Features of Microchip Non-Volatile FPGA -Low Power Consumption-

I would like to introduce [low power consumption], which is a feature of Microchip's non-volatile FPGA, in three parts.

In the first installment, Microchip's non-volatile FPGA is [Why low power consumption? ] I would like to explain.

Advantages of low power consumption

Now that FPGA and heat generation are inseparable problems, the demand for low power consumption FPGA is increasing.

The low power consumption of FPGAs has the following advantages:

  • Solving thermal design problems
  • BOM cost reduction for heatsinks, fans, etc.
  • BOM cost reduction around power supply by reducing power consumption

 

Among the advantages of using Microchip's FPGA, the biggest advantage is low power consumption.

From now on, I would like to introduce why Microchip's FPGA has low power consumption.

Why low power consumption

There are two main reasons for the low power consumption of Microchip FPGAs:

 

1. Low power consumption with Flash-based technology

2. Transceiver optimization

 

First, I will explain from 1.

 

1. Low power consumption with Flash-based technology

To begin with, PLDs are broadly categorized into the following technologies:

 

a. Antifuse

b. UV erasable type (EPROM)

c. Electrically erasable ROM (EEPROM, Flash memory, etc.)

d. Static memory (SRAM)

 

c and d are programming technologies widely used in current FPGAs.

Next, a comparison of FPGA routing technologies is shown below.

Figure 1 Comparing FPGA routing technologies
Figure 1 Comparing FPGA routing technologies

SRAM becomes volatile and requires programming from an external configuration ROM. SRAM-based FPGA and Flash-based FPGA have lower power consumption due to the difference in cell structure. How are the cell structures different? Let's compare a competitive SRAM cell with Microchip's Flash cell.

Figure 2 Comparing Competitive SRAM Cells and Microchip Flash Cells
Figure 2 Comparing Competitive SRAM Cells and Microchip Flash Cells

As you can see from the structure of the SRAM cell, six transistors are used to configure one SRAM cell. A Flash cell, on the other hand, consists of a single transistor. The leakage current of each transistor in an SRAM cell is more than six times that of a Flash cell, resulting in very high static power consumption, including leakage current differences due to manufacturing processes. This is the difference from SRAM-based FPGAs, and one of the main reasons why Flash-based FPGAs have low power consumption.

 

At this point, some of you may be thinking, "Huh? Vendors other than Microchip also have Flash-based FPGAs." That's right. Vendors other than Microchip also have Flash-based FPGAs.

 

However, the structure inside the FPGA is different. Below is a structural comparison of Flash-based FPGAs.

Fig. 3 Structure comparison of Flash-based FPGA
Fig. 3 Structure comparison of Flash-based FPGA

As mentioned above, other companies' Flash-based FPGAs use Flash for configuration memory, but SRAM for logic cell/wiring storage elements. In comparison, Microchip's FPGA uses Flash as the storage element for logic cells/wiring, eliminating the need for memory for configuration.

 

As I mentioned earlier, there is a large difference in power consumption between SRAM cells and Flash cells. Even with the same Flash-based FPGA, Microchip's FPGA can achieve lower power consumption.

 

In addition, since no configuration is required, you get the following benefits:

  • fast boot
  • Reduction of inrush current
  • Reduced power consumption during configuration

 

The figure below shows the rise time and power consumption after power-on for SRAM base and Flash base.

Fig. 4 Comparison of rise time and power consumption after power-on
Fig. 4 Comparison of rise time and power consumption after power-on

Compared to SRAM-based systems, Flash-based systems can reduce the inrush current that occurs when the power is turned on, and because they do not require configuration, they can start up faster.

 

As for 2., it's very simple.

2. Transceiver optimization

Other companies' FPGAs use Transceivers designed for High-end products also for Midrange products, so Transceivers are designed to support high-speed frequency bands. In other words, a large amount of power is consumed even in the frequency band used by Midrange products.

 

In comparison, Microchip's FPGA has a Transceiver optimized for the frequency band used in Midrange products. For example, in Microchip's latest mid-range FPGA, PolarFire, the Transceiver is optimized for operation at frequencies up to 12.7 Gbps and achieves low power consumption.

 

Seeing is believing.

Below is a graph comparing power consumption in order to see what we have introduced so far.

Figure 5 Power consumption comparison *Comparison with competitive mid-range FPGA (same scale device)
Figure 5 Power consumption comparison *Comparison with competitive mid-range FPGA (same scale device)

As mentioned above, Microchip's FPGA can reduce power consumption by up to 50% compared to other companies. You can see that Microchip's FPGA has overwhelmingly low static and Transceiver power consumption compared to other companies.

Summary

The reasons for Microchip's FPGA's low power consumption are as follows.

 

1.Low power consumption with Flash-based technology

  • Extremely low static power consumption
  • Reduced inrush current due to no configuration ROM

 

2. Transceiver optimization

 

In addition, low power consumption has the following advantages.

  • Solving thermal design problems. Demonstrates its power in heat-sensitive environments and in small sealed enclosures!
  • BOM cost reduction for heatsinks, fans, etc.
  • BOM cost reduction around power supply by reducing power consumption

 

The above is [why low power consumption].