12 June 2012
Design challenges of implementing USB into next generation portable communications hardware
Universal serial bus (USB) interfaces are becoming ever more prevalent in the design of portable communication products, including mobile phones. These interfaces not only offer a means of importing and exporting multimedia data, such as picture or music files, but also present a convenient way of carrying out the charging process. As a result, USB is now being widely tipped to become the ubiquitous interconnect within this sector.
However, there are certain issues that still need to be resolved and highly integrated semiconductor solutions will be needed to help to make this transition happen.
Engineering teams charged with designing portable communications products face a range of challenges, including:
Consumer expectation. This dictates that portable communications products must have sleek, lightweight form factors. Thus, the board real estate that engineers have available to them has a high premium associated with it. Any opportunity to save space should be fully exploited. This means that a USB connectivity solution must have the ability to be applied successfully into space constrained environments.
Power consumption. As portable products are battery powered, their current consumption (both when operational and idle) is another key concern, with reductions in this area being advantageous. It is therefore important that adding the desired USB connectivity does not cause some of the portable product's precious power budget to be sacrificed.
Shelf life. Cell phones and other handheld communication equipment designs tend to have very short shelf lives. As a result, a platform approach is often taken to maximise the design effort expended by using it over multiple products in a family.
Current levels. With the USB interconnect increasingly becoming the primary method for powering portable electronics equipment, not just computer peripherals, the current levels supported by the standard have now been raised.
Previously, USB could only deal with 500mA through a standard downstream port (SDP). Now, in order to keep recharge times for portable products as short as possible (in line with what consumers are already used to through traditional charging methods), a new version of the USB charging specification has been introduced which permits currents of up to 1.8A to be transferred.
This is done through a dedicated charging port (DCP), where both the data lines have been shorted. It is vital for portable products to differentiate between DCPs and SDPs if they are to benefit from a DCP's increased charging capacity. So far, USB charging circuits with DCP detection functionality have been relatively complicated, relying on a large number of discrete components.
They also need the system microcontroller to allocate some of its processing power to the task of detection. This means that it is no longer focused fully on its main function and impinges on overall system performance. In addition to being inefficient, this approach calls for many hours of tedious code creation and circuit design implementation.
More sophisticated controllers
The development of a new breed of more sophisticated USB controller chips offers OEMs within the portable communications sector the opportunity to take advantage of the enhanced performance levels and broader feature sets needed to address the market demands placed upon them. Such devices can help to alleviate the various board level challenges discussed earlier, as well as retaining the opportunity to add advanced USB battery charging to their products in the future without system performance being impinged upon or engineering resources being overstretched.
The devices in FTDI's recently released X-Chip series integrate functionality for detecting automatically whether the portable device is connected to a DCP or an SDP, so that system logic can be switched from data transfer to charging mode. Once a DCP has been detected, the device asserts a signal on one of its output pins to indicate this to the portable product. The discrete component count required for this task is significantly lower and the system mcu does not need to get involved; instread, it can concentrate on executing its core activities. As a result no additional software or drivers need to be created. Figure 1 shows the X-Chip's basic functional blocks (including protocol handling, voltage regulation, memory and clock sections), while figure 2 compares the conventional DCP detection circuit with the more spartan approach that is possible using one of these devices.
As already mentioned, one of the key objectives set for development of next generation USB controllers is to lower the power consumed. By migrating to a smaller process geometry and having a core that runs at 1.8V, rather than 3.3V, X-Chip devices draw less than 8mA when fully active (between 30% and 45% less than many competing devices currently available) and less than 125µA while placed in suspend mode. A built-in clocking mechanism helps to save further board space, as well as reducing procurement effort, since fewer components have to be sourced. Furthermore, by following an integrated PLL approach, rather than having a discrete oscillator, improvements to system reliability are also realised.
The platform strategy favoured in the portable consumer sector is assisted by having a multitime programmable (MTP) non volatile memory incorporated into each of these devices. This allows storage and configuration of descriptors covering the product (such as model type, feature set, production date and production site). Engineering teams can alter descriptors as needed, so that they can spawn different product offerings from one original stem.
The main factor dictating the footprint of a USB controller is not the size of the semiconductor die, but the number of I/O pins that must be supported. As, in this case, the memory and clock are both on chip, the need for eeprom and clock pins is negated. The pincount is lowered further by getting rid of some of the handshaking pins. By doing this, X-Chip devices can be supplied in more compact packages than existing USB controllers on the market, so that less pcb space is taken up. The smaller dimensions of the package do not have a detrimental effect on thermal management however, as any reduction in pad size is more than compensated for by the lower core voltage used.
Specifying highly integrated USB interface solutions which have timing, memory and DCP detection functionality on the chip will greatly benefit the portable communications sector. More sophisticated controller devices will lead to better use of both hardware and software design elements, as well as keeping power consumption in check, saving space and reducing the overall bill of materials. They will allow advanced battery charging to be initiated without there being a need to redirect some of the MCU's processing capacity, as well as supporting the development of multiple products from a single platform.
Gordon Lunn is a senior application engineer, Graham Brown is an applications engineer. Both are with FTDI (www.ftdichip.com).
Gordon Lunn and Graham Brown