Design Article
Case study: Practical in-system programming for PCM
Natalie Hunter, BPM Microsystems
12/26/2012 9:45 AM EST
The key advantage of in-system programming is that it allows design engineers and production manufacturers to integrate semiconductor device programming and testing into a single step, eliminating the need to program a device before board placement. Flash memory is often preprogrammed in order to maintain high throughput because traditional in-system programming solutions do not have sufficient speed for typical high-density flash memories. Phase change memory (PCM) is a non-volatile memory technology that features increased system-level reliability, byte-alterability and higher programming rates compared to flash devices. By using BPM Microsystems’ high-speed 2800ISP in-system parallel programming system, customers can apply code or firmware just in time in a production environment, while maintaining a high manufacturing beat rate.
Embedded systems typically contain a microcontroller as the processing core, a non-volatile memory chip for firmware, and a managed NAND flash device for data storage. For this application, BPM Microsystems used a panel of eight PCBs, each containing one 8 Kb microprocessor, one Micron NP5Q128A13ESFC0E 128-Mbit PCM for application firmware, and one Micron MTFC4GLVEA-0M WT 32Gbit eMMC device for data storage (see figure 1).
Figure 1. BPM Microsystems’ eight-panel PCBs with Micron’s
PCM and eMMC device for data storage.
BPM Microsystems’ semi-automated 2800ISP in-system parallel programming solution has the ability to program flash memory devices at high speeds using its Vector Engine Co-Processor. The co-processor accelerates waveforms during the programming cycle. The faster speeds are achieved through synchronous operations that eliminate the dead times so the device under test no longer waits for the programmer. The result is programming near the theoretical limits of the silicon design—the faster the device, the faster the device is programmed.
With Vector Engine technology, high-density flash memory devices are able to achieve read/write speeds up to 140 Mb/s. By programming in both serial and parallel mode the 2800ISP solution is significantly faster than traditional serial in-system programming solutions.
Result
The 2800ISP programmed the serial PCM in 24.29 s, corresponding to a rate of 5.27 Mb/s. It was verified in 3.33 s or at a rate of 38.44 Mb/s. The signal integrity and speed is demonstrated by the waveforms captured during the in-system programming process (see figure 2).
As the table illustrates, with these high-speed program and verify times, over half a million boards can be programmed per year with a single 2800ISP. This solves the bottleneck typically seen with traditional test equipment, making the 2800ISP an efficient in-system programming solution for PCM devices.
About the author
Natalie Hunter is the marketing communications manager at BPM Microsystems.
Embedded systems typically contain a microcontroller as the processing core, a non-volatile memory chip for firmware, and a managed NAND flash device for data storage. For this application, BPM Microsystems used a panel of eight PCBs, each containing one 8 Kb microprocessor, one Micron NP5Q128A13ESFC0E 128-Mbit PCM for application firmware, and one Micron MTFC4GLVEA-0M WT 32Gbit eMMC device for data storage (see figure 1).
Figure 1. BPM Microsystems’ eight-panel PCBs with Micron’s
PCM and eMMC device for data storage.
Because of the attributes of PCM technology, any preprogrammed data to the device would be lost after reflow, therefore requiring in-system programming equipment such as in-circuit test systems, JTAG or memory interfaces. Most of the in-system programming solutions currently on the market do not have the ability to match the high programming speeds that PCM is capable of achieving, thus creating a manufacturing bottleneck.
BPM Microsystems’ semi-automated 2800ISP in-system parallel programming solution has the ability to program flash memory devices at high speeds using its Vector Engine Co-Processor. The co-processor accelerates waveforms during the programming cycle. The faster speeds are achieved through synchronous operations that eliminate the dead times so the device under test no longer waits for the programmer. The result is programming near the theoretical limits of the silicon design—the faster the device, the faster the device is programmed.
With Vector Engine technology, high-density flash memory devices are able to achieve read/write speeds up to 140 Mb/s. By programming in both serial and parallel mode the 2800ISP solution is significantly faster than traditional serial in-system programming solutions.
Result
The 2800ISP programmed the serial PCM in 24.29 s, corresponding to a rate of 5.27 Mb/s. It was verified in 3.33 s or at a rate of 38.44 Mb/s. The signal integrity and speed is demonstrated by the waveforms captured during the in-system programming process (see figure 2).
Figure 2. Waveforms captured during the in-system programming
process demonstrate signal integrity and speed.
process demonstrate signal integrity and speed.
As the table illustrates, with these high-speed program and verify times, over half a million boards can be programmed per year with a single 2800ISP. This solves the bottleneck typically seen with traditional test equipment, making the 2800ISP an efficient in-system programming solution for PCM devices.
About the author
Natalie Hunter is the marketing communications manager at BPM Microsystems.
|
|
|
|
|
Navigate to related information