Supercharger Power relative to Start SoC and Amount Charged

Given that I just hit 40 total supercharges over the weekend (not 40 superchargeRs - I've hit 18 unique ones), I figured I'd do some data crunching to figure out what affects the average power delivered by a supercharger based on different variables.

To spare you the details, I'll jump right to the conclusion that the highest two variables that affect average power from a supercharger are: state of charge (SoC) at the start of the charge cycle, and the amount of energy delivered during said cycle.

And since a picture is worth 1,000 words, I'll share a graph in which I've sorted average power from lowest (on the left) to highest (on the right):




Now, there is strong evidence and it's common wisdom that the lower start SoC is, the higher initial power output will be. Also, because of the strong tapering effect above a certain SoC, lower amounts charged (and hence staying below higher SoC's) helps keep a higher power output. Of course, a higher end SoC is sometimes needed to arrive at the next supercharger or the destination.

However, it seems there's a sweet spot to hit about 80 kW of average power where start SoC is around 10%, and end SoC is 70% (hence charged amount is 60%, which is about 45kWh in an 85kWh battery since the usable amount is around 75kWh).

So it appears that the best "bang for the buck" in terms of getting the most out of a supercharge is to arrive with ~10% and only charge to ~70% (if that's enough to make it to the next destination, of course). This will minimize the amount of time spent waiting for a charge: around 34 minutes.

By contrast, arriving at a supercharger with 30% and charging to 90% will only charge at an average power of about 55 kW, which will take around 49 minutes, or ~1.5 times longer than the "ideal" supercharge scenario.

So, over to you: what has been your experience with supercharges at different start/end SoC and amounts charged?

Benjamin Brooks said:

Was all this ~40x charge cycle data taken exclusively on unpaired stalls? (i.e. you had access to full 120kW of charging capability during the charging sessions, whether or not the battery got it).

Click to expand...

Good question. Until this last weekend's supercharge at Mountain View, I always plugged in to an unpaired stall that had the "A" marking such that, even if another Model S plugged in to the "B" stall of the same number, mine would continue to get as much juice as it needed. I was able to validate this with the VisibleTesla logs, whereby the graph of decreasing current (as SoC increases) was always a smooth log curve, confirming no current cap from the supercharger (as opposed to a step curve with a noticeable drop at the moment another S plugged into the "sister" stall).

As for this last weekend's event at the Mountain View supercharger, the place was really packed so I initially plugged into a "B" stall of pair #2, where 2A was occupied by an actively-charging S85 (the owner was waiting in his car). Seeing as I was capped to about 200A, even though my SoC was under 30%, I moved to another unpaired stall when it freed up and got over 200A (at least initially). At some point later, another S pulled into the stall paired with mine, but I didn't notice a sudden drop in current (I was watching it closely and confirmed with VT later).

Anyway, I did not log the first short supercharge at Mtn View as I knew it would skew my data, but I did include the second unhindered charge. So the 40 samples are all based on the car pulling as much power as it needed.

Benjamin Brooks said:

It probably does provide higher average power. However none of MarcG's real world SCing datapoints start at 0% SoC.

Click to expand...

The only data point I have that starts close to 0% was when I had 2 rated miles remaining, so about 0.8% SoC. However, I charged to 80% (not 60%) so it wouldn't be a fair comparison. I honestly don't know if 0-60% would be faster than 10-70%, but if I do run into this event I will be sure to log it and report back here.

Candleflame said:

I think comparing 1% to 60% with 10 to 70% would be quite interesting for comparison. I'm gonna assume the latter is perhaps 30 seconds to a minute slower.

Click to expand...

Yes it would be interesting, and I wouldn't be surprised if your assumption was spot-on.

Here's the graph from Tesla's website that shows how a supercharge tapers, and even though it's probably not to scale and it's not obvious what start SoC is assumed, to me it seems the first 50% or so doesn't see much taper at all:

Here's another interesting graph I just put together, which shows the average power delivered over a supercharge relative to start SoC and end SoC.

It doesn't take a regression analysis expert to see (graphically) that the average power over a supercharging session is much more closely correlated to end SoC (red dots) than start SoC (blue dots).

I went ahead and added the R[SUP]2[/SUP] numbers of each trendline anyway, which confirm this graphical indication: