Bristol East Filton Supercharger Closing 30-Nov-2016

[arg]

Sorry for being a bit dim but can you explain that?

Since people are turning up essentially at random, sometimes (actually a lot of the time) there will be empty spaces, while at other times there will be a queue even though there would have been enough capacity for those people if they'd turned up slightly earlier or later.

This subject has been extensively studied mathematically ( Queueing theory - Wikipedia ), most particularly for telephone systems - how many lines do you need to carry a certain amount of telephone traffic? This is very close to the same question as "how many supercharger stalls do you need to serve a certain amount of Tesla traffic".

For telephones, the numbers are normally stated in terms of how much traffic can be carried by a given number of lines for a certain quality of service - where quality of service means the probability that you won't be able to make your call because all the lines are in use.

So for a quality of service of 0.01 (1% chance of finding all lines occupied)

1 line can carry 0.0101 units of traffic
2 lines can carry 0.1526 (15 times as much as 1 line)
4 lines can carry 0.8694 (5.7 times as much as 2 lines)
8 lines can carry 3,128 (3.6 times as much as 4 lines)
12 lines can carry 6.615
90 lines can carry 74.86

If you are prepared to accept a poor quality of service, say 10% chance of finding all lines occupied, things get more even but still capacity climbing faster than number of stalls.
1 line can carry 0.111
2 lines can carry 0.595 (5.3 times as much as 1 line)
4 lines can carry 2.045 (4.4 times as much as 2 lines)
8 lines can carry 5.597 (2.7 times as much as 4 lines)
12 lines can carry 9,474

Intuitively you can see that it would be like this - you always need to keep a couple of stalls free to handle new arrivals that might or might not turn up, but the more stalls/lines you have, the more of them can be working away solidly rather than kept in reserve.

The Supercharger numbers won't be quite the same as for telephones - there's factors like the sharing of cabinets between pairs of stalls, the time taken to enter/exit stalls, the behaviour of drrivers when all stalls are taken etc. I expect Tesla has their own numbers that are more accurate, but it's going to look something like this.

 

[DrBobM]

Since people are turning up essentially at random, sometimes (actually a lot of the time) there will be empty spaces, while at other times there will be a queue even though there would have been enough capacity for those people if they'd turned up slightly earlier or later.

Thanks for you patience, helpful as ever

 

[busaman]

superchargers are wired in pairs a&b =130kw ish available A alone full power and B. both used A gets the lions share and B reduced current if that makes sense at least thats how i understand it..

 

[arg]

superchargers are wired in pairs a&b =130kw ish available A alone full power and B. both used A gets the lions share and B reduced current if that makes sense at least thats how i understand it..

There's no preference of A over B (or vice-versa) - it's the first car to arrive that gets preference.

The first car to connect gets initially as much as it can take. When a second car arrives, it is given whatever is left over by the first car, but a minimum of 1/4 of the capacity (about 30kW), with the first car's share being reduced if necessary to give that minimum.
As the first car's charging slows down, spare capacity can be transferred to the second car, but only in large chunks - it's not possible to split (say) 26%/74%.

The exact size of these chunks is not totally clear. The supercharger is made up of 12 separate chargers, so the minimum unit possible is 1 charger (1/12 of total capacity, about 10kW). I believe that the original superchargers could only switch in groups of 3 (1/4 of the total capacity, about 30kW); it is not clear whether the Gen2 superchargers improved on this and allowed finer grain switching, or if they are still switching in groups of 3.

The actual power per charger depends on the input voltage at the site - it's nominally 10kW per module, 120kW total output for "European" 230/400V, but in the UK close up against a transformer it might be getting 250/433V and so giving 130kW output; in the UK for a single cabinet on the end of a piece of damp string maybe 220V -> 115kW, or in France under similar worst conditions as little as 207V -> 108kW total.

 

[WannabeOwner]

close up against a transformer it might be getting ...

So, all other things being equal, choose the stall nearest to the transformer?

 

[arg]

So, all other things being equal, choose the stall nearest to the transformer?

I know you said it in jest, but it made me go calculating...

Normally the supercharger cabinets are all in one place and the stalls various distances away (maximum permitted distance of 30m). Given the cable size Tesla use, when charging at maximum power and at a stall the maximum 30m from the supercharger there's a voltage drop equivalent to about 1%. So maybe a 0.5% difference between the best and worst stall at a large site like South Mimms - and that only when charging at max power (where you are usually limited by some other factor anyhow). Given that there are separate cables to the two stalls, in a sharing situation it's going to be even less than this - maybe 0.25%.

At South Mimms, the stalls closest to the equipment enclosure are furthest from the entrance, so the tiny saving in charge time is probably offset by the extra time to drive to the end stall and park.

What I was really talking about was the cabling from superchargers to supply transformer - which is almost nothing at sites like South Mimms (which has its own dedicated transformer) but could be quite significant at the smaller sites at hotels etc which are just using the hotel's electrical supply. Could mean these sites have as much as 10% less power, with correspondingly longer charge times - particularly in the shared-stall situation.