Well, my 2017 Zero SR has non-stock chargers that are programmed by Elcon. These are normally used for electric forklift charging. The program needs to be different to charge an EV motorcycle battery. Even the output voltage is different, which I think is set by the program.
Right. My point is the program is extremely simple (as tends to be the case with a lot of things). If it had to change, all that needed to change was a few simple parameters.
Also, CCS runs a program. In fact, many of the broken EV-Go programs show they are running a charger program from Windows XP (IIRC) when they are rebooted. That is the program I am referring to. But I think I understand what you're saying. Maybe . . . .
Well yes, there is a lot of software at work. But let's limit the scope of the discussion to that which handles the actual charging process, and not the user interface, and payment systems, etc.
During the initiation process before the charge starts, what is happening during that15 seconds or so?
Every charging station is different, but there are probably three things that must be done that take some time.
One is doing a safety check of the path from the high voltage equipment inside the charging station to the inlet on the vehicle. I think this is usually done by ramping up the voltage on the cable and trying to detect any shorts to ground. If it finds any, it will immediately shut down. There are other safety and environmental checks as well (making sure connectors are properly seated and latched, etc.) but those probably don't take much time.
Second is negotiating with the vehicle itself, and by that I mean the BMS which is in control of the process. The handshake process should not take long, but sometimes it does. With Tesla Superchargers, for example, it happens almost instantaneously, but that's probably because they devised a very simple communication protocol, whereas the committee that designed CCS probably through all kinds of stuff into the negotiation handshake. Plus if there are communication errors or unreliable communication, it probably has to try several times.
And third, which would apply to Plug & Charge and similar setups is the authentication process to securely identify the vehicle to the charging station and then the charging station goes off into the cloud somewhere to retrieve the identify and payment method assigned to the vehicle. Obviously this step is going to rely on the internet connection available to the station, as well as the responsiveness of the server on the other end, which may or may not be fast.
I probably have the order in which these are done reversed, but these are the likely culprits in the delay you see before the charge commences.
Before the FW was charged in the CP CCS chargers, it would stay stuck in the initiation process on my Energicas. And forever. It would not time out and it would not charge. I left it in the initiation process for more than a half hour once and no change before I disconnected and gave up with it.
Yeah, likely got stuck in the negotiation process. If you've had the pleasure of trying to connect two Bluetooth devices together, you probably understand this. Granted, in your case it's different because it wasn't an intermittent fail, but an entire incompatibility. That's the kind of thing that can happen when the negotiation/handshake process gets overly cumbersome: it's prone to failure and even complete incompatibility. It will be interesting to see how well Tesla's experiment with opening up Superchargers to other brands ultimately works out. Will their relatively simple interface win the day (a la how USB is today) or will it prove to not be sufficient for all the fancy things that other car manufacturers may want to do over the connection?
Is it the program that decides when to switch from CC to CV? I assumed that was in the BMS. I also assume it decides that by when the battery current gets so low at a certain point.
Yes, the BMS is 100% in charge of the charging process. It "is" the program. Other than the charging station shutting down itself for whatever reason (it detects a fault in its own systems for example), the BMS is what is calling for everything.
Is it the BMS or program that decides the charge rate based on battery temperature?
The BMS, although there are really only two temperature related aspects to what the BMS is going to do based on temperature.
The first is when it's too cold to charge. Li-ion batteries cannot be charged below a certain temperature without damaging them. So the BMS will call for power from the charging station, but it will use that the heat the battery, not actually charge it. Once the battery is above a certain temperature, it will do its normal CC/CV thing.
The second is if the battery gets too hot to safely charge, and it will shut down the charging process. This should never actually happen in a modern EV with working thermal management, but in a car like the Nissan LEAF without it, it could cause the charge to terminate early. This is really just a safety mechanism though, not an active technique used to modulate the charge rate.
The one exception to this is the Kia/Hyundai & Porsche thing I've mentioned, which do dynamically adjust the CC current based on pack temperature. I do expect that technique to be more widely adopted in the future. But again, it's a pretty simple adjustment on the CC current. It doesn't happen near the end of the charge session when the power being delivered is naturally tapering and the pack thermal management should have no trouble keeping up.
In the past, I assumed it was a complicated program and was one of the main reasons we have so many CCS failures. But perhaps not. What is the main cause of the failures? Is it something in the hardware?
I saw an excellent post by someone on
TMC forums with a pareto of CCS charging station failures. I don't know if this is anecdotal or if there was a report this came from, but here is a list:
- Broken CCS1 connector.
- Mainly the latch, this is a CCS1 design problem.
- Failed cable cooling system. (Limiting the output to 50kW.)
- Other companies go with air cooled cables.
- Failed temperature sensor in CCS1 plug. (Limiting the output to 50kW.)
- Bad design in Huber+Suhner cables. (Old design required entire cable replacement, newer design supposedly allows field repairs.)
- Failing power modules. (Limiting the output.)
- Mainly in SK Signet equipment.
- Failed payment terminals/network
- EA seems to use Nayax ones. (It looks like Tesla is going to use Payter in V4 sites where regulations require card readers.)
- Vending machines have had payment terminals forever and they don't seem to have the same failure rate...
- Signet surge: Likely a bug in Signet equipment that impacts >800v vehicles with bouncing output kW.
- If it is a bug, they should have been able to roll out a fix by now.
Again, I'm not saying that there isn't software on the charging station itself. There is plenty. And as you can see by the above, most of that is related to safety and managing the high power equipment itself. And some of those issues above are actually hardware (not software) issues.
I would think if the charger program was simple, we would not have all these FW updates and incompatibility issues. It seems to me it never ends.
The charger program is simple. The charging STATION software is certainly not. And yes, that's a huge problem (maybe a necessary evil though). Still, it seems as if the developers of this hardware and software are simply not used to consumer-facing devices and not used to making them bulletproof and ultra-reliable. I experienced my first fastcharger in 2012 and it was a real piece of garbage. The screen was absolutely unreadable in the sun, and it had an emergency off switch that looked like a regular off switch, so people would press it to stop the charge. The problem is, when this was done, someone had to come and unlock the cabinet to reset the switch before it would operate again. And that's just two human factors examples. The software that runs the payment and authentication systems on these things is similarly horrible (although fortunately back in 2012 the only vehicles that could use that station was the Nissan LEAF, and the stations were free, so we didn't have to worry about payment authorization).