Wireless Portable USB HUB - Project 13 Report | ECE 445, Study Guides, Projects, Research of Electrical and Electronics Engineering

Download Wireless Portable USB HUB - Project 13 Report | ECE 445 and more Study Guides, Projects, Research Electrical and Electronics Engineering in PDF only on Docsity!

WIRELESS PORTABLE USB HUB

By MATHEWS ANTONY CHARLES HSIEH STEVE LI ECE 445, SENIOR DESIGN PROJECT SPRING 2007 TA: DWAYNE HAGERMAN MAY 1, 2007 PROJECT NO.

ABSTRACT

The project's goal is to make a consumer-friendly device that would convert standard wired USB devices into wireless USB devices, thereby providing customers with an affordable alternative to costly wireless replacements. Our device consists of a PIC, a VFIR transceiver, a USB Peripheral Controller and a power controller module. The PIC is used to configure the VFIR transceiver and the USB Peripheral Controller. The USB Peripheral Controller is used to convert the differential USB data into single-ended serial data via its Serial Interface Engine (SIE), which the VFIR transceiver then transmits/receives wirelessly. The Power Controller is used to supply the entire system with power, while also monitoring and counteracting over current conditions. To test the Power Controller we artificially caused an over-current condition in which the PIC must turn off the correct port on the Power Controller. Testing the transceiver consisted of making sure the correct programming sequence was loaded by the PIC. ii

[This page is intentionally left blank] iv

1. INTRODUCTION....................................................................................................................

1.1 Purpose...............................................................................................................................

According to USB.org there are currently over 2 billion wired legacy USB devices in the world today. This number is still growing as USB proves to be an effective means of adding peripherals to one’s personal computer. From the common USB mouse and flash drive to the USB heated slippers, the ubiquitous technology has found its way to every part of our lives. With the increase of wired USB devices however there is also an increase of clutter and safety hazards that are inherent to dangling cables. Each device is constrained to its plug and does not allow the devices to have much freedom of movement. With the new Certified Wireless USB standard still young and mostly unused, we propose a technology that will smoothly transform the wired USB world into a wireless one. In the process of designing this device, we will need to apply our theoretical knowledge as well as technical experience of electrical engineering in specific areas such as digital signal processing, wireless communications, circuit design, power management, and PIC programming. This document is not simply a record of our product at the end of the allotted design and build period it is also a record of our research into the various means of wireless communication to further the development of a product with important end-user specifications in mind.

1.2 Specifications......................................................................................................................

Our initial goal is to build a system for which existing USB devices can be adapted with wireless functionality. Along with wireless functionality, we tried to make our USB adapter small enough so it is practical as a portable device. Documented below is a list of our design specification:

1. Powered by one 9V PP3 battery
2. Battery life: 2 hours @ 500mA
3. Size: a) Three printed circuit boards during design and testing process: i. Power Controller PCB: 1.78 x 1.28 inches ii. VFIR PCB: 3.35 x 1.58 inches iii. Peripheral Controller PCB: 1.78 x 1.28 inches b) Final Product measures no more than 2.24 x 0.67 x 1.0 inches (about twice the volume of a standard 9V battery).
4. Weight: 70g
5. Compatible with USB2.0 Full Speed specifications a) Transfer rate for each device should achieve 16Mb/s b) The maximum amount of current for each device is limited at 500mA c) The device should be able to support remote wake-up.
6. Total device cost should not exceed $

1.3.4 Power Module......................................................................................................

The power controller comprises of a 5V regulator, an N-channel MOSFET, and a TPS USB Hub Power Controller. The Power Controller must convert a 6VDV~9VDC input to 5VDC outputs on several different ports with a current limit of 500mA. Moreover it must have one 3.3V output to supply circuit components such as the USB Peripheral Controller. In addition to supplying power, it has an over-current function that is capable of noticing when the controller is draining too much current.

1.3.5 Mechanical...........................................................................................................

The mechanical components of the device consist of three printed circuit board adapters (PCB) and two protoboards. The three PCB adapters allow for ease of trouble-shooting and testing before all of the components are integrated into one final PCB before potential production of the device. The three PCBs are for 1) the Vishay TFDU8108 VFIR Module 2) the Maxim 3420E USB Peripheral Controller and 3) the TPS20701 USB Hub power controller.

2. DESIGN PROCEDURE AND DETAILS................................................................................

2.1 VFIR Module......................................................................................................................

The wireless component of our design was meant to transmit and receive the transformed differential USB signal. Initially, we had hoped to employ an approach where the signal would be transmitted at very high data rates, with a significant immunity to signal corruption. In our research, we came across a technology implemented by Motorola’s Freescale Semiconductor Division, titled “Ultra Wide Band (UWB).” The technology utilized the unlicensed 3.1 to 10.6 GHz frequencies of the communications spectrum [1]. Of special interest was UWB’s ability to spread the transmission power over a sufficiently wide bandwidth to make the signal appear as noise to a narrowband receiver, while still being able to transmit very high data rates over short distances (Fig2.1.1). With these characteristics, UWB seemed the perfect solution for our implementation. Fig 2.1.1 Comparison of UWB’s spectrum allocation with that of other communications systems [2] Upon contacting Freescale and a couple other manufacturers -- notably NEC and Wisair – we were told that the technology was still in the experimental phase and that development kits would cost in excess of $10,000. This forced us to seek out alternatives. We next came across the 802.11 standard, which promised a max throughput of 11Mbps using the “b” standard and 54Mbps with the “a” standard. The range of the device was stated to be along the lines of 25-30 m [3]. 802.11 would provide us with the necessary throughput and range, but the amount of power consumed was too much as shown in Equation 2.1.1. Uplink Power Consumption = 0.6 x Transmit + 0.2 x Listen + 0.2 Receive + 0 x Sleep [4] (2.1.1) = 0.6(2 W) + 0.2(0.8 W) + 0.2(0.9 W) = 1.54 W Since our design called for the entire system to be powered via 9 V battery, the 802.11 standard was not a very viable option. There was also the added problem of interference that can be incurred from devices such as microwave ovens, cordless phones, and Bluetooth peripherals which also utilize the 2.4GHz frequency band [5]. Our search eventually brought us to Infrared and the IrDA Standard. We had initially dismissed infrared, thinking it would not provide us with the necessary throughput. Upon further research, however, we discovered that the IrDA standard is a low power alternative to RF communications which is capable of providing a data rate of 16 Mbps, exceeding our 12 Mbps goal. The disadvantages of this

idea consisted of using an op-amp network where the differential signal would be converted to a single ended signal. The downside to this was that we would be negating the benefit of differential signaling— namely effective noise cancellation. Upon further research we came across 3 methods of effective signal conversion: USB to I2C, USB to RS-232, and USB to SPI. Each of these methods provide a single ended signal without the reduction in signal integrity one would experience using the op-amp network. The I2C and RS-232 interfaces had the advantage of hardware flow control, wherein the sender receives some type of feedback from the receiver; this helps in preventing the sender from transmitting data at rates that are too high for the receiver to process. The SPI interface, on the other hand, lacked a method of flow control which in some cases might cause the sender to transmit even though there is no receiver, effectively consuming more power than is necessary. The benefit of SPI, however, lay in its simple architecture and high throughput (13 Mbps), neither of which the RS- (115.2 Kbps) nor I2C (200 Kbps) could match. Since throughput was a major factor in our design, we chose SPI. The lack of flow control, as it turns out, only poses problems when implementing a point to multi-point communications system, since ours was of the point-point variety, we were able to disregard this limitation. We eventually decided on the MAXIM 3420E USB peripheral controller with SPI. The main reason for choosing this specific model was its Serial Interface Engine. The SIE was a very efficient component of the peripheral controller which aided in converting the USB signal into data bytes which could be used by external applications (or in our case the IrDA transceiver). The SIE also had the additional advantage of being a low level error checker and bus retry manager. The peripheral controller itself was a 32 pin device, but we only needed 12 pins. They were: pins 10 and 11 (MISO and MOSI), which regulate the data transmission between master and slave, pins 15 and 16 (D- and D+) which are the USB differential signals, pins 14, 17, and 18 (power pins), pins 19 and 20 (which were connected to a 12 MHz oscillator, and pins 7 and 8 (Reset’ and Slave Select’) each of which was tied to +5V and GND respectively. The completed circuit diagram can be seen in Fig A.3. Much of the passive component values were chosen so as to reflect values provided in reference circuits.

2.3 PIC Microcontroller............................................................................................................

The microcontroller module was responsible for selecting the multiplexers’ outputs, and providing both the clock signal and programming sequence for the peripheral and transceiver interfaces. Due to the basic duties of the microcontroller, a standard PIC18F877 proved to be sufficient for our task. The duties of the microcontroller include programming the VFIR module and the USB Peripheral Controller, switching between programming mode and data transmission mode, and monitoring and regulating power consumption. Fig A.1 is a flowchart of the PIC program’s functionality. The actual code is attached underneath Fig A..

2.4 Power Module.....................................................................................................................

The power system module was responsible for providing a stable source of power to the other components while monitoring the amount of current drawn. If more than 500mA was drawn, the system must cutoff power to prevent damage to the USB device. Since USB devices operate off a 5V source we had to either find a 5V battery or step down a 9V. Seeing as how there aren’t any 5V batteries, we chose the latter. In order to step-down the 9V to 5V we used the MC7805ACT voltage regulator from Analog Devices. The regulator employs a feedback loop to maintain its output voltage at 5Vdc. With a 3Ω load resistance, the power supply circuit was able to supply 4.99V with an almost negligible voltage ripple of 92mV. Not only did it output the necessary 5V, but it also limited the initial current to 1 A.

In order to limit the drawn current further and provide some type of over-current protection, we decided to use the TPS2071 Four Port USB Hub Power Controller from Texas Instruments. The power controller itself has four 500mA N-Channel MOSFET current limiters each having a 5Vdc output, and one 100mA N-Channel MOSFET current limiter having a 3.3Vdc output. When the system tries to draw more than the allotted 500mA, the over current flag in the power controller will be switched on and pin 22 displays a low voltage value of 0V. During normal operation pin 22 displays a high voltage signal of 1.51V to 3.3V. The completed power controller module consisted of a 9V battery, the MC7805ACT step-down voltage regulator, a 0.06Ω N-Channel MOSFET, and the TPS2071 Four Port USB Hub Power Controller. Together, they provided the stable 5V required for the PIC, the wireless transceiver, peripheral controller and the connected USB device. Refer to Fig B.1 in the Appendix for the complete Power System schematic.

2.5 Overall Design....................................................................................................................

Please refer to Fig A.2 in the Appendix for the overall design’s block diagram and Fig A.3 for the overall schematic. Fig A.4 is a scaled packaging layout for the end product; the extra space is left for passive components such as surface mount resistor and capacitors.

3. DESIGN VERIFICATION AND TEST..............................................................................…

3.1 Power Controller and its interface with PIC

After making sure we did not make any wrong connections in the VFIR module’s supporting circuitry we began testing conditions on the datasheet [cite datasheet] that could lead to an error in programming and thus inability to activate the module.

1. Setup time and hold time: The datasheet specified that the setup time and hold time (fig: from datasheet) must be at least 10ns without specifying a maximum setup and hold time. In order to make ensure the minimum time limits we had programmed the module with a 10us setup time and a 100us hold time. The output waveform can be verified with an oscilloscope ( Fig C.1 : scope output for TS and TH)
2. Correct Input sequence and acknowledgement from module: Next we checked whether the input sequence we gave the VFIR module was recognized by the module at all. This would tell us that our input sequence was at least in the correct format. ( Fig C.2 : Tx vs Rx with Ack)
3. Reading from data programmed into registers on the module: The last software test we performed on the VFIR module is to try to read from the configuration storage registers that we had tried to write to. This test would tell us whether or not the registers are correctly storing our configuration values. ( Fig C.3 : Tx vs Rx with Read) 3.2b Hardware testing of the TFDU8108 VFIR Module Through the datasheet and through contacting the Vishay technical support engineers we have gathered the following requirements to check against whether or not he VFIR module is function correctly. This was done so in two major parts. The first part was to make sure all of the external supporting circuitry of the module was connect correctly. The second part was to make sure none of the internal connections of the module had been damaged. We measured the pin voltages on the VFIR module to make sure the external support circuitry was wired properly. It was noted to us that pins 1, 4, 6 and 7 should be Vcc, pin 2 should be floating, and all of the remaining pins should be ground. (see Table C.1 : Pin voltages) Next we measure the diode voltages from each pin. Technical support as Vishay made clear to us that there are diodes from each pin except pin 1 and 2 to the ground pin. There is an expected diode voltage of 0.5V from pins 3 through 7 to ground (pin 8). Pin 1 should have a diode voltage of 1.4V to Pin 2. If the expected diode voltage was zero for a pin then that pin’s internal connection has been damaged. (see Table C.2 : Diode voltages) 3.3 Conclusions

The test results for the Power Controller had been successful. The device was able to achieve a current output of about 400mA and have a relatively stable 5V DC output. Furthermore it is able to successfully monitor and report overcurrent. Using the PIC and the Power Controller’s over-current output signal we were able to successfully shut off the port that was drawing too much current without shutting off the rest of the circuit. Through our testing of the VFIR module we are able to conclude that the problem we had been experiencing was not due to a programming error. Through further hardware testing we were able to see that the VFIR module itself was not damaged nor was it supporting circuitry connected incorrectly. Yet with those concerns dispelled the module was still unable to turn on. Furthermore we have found several discrepancies between the programming sequence Vishay suggested in the product’s appendix and our implementation. For example, to Enable All, Main Control Register must be set to 0b11101000. The suggested Main Register value is 0b11110000. However another section of the datasheet implied the main register’s value be 0b11101xxx [9]. Furthermore there was little detail about how to read from registers and the RxD line does not echo the TxD line in SD mode. Since the datasheet’s accuracy is now in question, we have decided to look for other manufacturers of VFIR modules with more solid technical support and documentation. Since we were unsuccessful in sending and receiving data with the VFIR module performing tests on the USB Peripheral Controller did not seem practical for the amount of time we had remaining. The remaining time had been used for redesigning our device for better performance and searching for alternate manufacturers of parts such as the VFIR module.

4. COST

We were successful in programming the VFIR transceiver through its SPI interface by bit- banging the output pins of our PIC microcontroller with the data bits and the clocking signal. Although it was unfortunately our transceiver did not function properly to allow the further testing of USB connectivity we were able to show that our design can in fact circumvent processing of USB signals for wireless transmission. Furthermore we were able to demonstrate the power monitoring and regulating functions of the USB Power Controller and PIC Microcontroller combination. Had the VFIR module worked, we were uncertain whether or not it would be able to connect successfully to a USB device and communicate its presence to the USB host dongle (computer-side dongle). Further research showed us that the peripheral controller alone would not be sufficient to construct a host-slave relationship between the USB device and our wireless adapter dongle. There need to be a USB Bridge controller on the dongle’s side that will establish the USB device as the host and the dongle as the slave. Otherwise USB handshake cannot happen and the USB device and the dongle will not be connected. To conserve power we can add IR Encoders to the input data stream (fig: improved block diagram) which would limit each signal to a pulse with width of no more than 100μs. In addition to s. In addition to conserving power, this add-on would make it less likely for the IrED to burn out resulting in increased product life-span. (see Fig D.1 for Revised Block Diagram) Since the USB host takes care of the bus timing, the resulting network with potentially many of these dongles would effectively be Time-Division-Multi-Access (TDMA) for multi-point to point communications. However, since the transmission mode is IR, we would experience some common limitations of IR transmission. Dongles not in direct line of sight of the host will either have poor connectivity or no connectivity. Dongles at different distances from the host will take different amount of time for signal to reach the host. Transmission errors can be detected and corrected to some extent by coding the transmission sequence. However, this would mean direct interaction with the data stream by breaking up the data stream into packets adding parity bits to each packet. That would imply the use of an expensive and more power-hungry DSP. Transmission lag will not affect this product. A rough analysis shows that for a significant difference in the data stream’s timing to be caused by transmission lag, the difference in distance between the nearest peripheral device and the furthest peripheral device must be in the order of hundreds of feet. That is, since light travels approximately 1ft per nanosecond, and USB transmission is on the order of microseconds, a difference of one tenth of a microsecond would mean hundreds of nanoseconds which means a distance difference of hundreds of feet. However since the effective range of our IR transceiver is rated for 1 meter, which is about 3 feet, no transmission at hundreds of feet through this device can be detected. On the other hand, transmission using IR has its advantages too. The transmission of data would not interfere with radio-frequency signals used by common house-hold consumer electronics such as wireless access points, cordless and mobile phones, or Bluetooth devices. In fact this takes away some fairly crippling design overheads such as allocating bandwidth inside the crowded RF band or worrying about EMI. The small package, low power consumption and simple design of the IR modules make IR ideal for a device aimed at saving existing USB devices and competing against future Certified Wireless USB devices. Therefore our product would ideally reduce the pollution associated with older generation USB devices becoming obsolete with the future dominance of Certified Wireless USB products. Our product is designed as a cost-effective alternative to buying new devices and throwing away the replaced devices. Also since this device does not operate in the RF band it will not cause EMI for other house- hold electronics devices the consumer might have. Another ethical consideration is eye protection. Since the human cannot detect IR, there is no blink reflex to shut out IR signals that are too strong. We have limited the transmission power of the IR transceiver to be 250mA at 5V (1.25W). However this power is pulsed at intervals of less than 100us there should be no eye damage so long as the user does not stare at the transceiver in any prolonged period of time REFERENCES

[1]Reed, Jeffrey. Buehrer, Michael. Ha, Dong. “Introduction to UWB: Impulse Radio for Radar and Wireless Communications” Mobile and Portable Radio Research Group at Virginia Polytechnic Institute and State University [2] UWB Radio Systems Committee, “About UWB Radio Systems” pp 1-3 April 28, 2004 http://www.soumu.go.jp/joho_tsusin/eng/Releases/Telecommunications/040324.pdf [3]Wikipedia, “IEEE 802.11” http://en.wikipedia.org/wiki/802. [4] Atheros Communications, “Power Consumption and Energy Efficiency Comparisons of WLAN Products” 2003 p http://www.atheros.com/pt/whitepapers/atheros_power_whitepaper.pdf [5] P. Fuxjäger, D. Valerio, F. Ricciato "The Myth of Non-Overlapping Channels: Interference measurements in 802.11" pp 1- The Fourth Annual Conference on Wireless On demand Network Systems and Services, Obergurgl, Austria, January 23-26, 2007 [6] Barna, Paul and Schlanger, Steve. “Fundamentals of the Infrared Physical Layer AN243” Microchip 2004 [7] Freescale Semiconductor, “SPI Block Guide v3.06” Motorola Inc February 4th, 2003 [8] Dallas Semiconductor, “AN3891: Isolating USB” August 8, 2006 [9] Vishay Americas Semiconductor Technical Staff, Very Fast Infrared Transceiver Module , Vishay Semiconductor, January 2006 [10] Maxim Integrated Products Technical Staff, USB Peripheral Controller with SPI Interface , Maxim Integrated Products, December 2006 [11] Texas Instruments Technical Staff, Quadruple 2-Line to 1-Line Data Selector/Multiplexer , Texas Instruments, 1993 [12] Texas Instruments Technical Staff, TPS2070, TPS2071 Four Port USB Hub Power Controllers , Texas Instruments Febuary 2001 [13] Analog Devices Technical Staff, 1.0 A Positive Voltage Regulators , Analog Devices, April 2002 [14] International Rectifier Technical Staff, HEXFET Power MOSFET , International Rectifier, May 1998 [15] Techlib.com, “Battery Capacity”, http://www.techlib.com/reference/batteries.html [16] Cisco, “Channel Deployment Issues for 2.4-GHz 802.11 WLANs”, 2004, http://www.cisco.com/en/US/products/hw/wireless/ps430/ prod_technical_reference09186a00802846a2.html [17] CRISP, “Electromagnetic Radiation”, 2001 http://www.crisp.nus.edu.sg/~research/tutorial/em.htm