wirelessZONE Products for the week of September 6, 2010
Linear Technology Says…
LTC5583: 6 GHz Matched Dual Rms Detectors Enable Accurate VSWR Measurements
Linear Technology introduced the LTC5583, a 40MHz to 6GHz dual channel, matched RMS power detector, offering over 55dB isolation at 2.14GHz. In RF power amplifier (PA) applications, the LTC5583 provides a simple solution for accurately measuring forward power, reverse power, and voltage standing wave ratio (VSWR). The device comprises a pair of 60dB dynamic range RMS detectors that are matched to 1.25dB. This provides accurate RF power measurement of high crest-factor signals such as those used in LTE, WiMAX, W-CDMA, TD-SCDMA and CDMA2000 3G or 4G basestations and other high-performance radios employing complex modulation waveforms. Each channel can detect signals accurately from as small as -58dBm to 2dBm, in a log-linear response with a typical linearity of better than +/-0.5dB covering all cellular frequency bands. At higher frequencies, the device is capable of providing 47dB of useful dynamic range up to 6GHz. Unique to the LTC5583, each detector simultaneously tracks the envelope of the modulated input waveform, providing on-chip capability to measure both peak and average signal power.
Analog Devices broke the back of the problems in measuring power transmitted with modulation schemes that only digital designers could have come up with. The essential measurement of rms power is the single most important characteristic in determining the efficiency of output PA stages, and controlling them for linearity and radiation effectiveness.
Linear ran with the technology in the form of the 6 GHz LTC5581, which won a Product of the Year Award from us in 2008. The 10 GHz LTC5582 closely followed – but with a dramatic increase in power consumption (from 1.4 mA to about 40 mA) – and we now have what is apparently a dual version in the form of the LTC5583. Power dissipaton over the LTC5582 has, expectedly, doubled (with a nominal 3.3 V rail) and the package has gone from a DFN-10 to a QFN-24. There is also a trimming back of the frequency range to 6 GHz.
The single-ended input capability (up to 2.4 GHz) remains a solidly-useful feature of the Linear parts and is used to contrast particularly, we believe, with the Analog Devices 60 dB range AD8364 (a dual AD8362) which dates from 2005 and is frequency-limited to 2.7 GHz. It is not completely accurate that the AD8364 can only be used with differential inputs but the measurement capabilities using single-ended feeds are limited. The AD9364 costs about $2 less than the LTC5583 but it has one severe limitation: the channel-to-channel isolation is only about 30 dB, compared to 40 dB for the new Linear part.
The LT5583 provides linear outputs from both channels and also has a difference amplifier that is cross-fed from both channels which, when connected as forward and reverse from directional coupler(s) or RF sniffers, will give a representation of VSWR.
This emphasis by Linear on VSWR measurement capability is strange for the primary target market of basestations. Yes, VSWR performance has become more important now that the radio transmit/receive heads are being mounted on top of the towers and there is, therefore, far less feeder to protect the power amplifiers from the effects of reflections. But, in the vast majority of fixed installations, all that can be done with VSWR monitoring is to set off an alarm for an out-of-limits situation and, at the extreme, to close down power before damage is done to the PAs. The data sheet block diagram of the feeds to the difference amplifier have been rather simplified and there is no graph, that I could find, showing actual VSWR numbers to track for the input RF signal powers.
It is more likely that the dual channels of the LTC5583 will be used as input and output power monitoring of basestation RF channels for the control of gain and predistortion linearity corrections. There may be uses for the difference amplifier functionality in instrumentation applications but the data sheet will need to be further supplemented. It is also a shame that there appears to be no temperature detection provided for further linearity corrections to the detectors although the news release states “the device has on-chip provisions for first- and second-order temperature compensation which enable easy calibration for improved temperature performance.” This is not followed through with any detail anywhere that I could see.
The LTC5583 is in production in a thermally-enhanced QFN-24 and is priced at $9.50 in 1000-piece lots.