"I'm on the train..Oh!"
Any idea why the train has stopped?
Shenzhen Metro is blaming customer Wi-Fi for disruptions to its service. The subway system for the city of Shenzhen in Guangdong province, China, depends on the unlicensed 2.4GHz band to link up its signalling systems. Following network failures in October, and a trial blocking of 3G signals earlier this month, the Shenzhen …
The Chinese version of Ofcom should allow the metro stations to switch to 5.8GHz (or whatever makes sense) and waive licensing fees but require a check for interference with any existing local users of that frequency.
It's a public safety risk and officialdom should respond by being flexible and sensible. Ah ........oh :(
Where does it say anywhere in the article that the signalling system uses WiFi?
They're almost certainly using a form of GSM-R sitting in the same band that these WiFi devices are using.
They can't run wires down the trackbed because modern signalling systems do away with trackside signals (fixed block working) and replace them with in-cab signalling that maintains separation between trains (moving block signalling). This allows for a greater throughput of trains with higher levels of safety.
Um. What is this "cab" thing you speak of? And what does it have to do with metro services?
I make daily use of the Lille metro system, and except during ticket-check festivals, there are no metro staff on the trains. They are VAL205[*] and VAL208 units running on rubber tyres both above and below ground.
[*] VAL originally stood for "Villeneuve d'Ascq à Lille" (Villeneuve d'Ascq to Lille) but now (due to adoption elsewhere) stands for "Vehicule Automatique Léger" (Light Automatic Vehicle). There are trackside signals visible from the front of the units, and each end has a locked cover of some sort, presumably concealing a manual control panel.
"2.4GHz is reserved, globally, for unlicensed ISM (Industrial, Scientific and Medical) use, largely because it was considered worthless as it gets absorbed by water and because the band is rife with interference from microwave ovens."
a) it doesn't get absorbed by water any more than any another band area around it.
b) it is used by microwave ovens BECAUSE it's an ISM band, not the other way around! :)
The point I was getting at - in the name of putting untruths right - my reward being downvotes...
... is that it's a minor urban myth that there's something special about 2.45GHz that makes it ideal for cooking (water-based) food.
It's not. The same effect would occur if it was 2.3,2.1,1.9,2.5 GHz - pick a number at random - the only reason they work at 2.45GHz is because ...... it was already an ISM band!
"How do you run a cable to a train moving at 40MPH? Big curly cord like a headphone extension lead from the 70's?"
They've already solved that problem once when they had to figure out how to get electricity on to a train running at 40mph
I'm sure they can do it again
Right, because several hundred volts of DC and intricate signalling information are almost identical aren't they? You don't think that the giant spikes and drops that must occur on that giant miles long bus bar might be detrimental to any kind of data transmission?
The solution to modern railway signalling requirements is wireless - it's GSM-R or one of its variants.
I didn’t even suggest using the power supply to deliver the network, you seem to have leapt to that conclusion yourself whilst also attaining a strangely high level of anger. Or is asking lots of questions whilst making a point normal for you?
However, to answer you points:
The bus bar won’t be miles long, they’re typically split into discreet sections which can be turned on/off
Spikes and drops are all part of the noise, if the signal to noise ratio is within acceptable limits then data transmission at a suitable bandwidth can occur
I presume you've never heard of GSM-R or ERTMS then?
Wireless is now a standard signalling method - and it has to be as train protection now centres around the trains themselves rather than dividing the track into blocks. It's the only way that high speed trains such as the TGV can safely run without drastically limiting the number of trains allowed in a given distance of line.
Instead of allowing only one train per fixed block (with an empty block between each occupied one) each train now has a speed dependent protection zone in front and behind of it that no other train is allowed to enter. That has to work wirelessly as the trains are moving.
To the best of my knowledge, in a dozen years of this being in operation in many countries around the world there's not been a single accident caused by failure of this signalling system.
Maybe the phrase "safety critical" is the phrase being sought. I'd question the 2.4GHz band's status of being for ISM as well, with WiFi, Bluetooth and Zigbee (however rare that is) shoe-horned in there it should have been a no brainer. It's only the RF modules they need to change, it's not as if you have to redesign the whole lot.
People rarely make idiotic decisions - they make the best decisions they can with the information and resources available to them at the time.
Isn't it far more likely that at the time the system was installed the band was essentially empty? All systems are built with a certain set of assumptions built in - that there is a country called Yugoslavia, that year variables start with "19", that currency is French Francs, that VAT is 15% and in this case - that the 2.4Ghz band is free of sources of interference. Along comes a host of unregulated MiFi devices a few years later and the presumption is proved wrong and the system breaks.
80% of the cost of a system, any system, is in maintenance and updates, not in the installation and development. The reason for that is that the world changes and your presumptions are proved false. Unless you are gifted with prescience, this will always be the case. Assuming that you work in IT, describing the designers of the system as "idiots" is doing your profession a dis-service. Or are you claiming that every system you have ever worked on, ever, works for all eternity with no changes?
People make *cheap* decisions that often turn out to be idiotic.
The simple case is any unregulated band is a disaster waiting to happen, which is why you end up paying for something that is (moderately) dependable like GSM if you need some quality of service. Even then, unless you have some extra QoS system in place for 'critical' use you can occasionally get swamped by other users (e.g. New Year calls & texts that saturate the network).
This post has been deleted by its author
It is an unregulated band true - but it's also line of sight, limited propagation and underground.
It would be hard to see in advance that anything could mess with that, unless you can determine that a decade in the future all your customers are going to start bringing 2.4Ghz transmitters onto the train with them. 2.4Ghz isn't a mobile band, so why would you even suspect that people might do that?
It's easy to criticise with the benefit of hindsight. I don't think these people were idiots.
I upvoted you because you (and someone else before you) make the only valid point in this saga: this seems to stem from a time where that band was empty and usable. You can't blame people for not knowing the future.
I fear their only solution will be either switching frequency or protocol, the former the better solution but also probably the costliest..
If you ever have the chance to use Shenzhen metro during rush hour, you'll find the Subject not so far off.
The original single line (opened 2005 from Luohu to "Window of the World") has been extended into a complex network in a gigantic effort, probably multiplying the volume of commuters by 10 or more. The former WoW endpoint platform now has 1.5 meter bottlenecks between the (thankfully glassed in) tracks and the cages around the stairwells going down to Shekou line. In a town of approx. 12 million...
The funny thing is: they'd only have to go (by metro) 1 hour South to see how one of the best systems in the world, named Hong Kong MTR, did it.
Don't get me wrong: The overall work is amazing, even the quality of the materials. Detail planning however must have been delegated to lower levels, see Subject.
"People rarely make idiotic decisions - they make the best decisions they can with the information and resources available to them at the time."
I think the AC has a valid complaint that putting a safety critical system into an unregulated band was not a wise thing to do, and certainly doesn't count as the best decision given information and resources available at the time. "Unused now" doesn't mean "unused forever" which is the assumption that they made. I haven't seen any complaints of similar problems elsewhere, which suggests that the majority of those responsible for wireless signalling systems didn't make this mistake, and putting together the erroneous assumption with its apparent uniqueness suggests that Shenzen metro planners were in fact idiots.
Would you assume that unregulated spectrum that is currently unused would remain so for the asset life of a major infrastructure project? I'll bet you wouldn't, both because you're probably smart enough not to, and because on a major project you would be controlled by a risk management approach that should highlight the consequences of all unproven assumptions.
So come on, admit they were plonkers.
Won't the wifi be using the same bands? Added to which, the 3G down the tunnels bit - well, I think we're focusing on urban rapid transit as a form of underground, when the article refers to problems being stopped if they blocked the 3G signal, and in that case the comments many people (self included) have made with tunnels in mind are probably irrelevant, because the problems must be being caused above ground?
The article could be clearer, but it seems to me that the problem is probably caused by mifi use above ground in areas of 3G reception - otherwise there's no backhaul.
Can any other commentards shed any light on this?
You don't think that they probably did use professionals? I've never been to China, but in the countries I have been to railway signalling is considered quite a difficult technical job and it tends to be given to people who know what they're doing. People who can prove through experience and qualification that they know what they're doing. It's unlikely that the Chinese authorities popped down to the local equivalent of PC World and asked them what they'd recommend and whether they guy would be interested in installing it after work for £100, cash in hand.
Running cable and fibre wouldn't do it properly because modern signalling systems use radio. They're safer and more effective than old fixed block systems.
At the time the system was installed, it would have been clear of interference. It's line-of-sight, of limited propagation and underground. Only a decade later when people start bringing MiFi transmitters onto the trains with them has there been a problem. The solution is to move band.
"You don't think that they probably did use professionals?"
What can be said of a professional that puts a critical system via a radio band that anyone can, at any time, set up another transmitter in your area and flood your frequency and you cannot do a thing to stop it.
There is a difference between professionals who set up train systems and professionals who set up radio systems.
I know of other systems that were put in by professional engineers (not professional radio engineers) which failed due to being on the unlicensed band.
Biting the hand that feeds IT © 1998–2020