The solution
In the late 1980s mobile telephony was an allocation problem: give each caller a frequency or a time slot, and guard bands between them. Qualcomm, then mostly a satellite-messaging company, argued for the opposite — let every handset transmit on the same frequencies at the same time, separated only by a spreading code, and make power control the load-bearing feature of the whole system.
The physics was against them. Terrestrial path loss follows an inverse fourth-power law: in a cell of five miles, distance can vary a hundredfold and path loss by more than 80 decibels, so an uncontrolled phone near the tower drowns one at the edge. Qualcomm's patent, filed in November 1989, closes the loop — each phone measures the received pilot and adjusts its power inversely, while the cell site measures every arriving signal and commands further corrections toward the minimum level that still communicates.
The reward for a whisper is that everyone gets to whisper. With every transmitter held at its floor, the interference floor stays low enough for all codes to decode, and the same spectrum is reused in every cell. The 1991 capacity paper by Gilhousen, Jacobs, Padovani, Viterbi, Weaver and Wheatley put numbers on the gain, and the scheme became the working core of cellular CDMA.
What it achieved
Every signal arrives equally faint
Why it worked
A CDMA receiver hears every user as noise, so capacity is set by total interference — each user who transmits less literally hands capacity to the others.
Open-loop estimation alone cannot track fading and distance changes, so the tower-side command loop does the fine trimming.
Frequency slots and time slots had been engineered to their limits; code sharing moved the fight from infrastructure to transmitter discipline.
The loop was a hardening of what Qualcomm already ran on satellite links, where path loss varies only 1–2 dB; a terrestrial cell spans over 80 dB.
What can be applied
The fix for a resource ruined by its loudest users is not more lanes: it is keeping every user at the lowest level that works. Capacity came from subtraction — no new spectrum, just restraint.
Aftermath
Granted in October 1991, the loop became the working heart of cellular CDMA. The same year, Gilhousen, Jacobs, Padovani, Viterbi, Weaver and Wheatley published the capacity analysis in IEEE Transactions on Vehicular Technology, and IS-95 — the standard built on this design — was approved in 1993 and went commercial as cdmaOne, with CDMA2000 descending from it.
FOLLOW THE EVIDENCE
The sources
- Method and apparatus for controlling transmission power in a CDMA cellular mobile telephone system (US 5,056,109) freepatentsonline.com
- On the capacity of a cellular CDMA system doi.org
- IS-95 en.wikipedia.org