REMINDER: Leave a gap on V2 superchargers

[J0hn.r]

How hard would it be for Tesla to provide a new feature that, when you arrive at your supercharger, the navigation map changes to a supercharger layout map that presents all used slots and picks the best available one for you? All in a nice graphical way.
Also: when you are connected to a bay that has another charging session paired, to notify you via the app that this is happening, tell you why you better use a different bay, and even suggest the correct free bay to pick?
But wait..farting and raindeers are much more interesting to develop!

 

[arg]

Maybe it’s site or SoC specific based on how dramatic the effect is...

It is.

The (v1 or V2) supercharger cabinet has 12 modules, of the same type used in pre-facelift Model S. The minimum allocation given to a 2nd car connecting is 3 modules (25% of the total capacity). I believe that the overall granularity for sharing is groups of 3 modules (so the split is 12:0, 9:3 or 6:6), though I am not as certain of that as I am about the initial 3.

These modules are limited to 16A per phase (or 40A single phase per module for V1: I don't think there are any V1 in the UK, but that might be the reason for switching in groups of 3 rather than individual and was never revised when they upgraded to the v2 modules). So like in the cars the higher the voltage the more power you get - and in the superchargers they are rated up to the 480/277V of a USA industrial 3-phase feed rather than the 400/230V nominal for EU-spec cars.

At 480V/277V, the modules are rated for 12kW output power each (not to be confused with the 11kW input power we are used to seeing for them at 400/230V: in that condition the output power is only 10kW).

So in the old days at a site with maximum voltage, it was true that plugging in a 2nd car had no effect on the 1st car's charge rate, as the 1st car couldn't use more than 9 units in the first place (note that at launch with V1 superchargers, supercharging was 90kW maximum; later they upped that to 120kW if the sole car at a V1 supercharger, then V2 came along and you could get just about 120kW off 9 modules and it was back to 2nd car doesn't take away from first).

Now they've done the same thing again and single car at a V2 can get 150kW (not sure if that's "marketing kW", or they've decided they can thrash the modules 3% harder, or if you can only get 150kW where the voltage is above nominal: the original spec sheet for the V2 claimed 145kW maximum at 480/277V).

So nowadays even at a max voltage site a 2nd car can drop the charge rate of the 1st car, but only if the 1st car was taking more than 120kW - which means only some models and only at the optimum SoC and temperature etc.

However, in the UK not all sites are max voltage sites. A few have Tesla-owned transformers and operate internally at 480V - I believe South Mimms is one such as it has battery storage and a non-standard looking transformer (the plans for the as-yet-unbuilt LFE site showed an internal operating voltage of 480V to suit the PowerPack battery units).

Many sites with a dedicated but not Tesla-owned transformer are probably 440/253V (max permitted UK supply voltage and most standard transformers are wound for this) so would expect to deliver 133kW to a single car, and 100kW to the first car once a second has arrived.

Old two-stall sites using an existing supply and/or a transformer at a distance from the superchargers would be more like 400/230V, so 120kW to a single car and 90kW to the first car of two.

And at any of these sites the position will be slightly worse if we are talking about (say) the 12th car arriving at a full site rather than a 2nd car badly parked at an otherwise empty site: the other 10 cars will have dragged the voltage down slightly.

 

[cezdoc]

@arg I'd be interested on your thoughts on this post

Pairing not quite what I'd been led to believe

which appears to show a 90D initially charging at 112 kW but then dropping step-wise to 69 kW (after a brief period at 84 kW) when a Model 3 plugged into the same pair.

 

[arg]

Oh, and as an addendum to the above:

• There's the case of faulty units. Superchargers can continue working when one or more of the modules has gone faulty. This can leave you with a bigger impact when a 2nd car plugs in (because it still gets 3 good modules and the first car takes the hit from the faulty ones).
• Superchargers can overheat. Probably unlikely in the UK for a healthy unit, but possible if the radiator is clogged, coolant level low etc. This can give low charge rates as the supercharger throttles back to limit the temperature. Again can make the 2nd car have disproportionate effect, also can cause effects after a period of charging (so it wasn't the 2nd car's fault entirely, the rate was going to drop after xx minutes anyhow).
• Some early sites were power limited to less than the supercharger cabinet rating. While this was definitely true, exactly what was done is the subject of rumour/speculation. Maybe it was just knocked down to a fixed lower rate; there were rumours of some sites having lower power during daytime when the host site was using more power for its own purposes. Theoretically possible for them to have implemented an overall limit on a total site, though I've never seen evidence of that. I'm not convinced any of this has ever applied to the UK apart from maybe 2-stall sites; probably larger sites where it has been speculated are actually just faulty, but it's conceivable there's some deliberate throttling going on.

 

[Glan gluaisne]

Just like to thank @arg for those enlightening posts. It's really useful to have some hard information on what's going on under the covers of this stuff.

 

[arg]

@arg I'd be interested on your thoughts on this post

Pairing not quite what I'd been led to believe

which appears to show a 90D initially charging at 112 kW but then dropping step-wise to 69 kW (after a brief period at 84 kW) when a Model 3 plugged into the same pair.

It's hard to analyse accurately because we don't know what the M3 was seeing.

It's apparently a V2 site in the USA, 150kW according to Tesla's website, so should be delivering approx 37kW per group of 3 modules if healthy.

If so, it should have been using only 9 modules initially and with the 'traditional' behaviour would not have expected any impact from a 2nd car plugging in. If it was standard hardware but the sharing algorithm has changed, then you'd expect the potential steps down to be 74kW and 35kW. However, that doesn't match what we see here.

XXX having spent some time analysing those numbers, I've looked closer at the actual graphs and the true powers appear to be 115kW, 93kW, 70kW. Not sure why the OP said 84kW. If he was reading the charge rate off the dash, it makes sense that the actual power drawn from the supercharger was a couple of kW greater due to cabin heat etc. so 115 vs 112 and 69 vs 70 makes sense, but 93 vs 84 seems like a mistake. Peak power (earlier in the session) was 122kW. At the time the M3 plugged in, the 1st car was slowly ramping down and had reached 115kW; it then sat at 93kW constant for about 30 seconds, then jumped to 70kW and slowly ramped down over the remaining 90sec until disconnected.

If we believe the two big steps in output relate to switching chunks of capacity from the 1st car to the M3, then the initial charge rate will have been the max the 1st car could take and so not related to the supercharger capability, but both the subsequent levels should have been supercharger limited and so must be a multiple of modules. 93kW has to be at least 8 modules, 70kW has to be at least 6 modules (we know for certain each module is 12.5kW max and believe they are switched in groups of 3 but might not be). If 70kW was 6 modules flat out, that implies these particular modules are only achieving 11.7kW (due to low voltage or temperature or whatever). 9 modules at that rate would be 105kW, but 8 modules is 93.6 at that rate (would be 100kW at ideal rate).

So the first car was definitely using all 12 modules to start with (couldn't get 115kW out of 9 modules even under the best conditions). Neither of the subsequent states look like any possible number of modules run flat-out. It looks very much like the first step was to 9 modules (which is expected as soon as a 2nd car plugs in), then to 6 modules (which requires explanation). Also requiring explanation is why the car didn't get the max possible out of the number of modules it was given.


Plausible theories:

• The sharing algorithm is no longer first car priority. So the minimum 3 modules was switched to the 2nd car, 30 secs were spent getting it charging and estimating the rate it could take, then it was decided that for some reason the Model 3 was more deserving and went to a 50:50 split. This seems quite plausible but not proven by this data. One possibility is that rather than waiting until a car isn't using a group of modules at all, they are taken away once the first car isn't making full use of them - at 93kW, the car was wasting some capacity. If the second car was able to charge at 70kW or more, then the overall throughput would be greater.
  
• The first car used a bit less than the capacity it was given, certainly after the first step. It is possible that the shutdown and restart to allow the switching to take place caused some recalibration of the algorithm - certainly the rate was already ramping down when this happened. It seems most likely that the reason for 93kW after the first switch is probably that the car was only asking for that much. It is possible the car then decided to further reduce the rate all the way down to 70kW, but that seems less likely.
  
• The supercharger was in some way impaired (faulty module(s), over temperature etc). One faulty module would explain the first step being to 93kW rather than 100 or 105kW. This doesn't explain the second step however; it's just possible some more obscure type of fault does so, but this seems far-fetched. Being at less than maximum voltage and so only delivering 70kW with 6 modules flat out is quite likely though.
• I assumed above that the 115/93/70kW were total power input to the car, including auxiliaries. I don't know what that app is actually displaying - if those figures are the actual battery current (slightly implied by the fact that it's shown as a negative value), then there's auxiliaries on top and the car might be taking much closer to the maximum available to it. One further related idea is that the car was initially running the battery heater (it definitely was earlier in the session), but after being reduced to only 9 modules it was no longer worth the trouble of heating the battery further (since the faster charging allowed by a higher temperature would not in fact be achievable) so it switched off the heater at that point.

 

[cezdoc]

@arg thanks for the forensic analysis! With so many variables it's clearly impossible to know for sure whether sharing algorithms have been tweaked. Data from repeated examples from drivers with cars capable of charging at 100kW+ might shed some light - or just confirm that nothing's changed

A couple of years ago I never really considered the possibility of "slightly" faulty chargers (e.g. impaired modules) - the equipment on the UK routes that I might occasionally use was all pretty new back then. It's useful to bear in mind that as it ages there might be one or two chargers that could under-perform. Looking at my own car charging statistics, I'm usually supercharging in the 50-90kW range so sharing is very unlikely to seriously hamper me, especially if I'm first to plug in to a pair before another car turns up.