Cellular phone handsets are required to set transmit power to within a specified precision. There are two predominant techniques. The first is the in factory calibration performed when the handset is being manufactured. In calibration, the handset is measured to ascertain the output power under various circumstances, and a table of the results is generated and stored within the handset. This table is used to set the power per the direction of the system. The accuracy of the power setting is then determined by how thoroughly this calibration is accomplished. This technique is not capable of responding to changes in the performance of the handset.
The second technique is sample and detect. The power out of the transmit portion is sampled and detected. The second technique requires a coupler, detector, and signal processing to measure the detected voltage as will be further described. This requires that a form of calibration be performed, but the detection circuit will accurately reflect any subsequent changes in the performance of the handset.
Using either prior art coupler, the detected power is then delivered to a detector diode. The diode rectifies the power and generates a DC level. This DC level is processed according to the system needs. The detected value is used to adjust the power level as required.
The process technology used to implement the coupler sets the minimum separation between the through conductor, e.g. first transmission line, and the coupled conductor, e.g. second transmission line. This minimum separation determines the minimum length to achieve the desired coupling. To illustrate, driving a diode directly requires about 15 dBm at 1 to 2 GHz, the range of interest for handsets. If the amplifier is transmitting 1 W (30 dBm), then the coupler must provide 15 dB of coupling. This requirement sets the minimum length of the coupler in any particular process technology.
There are two loss mechanisms in a coupler. The first is the ideal loss associated with the coupled power. This power leaves the through path and enters the coupled path. When half the power is coupled in a 3 dB, the through loss is at least 3 dB. In a 15 dB coupler, the through loss is at least 0.14 dB.
The second loss mechanism is resistive. The metals and dielectrics used in a coupler are inherently lossy. Consequently, the longer the through transmission line is the higher the loss.
Couplers are available in many form factors. The largest are instrument grade, made of machined metal, operable over many octaves. The smallest are built on ceramic, covering perhaps one octave usefully, e.g. small ceramic AVX 15 dB coupler having 0.35 dB loss at 2 GHz. To implement the detector function, the circuit includes the ceramic coupler, external diodes, a biasing network for the diodes, bypass capacitors, and terminating resistors, if needed. The resulting network is large and unwieldy.
The present invention is a coupler and detector integrated on a semiconductor substrate, e.g. gallium arsenide or silicon. Semiconductor processing allows for small trace and space rules. The tighter design rules provide for tighter coupling than can be achieved by ceramic processes. The greater coupling allows for a shorter through line and with less loss, thus closer to ideal coupling. The semiconductor substrate supports the addition of whatever supporting components are required to complete the detecting function, such as diodes, transistors, resistors, capacitors and interconnections.
The present invention is a coupler and detector integrated on a semiconductor substrate, e.g. GaAs. Semiconductor processing allows for small trace and space rules on the order of less than 3 μm horizontal and less than 1 μm vertical. The tighter design rules provide for tighter coupling than can be achieved by ceramic processes. The greater coupling allows for a shorter through line and with less loss, thus closer to ideal coupling.
The entire circuitry for detecting power may be fabricated on the same die. This provides two benefits. First, it greatly reduces the size of the detection function. Second, it supplies a new design regime wherein coupler loss can be traded off with bias current to increase the overall efficiency of the handset.
As an example, to provide 1 W (30 dBm) from a 50% efficient power amplifier, 571 mA from a 3.5 V supply is required when there is no coupler. If the 15 dB coupler has 0.35 dB of loss, the amplifier must deliver 30.35 dBm, at the cost of 619 mA. Thus, the coupler requires an additional consumption of 48 mA. Because one can integrate the coupler and detector, the loss in the coupler can be reduced while the detected output can be maintained. For instance, if the loss is reduced to 0.15 dB, resulting in a coupling of 25 dB, one can use a 10 dB amplifier to bring the equivalent coupling back to 15 dB. The power amplifier is now required to provide 30.15 dBm, and so requires 591 mA. This amplification would require perhaps 3 mA, substantially less than the 28 mA difference between 619 mA and 591 mA.
The power detection function is made significantly smaller and more efficient by using an active semiconductor substrate, e.g. GaAs. This substrate can contain the coupler, the detector diodes, the required passive devices for biasing and bypassing, and transistors for amplification.
In operation, the linear amplifier 20 amplifies the output signal of the coupler allowing for a coupler with less coupling, and thus less loss.
In operation, the charge pump 22 increases the voltage at node A. This compensates for the possibly lower coupling of an integrated coupler.
Number | Name | Date | Kind |
---|---|---|---|
3500255 | Ho et al. | Mar 1970 | A |
4789887 | Crossley et al. | Dec 1988 | A |
5001399 | Layden | Mar 1991 | A |
5036229 | Tran | Jul 1991 | A |
5105171 | Wen et al. | Apr 1992 | A |
5313175 | Bahl et al. | May 1994 | A |
5378939 | Marsland et al. | Jan 1995 | A |
5508630 | Klemer et al. | Apr 1996 | A |
5658132 | Akazawa et al. | Aug 1997 | A |
5786992 | Vinciarelli et al. | Jul 1998 | A |
5832374 | Birth et al. | Nov 1998 | A |
5960333 | Repke et al. | Sep 1999 | A |
6002375 | Corman et al. | Dec 1999 | A |
6542375 | Kuitenbrouwer et al. | Apr 2003 | B1 |
7034633 | Passiopoulos et al. | Apr 2006 | B2 |
20050073373 | Dupont et al. | Apr 2005 | A1 |
Number | Date | Country |
---|---|---|
0364879 | Apr 1990 | EP |
0511728 | Apr 1992 | EP |
1 521 363 | Apr 2005 | EP |
2003-324326 | Nov 2003 | JP |
Number | Date | Country | |
---|---|---|---|
20050231302 A1 | Oct 2005 | US |