery management comprise two different types of devices: intelligent battery systems with an ADC, V, I, and temperature sense, which only measure battery parameters without controlling them, typically with a microcontroller on board and battery cell controllers, each with an analog front end, high-speed bus, and no microcontroller.
The Power design team made some bold design decisions with the MC33771 battery cell controller, like having a 14-cell capability all on a single monolithic IC so that the 48 V level could be served, such as in the telecom industry. I see a great many 12-cell solutions or less out there by many vendors. The Freescale devices can be stacked and daisy-chained via twisted-pair wire to handle greater than a 1,000 V battery solution, which can cover a very wide range of battery system needs.
Since the Automotive Safety and Integrity Level-C (ASIL-C), as part of the ISO26262 Functional Safety Standard, needs to be met in automotive and other high-reliability systems, cell information must be measured and transmitted very quickly. The CAN bus has been the go-to solution, especially in the automotive industry. Freescale designers chose a bold move to lower cost while still maintaining bus speed. Fast data acquisition and communication to the pack controller is achieved in only 2.6 ms for the pack controller to acquire conversion from 96 cells. This is critical to automotive and industrial safety.
A new PHY and communication bus was created that meets these needs. Getting away from the capacitive isolation used in the CAN bus eliminates the requirement of needing the capacitors be identical in order to maintain good system isolation. The Freescale solution uses a sinusoidal differential bus signal with one transformer per BMS IC for 3,750 V isolation. The MC33664 is the Physical Layer part of this solution, and only one of these is needed for an entire series of stacked MC33771s totaling more than 1,000 V.
The MC33771 design also integrates 14 FET balancing transistors inside each IC. A current sensor with Coulomb counting is also integrated to determine the internal impedance of the cells in one shot. The V and I are synchronized on chip in order to get extremely accurate impedance measurement of the cells within 65 µs. This provides useful information about its performance and an estimate of state-of-charge and can detect hidden trouble spots. High resistance values are often an indication to replace an aging battery. Finally, there are seven ports available for external temperature sensors.
The 2 mV voltage measurement accuracy in this IC ensures that the system can quickly and accurately identify when the high-slope region of the battery charge/discharge curve is reached. See Figure 2.
Fast time to market is essential, so having hardware that will speed evaluation as well as hasten time to market by accessibility to Gerber files in these expertly designed boards is a must. Both the MC33771 battery cell controller and the MC33664 transformer physical layer have KIT33771SP1EVB and KIT33664EVB evaluation kits. See Figure 3.
Watch Planet Analog for more of my analysis regarding new industry offerings with an emphasis on what you as a designer need to know about trends and new products.