Torbjorn Johnson, known as T J, is a mobile base station pioneer. Since the 1970s he has been producing revolutionary technologies which have covered people and places. Now he is seeking to revolutionise the whole mobile base station sector through much lower power and cost RRHs, plus high gain passive Phased-Array antennas (PAA) mounted in a very few, but higher and stronger towers. These towers will be future-proofed to cover all band capacity multi-sector antennas, provide 130% green energy and full hardware SuperSiteHotel sharing. If what T J has developed comes to fruition it would be nothing less than a complete revolution in global telecom site performance and cost. It would deliver 10,000+ times higher signal/power efficiency, cutting energy and all costs. TowerXchange speaks with him about his history, the current coverage, capacity and US$/km2 and US$/GB (gigabyte), power Wh/GB and carbon pollution problems facing telecom networks and how he can solve them. Has the mobile industry got it all wrong? Can large passive PAA modular antennas in much fewer, higher, stronger towers make LTE+5G an affordable, 99.5% nationwide profitable reality? T J says yes, Electro-Mechanic-Aerodynamic-Power HW ’Dimensions Matters Enormously’ for antennas and the towers.
TowerXchange: For those who don’t know your history, please introduce yourself:
Torbjorn Johnson, CEO, Radio Evolution Sweden AB:
I am a TV and mobile radio innovation pioneer. After first installing, then designing UHF TV transmitters 40,000 ’ 2 W. I started my first mobile base station company in Stockholm when Ericsson was still only producing switches, fixed line equipment and military electronics.
In 1980 Ericsson didn’t believe in mobile telephony, when I had already developed my first MTD and NMT 450 complete 50W base stations and 2.5m panel antennas. In the beginning, I delivered to Ericsson for several countries and delivered and licensed to Nokia.
In 1988, my first stock listed company was acquired by Ericsson for US$60mn.’ I joined Ericsson’s management for base station technology. Until 1994 I analysed SW-radios, linear-amplifiers, phased-array antennas, energy and cost KPIs, high altitude platforms, satellites etc. ’ all areas of innovation ’ but product management was conservative, so we missed many leadership opportunities.
By 1994 I found that Ericsson and MNOs had neglected the new PAA antenna technologies and the important companion towers. High gain 40 dB vs old 18 dB antennas in few high, but highly efficient towers can multiply coverage area up to 131 times, and site capacities with PAA-lobes in all new mobile bands by over 300 times. This is comparable to the Wavelength Multiplexing innovation for fibre capacity. This is the sort of thing needed for MNOs to become future-proof and much more profitable with +50% GB/Y at lower cost, by sharing all optimised site hardware. Seeing the opportunity, I developed a first generation with 0.1 Watt LEMS low-emission handsets and low power TRXes. We demonstrated multi-sector PAA antennas in several countries and delivered 50,000 handsets with eight hours talking and two weeks of standby AA battery life, for a range of 30 km. This was around 1997 for NMT, but uptake was not significant when GSM was coming on higher bands.
I was also a SiteHotel pioneer 30 years ago with a 45m standard tower at his house along the entrance to Stockholm for all ferry ships, now having 6 MNO+ISP tenants and a RoE/Y of 140%. But only reaching 25 km2. The telecom sector has often faced disruptive change, 15 years ago iPhones changed everything. Steve Jobs wanted to cooperate with Ericsson, but the technical guys didn’t like his smartphone innovations, and thought that there was no market for it. So they turned him down and Steve went alone, and Ericsson and Nokia soon lost their handset market leaderships. The same could happen with expensive high power hungry 3GPP RRHs.
TowerXchange: Why have MNOs underperformed since the introduction of 4G?
Torbjorn Johnson, CEO, Radio Evolution Sweden AB:
The highest performing parts of the mobile industry value chain have become the three US listed old towercos with a market cap of US$17mn per employee.
These are followed by Facebook, Google, Microsoft and Apple at US$12-7mn/empl. Then MNOs rank at US$2-0.05mn/empl. Then come the listed 3GPP RRH suppliers Ericsson and Nokia with US$0.40-0.20mn/empl. facing price competition from China. The towercos with old low-tech towers have gained over 150% in five years but their struggling customers, the majority of MNOs have declined by 13-91% in the last five years! It is a paradox that the towercos, highest in the MCAP/empl. have the least radio network and antenna competence, relative to the MNOs and RRH suppliers.
The costs for MNOs are too high relative to revenues. MNOs today have few towercos to choose between, and can buy LTE equipment from just five vendors. A report from the GSMA and ITU-D Q5/1-308, 4 October 2019, illustrates this ’ with 4bn LTE customers globally on 4mn voice-built towers that equates to 1,000 customers per site, with 1.9 tenants per site delivering from 10 RRHs of 40W/10MHz (with total energy costs of US$25bn/year), and only 9GB/month of data per customer. The average site costs US$34,519 per year/ 1000 customers. It costs MNOs US$2.87 per customer per month using 9 GB of data.
The cost breakdown of the average global LTE tower site into business units is as follows:
BU1: 58% of costs relate to tower space and land - US$20,000
BU2: 1% is for antennas - US$420.
BU3: 6% is for transmission, mostly microwave - US$2,205
BU5: 23% is for energy, sourced from gensets, or grid coal+gas. Mostly from carbon-intensive generation. The cost is a very high for 97 Wh per GB delivered - US$8,054
BU4: 11% is for RRHs + BBUs - US$3,840
This is very high 810% overhead/RRH, unbalanced performance per dollar between high tower space cost for few and low antenna Ael area, and has been very detrimental for most MNO’s market capitalisation.
MNOs with mostly falling ARPUs are looking at OpenRAN, C-RAN, to reduce RRH+BBU costs. They are hoping for a 50% reduction by looking to new vendors and new technology. But RRHs costs are only 11% of their total outlay. The big problem is cost of low coverage and capacity towers and energy.
With voice it was very easy to grow coverage and customers, with towers built for voice by MNOs for 1 or 2 low bands and small 18dBi 3 sector antennas. LTE for full data rates needs 30dB stronger signal vs GSM voice. LTE sites have therefor less coverage areas, limited capacity, low SEaver, less customers per RRH, more expensive capex/opex/ and lower returns. Built as Macrosites for voice, they have become Microbases for LTE+5G.
TowerXchange: What problems do you identify with the present telecom towercos?
Torbjorn Johnson, CEO, Radio Evolution Sweden AB:
I have analysed the US listed towercos. Their market cap is US$231bn and has skyrocketed since 1995. Investors think that 5G, higher bands, data growth et cetera will need more towers. But I do not think that is accurate, or affordable to MNOs and end customers.
Radio Evolution has compiled a ’Global MNO’s Wish List to Super TowerCo in 12 items’ (Doc 7072A).
It solves how MNOs can turn-around from suffering, to highly profitable.
Presently GB costs are far too high, and this is driven by the cost of tower space leasing and low service areas. In Southeast Africa the cost in 2017 of 1GB prepaid mobile data varied from US$2.3 to US$35.3 ’ this is completely unacceptable, and a discrimination of free IP information for the countryside/rural populations.
The world’s 4mn towers were originally built and planned for voice at 960 or 894MHz with -85dBm planning average edge outdoor signal, GSM works indoors at -104dBm into handsets, but LTE/5G would need -74dBm/10 MHz into indoor smartphones, for full Data Rates SE 4.0 Mb/s/MHz. 2170/3800 MHz higher bands to Oku-Hata have 13/22 dB higher losses. And higher DL-BW/Data Rates need still more signal according to Shannon. The Coverage+DR+Capacity+Cost challenges will not be resolved by using same tower heights, removal of 3.5 dB coax feed lines, and slightly improved antenna gains, or Multibands, or M-MIMO 64T64R.
They need to, and can be drastically improved in every site BU, by optimising all electro-mechanical-aerodynamic-power HW dimensions.
TowerXchange: How are you trying to move the industry forward? What is your solution?
Torbjorn Johnson, CEO, Radio Evolution Sweden AB:
By a cascade of optimised RE efficiency innovations for 10,000 to 28,500 times more signal/power.
All hardware dimensions matter enormously to multiply Coverage Area, Data Rates and Capacity, with disruptive energy and cost reductions in all site BU cost centres. This is better than building millions more of traditional low, weak, obsolete towers with small antennas. See ’Important 37 Techno-Economic Tower KPIs for all TowerCos’ (Doc 5006A) on my website.
RE’s webpage ’www.radioevolution.se gives an introduction. It is my history of five unique pioneering R&D companies and techno-economic market optimisations.’ With stepwise of over 25 years of R&D and US$100mn investments and several license takers and several thousand man-years and binders of technical methodical analysis of all hardware.’ Gradually earlier PAA antenna versions with 30dBi in 100/166m standard towers have in five countries vs 18dBi at COR 55m, verified CovF 16-20-30x. Having audited traditional tower 1 page specifications and pushed CTOs of tower manufacturers in 6 countries, and site audits of 100s of imperfect site installations, RE has developed the SW design parametric specs for the SuperEfficientTower family, and the companion SAS Super Antenna Modular plug-in from inside, passive robust antennas, and low loss feeding LSUs.
With strategic partners to implement RE’s 9 Goals is SuperSites where each BU is always best in class, modular and highly profitable for MNOs and the SuperTowerCo. And dynamically reconfigurable 50Y future-proof from initial SuperCovMode1, CovF 131, to gradual infinity to provide over 2,000 GB for US$1 total site costs per customer with the SuperCapMode5.
To do this, sites need to share all site hardware. Plus finally sharing, borrowing, pooling, spectrum for higher peak data rates. And Radio Evolution has systematically developed and optimised down to the last BOM steel tube and bolt dimensions, the universal platform of SuperSiteHotels. By using the cascade of optimised signal/power efficiency innovations, that i.e. reduces the RRH W/10 MHz 20x less, makes Wh/GB 97x less vs 2019, and then energy US$/GB 358x less, plus ZERO CO2 emissions from Y1. ’Summarised below:
Efficiency 1: Increasing tower height to COR 225m vs old 55m. This increases signal efficiency = power 18dBi = 63x
Efficiency 2: Aerodynamic efficiency. SAS antennas mounted as 1 large Diameter cylindrical body (EPAtot =Cd 0.45*H 25m*D10.5m/960 pnls = 0.125m2/pnl), vs old open ’Birds Nest Mounting’ with many additional flat exposed elements with EPAtot = Ael5 m*0,30 m * Cdeff 2.5 = 1.90 m2/pnl. = 15.2x.
Efficiency 3: Increasing Ael area 960*2.5m*0.33m = 792m2 vs 9m2 = 88x
Efficiency 4: Vertical 64 dipoles feeding Robin-Hood 1 degree ideal lobe, vs 8 dipoles std 7,6degr = 7,6x.
Efficiency 5: The SET tower has EPAtot strength 120/22.8m2 = 5,2x.
Efficiency 6: PAA multicolumn for feeding increasing signal 65/3 degr = 22x, and lobes capacity by 75/3 = 25x/Ael. By 10,000Gx*75lobes*8MIMO/768p*2.5m*0.33 m = 9,470x, vs 63*3*8/12p*2.5m*0.30m = 168x = 4x
Efficiency 6: Side/Backlobes reduced 15/22 dB vs narrow std antennas, increasing capacity 1,3x.
Efficiency 7: MIMO3,4,5,6,7,8 efficiency improved by 5x wider COR separations from Dia 10.4m, = 3x
Efficiency 8: Vertical 4-8 Multilobes 35 degr, gain 3X, plus 4-8x vert capacity. Above 1710 MHz.
Efficiency 9: Site Hotel Lifts access. In base+top tower. Enable convenient, weather protected, safe installation of all equipment from day 1, adding capacity modules and service anytime.
Total Signal/Power Efficiency factors in ?Ga+COR dB: ’(40’45)-18rel+18 = 40+ dB, or 63*7.6*22 = >10,500x.
Employed to get CovF 131x at 40W/10MHz same DR, or 66x/10W, or 26x/2W. Wh/GB from 97 to 1 = to 1%.
Each MNO SuperSite Capacity CapF improvement: 25*1.3*1.3 = 42 x. For a given spectrum price,
SuperSite Total capacity, by all DL bands 470 -7,150MHz (+MW) = >1,000 MHz, vs std site limited to typ 200MHz. ’’Total traffic BH >1,200Gb/s (*6 with VML), vs typ 4Gb/s = >300x = Future proof, very low cost.
TowerXchange: What does the 10,000- 28,500x signal/power Efficiency improvements mean for the cost reductions in the SuperSiteHotel 5 BUs?
Torbjorn Johnson, CEO, Radio Evolution Sweden AB:
Dramatic leasing cost reductions by the above efficiencies. The only drastic and future-proof solution:
BU1: SET Tower COR 225 m with height gain 18dBi = 63x.’ And capacity for 960 SAS panels vs std 12.
In LB: 40dBi, Area 3 260 km2. Lease US$72,5000/Y = US$ 22.2/km2, vs std COR55m, 40W, G 18dBi, Area = 24.5km2. Lease US$20,000/Y = US$816/km2.’ = 36x lower/km2.’ Improvement CovF = 131/66/26x at respectively 40/10/2W. Capacity 120 m2 EPAtot. ’Tower leasing/2.5m pnl = US$72,500/960 = US$75/Y, vs US$20,000/12 = US$1,667/Y = 22x less.’ Tower leasing Y8 full capacity US$72,500/1,200,000 Mb/s/12mnth = US$ 0.025/500GB/mnth, = 42x less.BU2: SAS-antennas. With the aerodynamic EFF2 of 15,2*120/22,8m2 EPAtot = 80x more Ael.’ ’’’’’’’’’’’’’’’’’’’’’’’’’’’’’Enabling with the EFF1-9 Dramatic passive Antenna 10,000-28,500x performance multiplications BU3: Transmission. Located along fibre grid locations, US$/Gb/s to GW at 1,000 vs 5 Gb/s, with cap exp 0.30 is typically 20.4x lower, or exp 0.20 is typically 69x lower.
BU5: Green 130% site energy+ups. At 100kW US$0.10/kWh, vs 5 kW, US$ 0.37/kWh = 3.7x lower/kWh. ’’’’’’’’’’’’’’And at 1,0 vs old 97 Wh/GB becomes US$358x/GB lower.
BU1+2+3+5. SuperTowerCo leasings at full capacity becomes:
= US$0.005+0.009+0.001+0.006 = US$0.105/5Mb/s/mnth. = 25.3x less = -96.0% vs standard costs of US$0.214+0.018+0.100+0.200 = US$2.660/5Mb/s/mnth.
Adding BU4: LP-WB-SuperRRHs+LSUs+PAs+BBUs, at 2W vs 40W/10 MHz, US$60/300/Y = 5x less.
Reducing COR225m height or antenna panels qty and gain, to reduce leases, actually multiplies US$/km2 and US$/GB. Easily found by RE’s systematic performance/cost design analysis. TowerXchange: How do you solve the present high Wh/GB Power and CO2 problem, with 111x more traffic <10Y, and bring CO2 emissions to NIL?
Torbjorn Johnson, CEO, Radio Evolution Sweden AB:
Diesel driven sites burn approximately 1.0 kg CO2/kWh. This means one site if drawing (97 Wh/GB) 1.2 kW and 80% is fossil based, emits 8.4 ton/Y for 1,000 cust *9 GB/month = 0.93 kg CO2/GB, and from 4 M sites 2019, 33.6 MtonCO2/Y, costing US$25bn/year, US$0.59/kWhaver.
The same type of sites for 8bn customers using 500GB/month and 111x more traffic than today would result in CO2 emissions of 3.7Gigaton CO2/Y. Compare to 3.9Gton/Y of 1.4bn of petrol cars running 20,000km/Y at fuel efficiency of 5litre/100 km. The 3.7 Gigaton CO2/Y would be a CO2 disaster and breach against the Paris accord, and would nullify the electric car replacement of petrol cars in the coming 20-30 years!
It is RE’s environmental goal to stop NEW 3.7 Gigaton CO2/Y happening, enabled by our efficiency actions.
Then we have with 111x more traffic, reduced the 2019 global energy cost to US$25B*111*1/97* 0.10/0.59 = US$ 4.8 B/Y = -80% relative the 2019 global site energy costs! And ZERO CO2!
TowerXchange: You are focused on the techno-economic-energy problems with LTE, do you think that these problems can be resolved through 5G?
Torbjorn Johnson, CEO, Radio Evolution Sweden AB:
5G has been hyped to do everything, but I do not see it making operators profitable using towers.’’’’’’’’’’’’’’’’ ’We have to get LTE lower band towerco platforms right and very economical and future-proof, before overspending on high band 5G, which always has shorter coverage and higher energy costs.
TowerXchange: Give some 99.5% pop Coverage+Cap-$ examples for LTE+5G with SuperTowerCos?
Torbjorn Johnson, CEO, Radio Evolution Sweden AB:
Kenya examplev3B SuperCovMode2. LB only, 99.5% nationwide. For 3 tenants Y3-4 covering 564’477 km2, from only 291 SuperSiteHotels.’ 1,939km2/site/aver, 52.6Mcust*0.5Mb/s/BH, 50GB/mnth/aver. 180,824cust/site. (Doc 5068KA) With the average SuperSite as below: BU1. SET Tower COR225m US$72,500 (10%), With 2 convenient enclosed Lifts. And capacity 960 panels.
BU2. SAS-Panels 523x, US$41,837 (6%), Models from 470’.7,150 MHz
BU3. Fibre, US$35,555 (5%), located close to national fibre grid.
BU5. Energy+ups, 2,933*13Waver RRHs = 38kW, US$50,069 (8%). ’100% Green and NIL CO2 from Y1.
Leasings US$199,961/180,824cust = US$0.092/mnth (29%)
Then BU4. with 13Waver/10 MHz RRHs 2,933*$169/Y = US$495,667/Y (71%) = US$ 0.229/mnth
Totals = US$0.321/mnth. ’= -78% vs std Hotel below. Saving MNOs US$724M/Y. And 302x less CO2 emitted.
(Std 55m towers,12*18 dBi ant, would require 59x = 17,067 sites*33.1km2aver, 3,084 cust and 56 RRHs*40W at US$300/Y. BU1+2+3+5 US$37,601 + BU4 US$16,800 = US$54,401/Y. ’Totally US$1.47/mnth) This means that all 47 counties with pop densities from urban/average/rural 6,141/93/5.9 per km2 and 99.5% pop will be 50GB/month covered with the ITU DL low bands 960/35 + 821/30 + 788/30+20 MHz. Supporting all MNOs with total leasings of US$58.2mn/Y = 9.1% vs US$642mn/Y. Together with fewer and lower Power RRHs, the total savings would be US$724mn/Y. And the new SuperTowerCo would make an EBT of US$29mn/Y * PE15 = MCAP US$437 M. And polluting ZERO CO2. When more traffic towards 500 GB/month at < US$1.00 are needed, the higher bands are added together with some hundreds of densified FCF self-financed Daughters for SuperCapModes. Future-Proof.
India example SuperCapMode4 All bands Y8. Nationwide 99.5% 3,200,000 km2 from 15,000 SuperSites. ’Capacity: ’SEaver 2,0*DL1,000MHz*75Lobes*MIMO8 =1,200,000 Mb/s to 240,000cust*5 Mb/s/BH, 500GB/month. See also examples in (Doc 5700A.) Gives: BU1. SET Twr, COR225m US$72,500/Y (5,7%), Capacity for all coming bands and all MNOs and ISPs.
BU2.’SAS-panels 960x, Ga 35-40-45 dBi, supporting all bands 703-7,150 MHz + 5G MW LOS. US$127,680 (10.0%),
BU3. Fibre trm US$ 61,933 (4,8%), along fibre grid. And often co-located with edge computing containers.
BU5. Energy green, 120 kW for 2W/10MHz Low Power-WB Super-RRHs ’US$105,100 (8,2%).’’’’’’’’’’ Saving India several GigaWatt coal plants and reducing to ZERO CO2 emissions.
Leasings US$367,213 /12/1,200,000 Mb/s*5.0 Mb/s = US$0.127/500GB/mnth. ’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’ ’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’Then BU4 RRHs with 2W/10 MHz LP-WB-SuperRRH+column 41 Waver 2,880*$350 = US$1,008,000/Y (79%).
Totals = US$0,47/500GB/month.’ Saves all Indian MNOs >US$35bn/Y, when employing all 4G+5G spectrum nationwide vs old standard sites of 2019.
This also means that with the SuperCoverage and SuperCapacity support, nationwide 99.5% and the higher data rates and very low cost 500GB capacity, India could also become the future-proof capacity multiplying affordable model for many other countries.
Upgrade opportunity for obsolete old towers. Limited by COR 55 m and EPAtot 22,8m2 specified built as ’Open Bird’s Nest’ with 18dBi few panels limited Ael area 9m2. Upgraded by enclosing the top 7,5m with 3 SET D5.4m platforms 3*48 SAS panels, we get Ael =119 m2 =13,2x. Inside accommodates 100eds of RRH modules. This enables CovF2-3x and CapF43x, and -70% US$/GB. And they can have a second profitable life. SuperTowerCo’s Economy, highly profitable and self-breeding 10 daughters/10Y: ’Y1 capex US$300k, Equity US$120k, Loan US$180k, IR 6%/Y. Leases US$150k, opex US$30k, IR US$11k, tax US$0, FCF US$90k, Depr US$15k, Net US$75k. ’The FCF every year is shared to capacity modules, and self-breeding daughters with Y1 equity of US$120k. After 10Y, the mother has breeded 10 ditto daughters, for next densified CapacityMode generations. Mother with a Net US$133k*PE15, MCAP valued =$2.0 M, plus 1-9 Y old daughters, the family together MCAP = US$10M, from the Y1 equity of US$ 120k. Gives RoE/Y 153%.
A regional SuperTowerCo with 10,000 sites over 2,000,000 km2 with 960 SAS panels all bands serving 100,000cust/ea*500GB/mnth could reach MCAP of >$20B. ’
TowerXchange: What are the barriers to rolling out this technology as described?
Torbjorn Johnson, CEO, Radio Evolution Sweden AB:
Any disruptive change will face barriers to adoption, but because my solutions are dramatically lower US$/km2 and US$1.00/500GB, then 2,000GB/$, MNOs really need dramatic cost cuttings everywhere in all 5 BUs, therefore it will be adopted, when they learn to compare the 37 KPIs. They can’t afford to wait and see.There is a potential barrier from some regulators, having old EIRP rulings. We disperse the towers, move the antennas high above ground, away from people. In any direction at full capacity all bands, we only radiate 1,000W real power. This actually reduces ICNIRP radiation density <<10 W/m2, relative low elevated standard sites by < 1/10,000x. Also, some regulators are slow to release spectrum, and charge a too high up-front price. Possible resistance from existing towercos and MNOs because they are tied up in the old towers. American Tower is making US$102,236/site in the US, costing US$9.57/month/customer, and Helios Towers US$85,405/site in the DRC. What I propose would completely change their model, if they don’t significantly upgrade their old towers for LTE+5G.
TowerXchange: ’You are proposing an enormous shake up in the telecom sector, what are the next steps?
Torbjorn Johnson, CEO, Radio Evolution Sweden AB:
RE moves forward with strategic strong partners, hardware licenses, franchising partners, and a minority shareholding in new SuperTowerCos. And we will support old towercos to significantly upgrade their existing small voice-built towers with SAS antenna platforms.
Bill Gates is now working on medicine and poverty. My role with partners is to implement 126,000 SuperSites on 6 continents to cover 99.5% of the population, open for all MNOs+ISPs, leasable at affordable cost of <$1.00/500/2000GB, with zero CO2.