By James Blackman

Private 5G – a mobile phone showing the Lufthansa Technik network, outside the Lufthansa base

Note, this post is continued from an article called, “5G helped us through Covid’ – Lufthansa Technik on private 5G with Nokia, Vodafone”.

But this is not about how a nascent technology will save your business. That is not what Lufthansa Technik meant, actually. 5G is an enabling technology, after all; a springboard, at best, for other new-age digital pyrotechnics – if, in the end, the hype holds up, and it ‘does what it says on the tin’. It means nothing on its own, and might mean nothing at all for your own private industrial revolution. It depends what you want it for.

That is the first question, which Lufthansa Technik answered a year ago, when it decided to build two private 5G networks in two aircraft hangars in Hamburg, to resolve two business bottlenecks. But did it work? Did 5G live up to the hype? Those are the questions the company sought to answer at the

Athonet shindig a couple of weeks back (see coverage here, also), revealing the results of 12 months of testing, through global lockdown and crisis, and reflecting on the differences between the two setups

So what did it find? Well, that new-5G works better than old Wi-Fi for network latency and point-to-point roaming, as we heard from the headline-making look at the first Nokia installation last time out. But the company’s review was more thorough, and its findings more varied. Because its two use cases – uploading and downloading HD and 3D streams, alternatively, for remote collaboration on engine maintenance and cabin overhauls, respectively – made different demands of its two 5G setups.

So, in practical terms, 5G is not about ‘transforming’ industry, at all; it is about the hard graft of making marginal operational gains in a grim economic climate. Other measures must be considered, too – like capacity, speed, propagation, and even just the partnership that supplied the network in the first place. And these measures should be evaluated in each use case, each time. This review process is more important because the technology is new, and the two (IT/OT) sides in the negotiation are still scoping out a common ground.

That, then, was the gist of the Athonet address. There were some neat home truths in there, too. “Last week was the first time I used a real 5G IoT device in the 5G network in Hamburg. So this is pretty new stuff,” remarked Claudius Noack, IT consultant at Lufthansa Industry Solutions, the business unit overseeing the two installations, and ultimately looking to resell Industry 4.0 consultancy from the findings. He has only just taken receipt of a 5G-based IoT unit he can plug his consumer devices into, he said.

Because private networks have not had their ‘Apple moment’, it seems; not even the iPhone 12, heralded in some circles as the breakthrough device for handheld Industry 4.0, works on private 5G, it turns out. The Apple iOS blocks private 5G SIM cards; newer iOS releases should resolve the issue. But, for now, for whatever reason, enterprises with their own 5G networks are required to either plug into 5G-based industrial IoT units, or else connect to them on Wi-Fi over a 5G gateway.

Hardly ideal, and hardly feasible to connect billions of ‘things’, as per the ‘massive IoT’ manifesto. The point is only that the dream of industrial 5G is out of reach still, untethered in an almost-total hardware vacuum; Lufthansa Technik’s own tests are, arguably, just the best attempt yet to make it real. (There is a whole discussion to be had about who blinks first over the supply and demand of industrial 5G chips, between the likes of Qualcomm and Bosch, say – and that, in the end, neither will budge until the delayed standards-releases for 5G-proper come into view.)

But back to reality, and the answers to those questions about bandwidth and propagation in private 5G networks. Besides latency, the key measure of the Nokia setup, geared for live video uploads to support remote collaboration on engine maintenance, is the transmission speed. Besides latency and speed, the key measure of the Vodafone setup, focused on software downloads for AR-assisted cabin refits, is signal degradation through aircraft fuselages. 

5G antenna – in the background, to the left, in the aircraft hanger in Hamburg (Image: Lufthansa Technik)

The key to these metrics, in each case, is the contrasting nature of the twin setups. Both are standalone (SA) 5G networks, both use privately-licensed spectrum, and both focus generally on industrial use cases. But that is where the similarities end. The first, with Nokia, employs two indoor 5G small cells; the other, with Vodafone, a kilometre away and out of range, uses a single outdoor macro cell, fixed to the inside of the hangar.

There was a good anecdote at the Athonet event about taking delivery of the hardware: the team struggled to find a room big enough to hold the Vodafone gear, whereas the Nokia kit arrived in the back of a Ford Fiesta, driven by Noack’s colleague, Maik Voigt, who led the project on behalf of Lufthansa Technik. Noack was quick to point out this was late-2019, when the deliveries turned up, and that the unnamed vendor in the Vodafone setup has since shrunk its radio solution to a comparable size – and that Vodafone, for the record, is doing just fine with private 5G.

But he referenced a previous talk at the event, about private 5G in the Middle East for city-wide coverage, and cell capacity for “a million users and end-devices”. Which is what Lufthansa Technik got, initially, in the Vodafone deal – a big-scale solution for a small-scale problem. “They installed a 5G core at our place for a million end devices. But we are using, maybe, 10 – and we might use 1,000 at some point in the future. But never a million.”

Noack restated that it was the only solution available at the time, and that the vendor has since revised its wares, but he left it hanging as a cautionary tale about shopping around, and stopping by the smaller providers, as well. “That is what you have to keep in mind, if you are going with the big players – that they might not be right for your campus… There can be huge differences in the hardware – in terms of what you get and what you need,” he said.

But what about these test results? Data rates across the two antennas in the Nokia network are good, it seems. Noack’s team has recorded a stable download rate of 500 Mbps, with highs upwards of 600 Mbps, and even 235 Mbps down in the corners. It is not 10 Gbps, as the market has promised, said Noack; but rather like with 5G latency, discussed before, the reality gnaws, rather than bites; it is probably good enough.

Except this is not the measure that matters for remote table inspections, either, where high-definition video of engine parts is being sent to far-away customers for consultation. The upload speed is what counts, and mobile networks are typically configured 5:1 for download/upload speeds. “It is very tough to get high numbers on the uplink,” noted Noack. So does private 5G deliver? It sure does, it seems.

Table Inspection – high-resolution live imagery of engine inspection is beamed back to customers (Image: Lufthansa Technik)

Speaking after the event, Voigt explained: “We thought we’d need 4×25 Mbps; we measured with multiple devices – and those numbers are not very high anymore [on paper], but they were incredibly high for us, and really solved our problems. Because now we had a stable 100 Mbps upload stream with 5G.” Reading between the lines, the message from Hamburg is the marketing strapline about ‘super/ultra/silly-fast’ 5G should be reworked for Industry 4.0, probably just to read ‘fast-enough, maybe’.

What about the second case, with Vodafone? The results are interesting because they break new ground. “When we started no one knew how well we could go through the hulls of those aircrafts. No one. Not Nokia, not Ericsson, not Huawei, not anyone,” said Noack. Not anymore, not now; Lufthansa Technik knows exactly how 5G degrades through aluminium alloy and carbon fibre, and whether it can stand-up reliable two-way comms between cabin ‘techs’ and edge servers.

The answer is: of course, it can. The challenge with the AR system is to move the 3D overlays with the technicians, in near ‘real-time’, as they make their way through the fuselage to guide them precisely on how to fit cabling and cabin furniture (“even showers”), and to reflect design-changes back to the server. The crew used beamforming to direct the stream via the antenna at the tablets in the hands of the technicians, and the transfer hit 1.2 Gbps outside the hull and 800 Mbps inside the hull.

Noack explained: “They are incredible numbers. We thought, ‘Okay, we’d never achieve that with Wi-Fi’, which we’d tested as well. It was like, ‘Okay, a 32 per cent loss in the hull; that’s it – 32 per cent. But it was funny because we turned the lights off inside the aircraft for one measurement, and the rates inside went up to 1 Gbps – these old fluorescent light tubes, with a metallic ballast, were interfering with the network.”

He added: “Which is important to know, if you want a campus network – that some equipment might create a magnetic field and disturb your 5G network, as it did for us. Because with the lights off, we got to over 1 Gbps, on a stable stream. That is a real result; the hull is nothing, a five percent loss. So we were really delighted.” Emboldened, the team measured everywhere, inside and outside of the hangar.

AR for cabin fixes – showing 3D design of a cabin interior in an empty aircraft fuselage (Image: Lufthansa Technik)

Talking over a series of slides, Noack commented: “We reached 900 Mbps over here, 700 Mbps over there. We went into the nextdoor hangar, and we still had 300 Mbps. We went outside; 600 Mbps over here, dropping very low over there – but that is like 350 metres away, and we still had coverage.”

By comparison, the public LTE/4G network in the airport area provides steady readouts of 128 Mbps; it is capped locally at 150 Mbps, said Noack – no good for cabin-based AR, requiring 300 Mbps for downloads. What of Wi-Fi? Lufthansa Technik is running Wi-Fi 5, which stalls when roaming between access points, as discussed. He said: “It’s a big hangar – Wi-Fi 6 would cost a lot of money, [and] we have it covered with one 5G antenna.”

The crew ran other tests, too, including with lower output power from the base station. They discovered download speeds went up, to a point, as output power went down. “At 50 per cent power, it went up. You think it would go down, but, no. We did so many measurements, because we weren’t sure. But we found, repeatedly, the best result was at 56 per cent output power, not at 100 per cent.”

Conversely, of course, the coverage area goes down with the output power. “At 100 percent output power, you still have a very wide area, like 500 metres to the antenna. At 50 per cent, you can still be very far away, but the data rate starts to get lower; at one percent output power, you can’t even go 500 metres. The lesson is you can’t go as far with lower power ratings, but you might be able to edge the speed higher.”

In the end, Lufthansa Technik backtracked from its headline-making about 5G saving the world, or whatever. It does not want to be attached, it implied, to the telecoms sector’s grandstanding in the industrial space. Its networks are real, their results are real, and their impacts are tangible – but their wider business effects are marginal, for now, in the scheme of things.

At the coal face, the company is not much interested in the great promise of industrial 5G, nor the grand talk of industrial revolution. Noack’s team is down in the weeds, trying to solve real business challenges. And, it seems, 5G – privately deployed, carefully considered – can help with those too. In some cases.