Model 3 scheduled/timed/smart charging

[Glan gluaisne]

Thought I'd try and make an attempt to keep all the various mentions of problems with scheduled/timed/smart charging in one thread, as at the moment they are mentioned in several other threads, but none have titles that refer specifically to the issues being seen.

First off, my (limited) experience suggests that the M3 fails to wake up and charge overnight, if the charge is being controlled by the charge point. My charge point is (rather was until yesterday - more later) 100% compliant with the internationally agreed protocols, IEC62196 (for the connector interface) and IEC61851 and JAEJ1772. The key point is that, for the charge point to be able to control charging, the vehicle must respond when the charge point advertises that power is available, and must set its own onboard charger (OBC) such that the advertised current from the charge point is not exceeded.

The normal sequence of operation should go like this (ignoring fault conditions and states):

1. With the car unplugged, and the charge point in a operable state (i.e powered on, but not necessarily ready to provide charge power) the signal wire connection to the car connector, the Control Pilot (CP) should be sitting at a steady +12 VDC. In this state, there should be no power on any of the other connector pins.

2. When the connector is plugged into the car, the car has a fixed resistance of 2.74 kΩ between the CP pin and the Protective Earth (PE) pin. Inside the charge point, the CP signal is fed through a 1 kΩ resistance, so this load to PE results in the voltage at the CP being pulled down to +9 VDC. The charge point detects this change, and, only if it's ready to supply power, it should start signalling the amount of current that is available by putting a 1 kHz pulse train on the CP, with a source voltage (before the 1 kΩ resistor) of +12 V to -12 V. The duty cycle normal range for 230 VAC single phase charging in the UK is from 10% (indicates up to 6 A is available) to 50% (indicates up to 30 A is available).

3. The car accepts the advertised charge current by loading the CP signal down with an additional parallel resistance of 1.3 kΩ, which reduces the positive-going part of the CP signal to +6 V. There is a diode in series with the CP where it comes into the car, so the negative-going part of the CP pulse train remains at -12 V. This is a safety mechanism, used by the charge point to detect faults, such as a connector being dropped into a conductive liquid.

4. The charge point detects this drop in the positive-going part of the CP signal, and as long as it doesn't detect any other fault condition, like excessive earth leakage, it will close a contactor (the thing that makes a click) and start supplying power to the car. The car may slowly ramp up the charge current initially, until the limit set either by the charge point CP duty cycle, or a limit set within the car, is reached.

5. If the car wishes to terminate the charge, it removes the 1.3 kΩ resistance loading the CP signal. The charge point detects this, and turns off the power to the car. The charge point should continue to advertise the available current, by maintaining the set duty cycle on the CP, until the connector is removed from the car.

6. If the charge point wishes to terminate the charge, then it just switches off the 1 kHz pulse train on the CP and takes the drive for the CP to +12 VDC (this will still be loaded down at the car end). The car detects the loss of the CP pulse train and gracefully shuts down charging. When it has done this it removes the 1.3 kΩ load resistance, which takes the CP from +6 VDC to +9 VDC. The charge point responds to the change in CP voltage by turning off the power, by opening the contactor (the contactor should open under virtually no load conditions).​

The protocol allows for a charge point being operable, but not ready to supply power, to allow for timed charging, sharing of charge points from a common supply (where cars may have to charge in sequence to avoid overloading the supply), or to allow for a charge point varying the available maximum current through a charge period. If a charge point is connected to a car, but not ready to supply power, then it should just sit with the CP at +12 VDC, which the car should interpret as waiting for further commands from the charge point. When the charge point is ready to supply power it just switches on the 1 kHz pulse train and advertises the available current to the car, as in step 2 above. The car should then immediately respond by loading the CP as in step 3 above, and charging as normal.

If the charge point needs to vary the maximum current available, perhaps due to a supply loading restriction, or to maximise use of self-generated electricity, then it can just change the duty cycle of the 1 kHz pulse train during charging, within the limits of the standard, and that set by the current carrying capacity of the cable fitted to the charge point. The car is required to adjust its current consumption so that it always remains within the current that is advertised as being available.

For most cars, this works fine. Both my previous plug in cars responded correctly, and would start and stop charging on request from the charge point and would vary charge current in line with the charge point advertised maximum current availability.

The Tesla doesn't seem to do this. If it is plugged in, set to charge immediately, and the charge point doesn't charge straight away, the car seems to go to sleep after a while and then fails to wake up and start charging when the charge point starts signalling that power is available. This problem seems to have been first reported to Tesla in 2018, as far as I've been able to find out, yet seems no closer to being resolved. For those who wish to make best use of any "time of use" electricity tariff, with their charge point doing charge control, then this is a major, and costly, inconvenience.

In addition, there seem to be reports that if the car is used to schedule charging, it fails to reliably charge at the current being advertised by the charge point. This may well be related to the failure to wake up and charge problem, as it seems as if it may be connected to the way the car responds to the advertised maximum available current if it is woken up specifically to charge.

Tesla seem to have little to say, other than a vague promise from around a year ago that they were looking to address the reported problem via a software update. It seems we're still waiting for this, around a year later.

As an added data point, I made some measurements on a Tesla UMC yesterday, to find out how that signalled to the car, and whether or not it followed the standard protocol. It seems that it doesn't, in that when it detects the plugged in state, instead of advertising the maximum available current (set by the type of adapter lead plugged into the inlet end) it initially sets a duty cycle of about 3.2% on the CP. This is supposed to indicate that the charge point is really a DC fast charger, and tell the car to use digital communications, rather than the loading of the CP with resistors as a signalling method. After 5 seconds, this non-standard CP signal is replaced with one advertising the maximum available current in the normal way.

It's unclear why the UMC does this. It may be that it's a legacy signalling standard, so that older models can work with the charge point, or it may be that this non-standard CP signal is a way of waking the car up from a deep sleep condition. To test this I've modified the firmware in my charge point, so that it exactly emulates the Tesla UMC, and starts the CP signalling, when ready to charge, with 5 seconds of the 3.2% duty cycle CP signal. So far I've tested this and found it works OK when the car is already awake, what I next need to do it find out if it works when the car is in a deep sleep. Right now I'm waiting for that to happen, as the car was woken up a while ago with a software update, so my guess is that I need to wait for a couple of hours for it to go back into a deep sleep. I'm planning to try and wake the car up from the charge point after lunch, to see if providing this oddball CP signal holds a clue as to the failure to charge problem.

 

[Cwmwd]

It's unclear why the UMC does this

Might be to do with the way Tesla (miss)uses the Type-2 connector on the S and X to do DC charging on a Supercharger over the AC pins?

 

[Glan gluaisne]

Might be to do with the way Tesla (miss)uses the Type-2 connector on the S and X to do DC charging on a Supercharger over the AC pins?

Others (in other threads) have suggested that this may be a hang over from a legacy signalling system, part of the non-standard Tesla implementation of digital signalling using the CP.

I can understand this, based on the fact that back when Tesla first started making cars the standards weren't that firm, or at least didn't cover as many charge control options as Tesla wanted to implement. What seems odd is that the second generation UMC initiates CP signalling in this way. It seems that the M3 most is probably only capable of using the Tesla non-standard digital signalling protocol when supercharging, not when charging using the UMC.

There must be a reason for this oddball protocol being used in the UMC. Be interesting to see whether it's just because the UMC uses the same circuit board as the TWC. If it does, then it may be that this is just a dead feature as far as the UMC is concerned, that is only there because it may be needed if the board is fitted to a TWC.

Part of me hopes that this unusual feature may wake the car up from a deep sleep, as for me that would allow me to use timed and variable rate, charging under the control of the charge point.

 

[Adopado]

Part of me hopes that this unusual feature may wake the car up from a deep sleep, as for me that would allow me to use timed and variable rate, charging under the control of the charge point.

If this turns out to be the case then you may be able to share (sell?) your "intellectual property" with the charge point manufacturers! Certainly Myenergi who make the Zappi are aware of the issue and are currently still waiting for Tesla to solve it.

 

[Glan gluaisne]

An update.

I've now tested to see if emulating the oddball protocol used by the UMC allows the charge point to wake up the car from deep sleep and start charging.

The bad news is that it doesn't. The car sits there fat dumb and happy, ignoring the advertised charge power from the charge point.

The next stage is either to emulate a "wake up" protocol I've seen documented in a data sheet from Phoenix Contact, or to emulate a power outage. The Phoenix Contact data doesn't relate to Tesla, but does look as if it may be intended to address wake up issues, as it's specifically referred to as a "wake up sequence":

It's possible that this is contained within the current version of IEC61851, but I don't have the current version, and it costs ~£300 to buy, as I no longer have free access to standards (maybe someone here does? @arg perhaps?).

This is dead easy to implement, just a few minutes changing and testing some code, but I need to wait until it stops raining, as I don't fancy messing around with the charge point in the pouring rain (makes me wish I'd designed-in wireless firmware flashing).

If this doesn't work, the next stage would be to emulate a power outage, to see whether or not the car does behave as the manual suggests or not. I suspect it won't, but again it's an easy thing to test, with no changes at all, as for test purposes I could just throw the isolator switch on the charge point circuit whilst the car is charging, wait a few hours for it to enter a deep sleep, then throw the switch back on and see what happens. Again this will have to wait until the rain stops, as the isolator switch is contained within a weatherproof enclosure outside, and I don't much fancy opening that up in the pouring rain, either.

 

[Jason71]

An update.

I've now tested to see if emulating the oddball protocol used by the UMC allows the charge point to wake up the car from deep sleep and start charging.

The bad news is that it doesn't. The car sits there fat dumb and happy, ignoring the advertised charge power from the charge point.

The next stage is either to emulate a "wake up" protocol I've seen documented in a data sheet from Phoenix Contact, or to emulate a power outage. The Phoenix Contact data doesn't relate to Tesla, but does look as if it may be intended to address wake up issues, as it's specifically referred to as a "wake up sequence":

View attachment 486774

It's possible that this is contained within the current version of IEC61851, but I don't have the current version, and it costs ~£300 to buy, as I no longer have free access to standards (maybe someone here does? @arg perhaps?).

This is dead easy to implement, just a few minutes changing and testing some code, but I need to wait until it stops raining, as I don't fancy messing around with the charge point in the pouring rain (makes me wish I'd designed-in wireless firmware flashing).

If this doesn't work, the next stage would be to emulate a power outage, to see whether or not the car does behave as the manual suggests or not. I suspect it won't, but again it's an easy thing to test, with no changes at all, as for test purposes I could just throw the isolator switch on the charge point circuit whilst the car is charging, wait a few hours for it to enter a deep sleep, then throw the switch back on and see what happens. Again this will have to wait until the rain stops, as the isolator switch is contained within a weatherproof enclosure outside, and I don't much fancy opening that up in the pouring rain, either.

Is "deep sleep" just a figure of speech? or is there really such a thing as "deep" sleep. I thought there was only sleeping or not sleeping. There is certainly not an API response of "deep sleep".
If you want to know if the car is asleep you don't necessarily have to wait hours. it is possible to make an API call that will tell you if it's asleep without actually waking it if you want to be sure of the state.

 

[Glan gluaisne]

Just going on what's been reported/mentioned in various posts on different forums, some of which mention a deeper sleep mode that the car enters after a period of time (~2 hours has been mentioned). This could be just inaccurate observation, or speculation, though.

All I've noticed is that if sentry mode is off, the car seems to disconnect the main battery contactor a short time after it's locked (it's possible to hear the contactor opening with a fairly loud clunk). With the HV battery isolated, then presumably the car relies on the 12 V battery for comms etc, as it seems it can still be partially awake with the contactor open.

At some point I'd guess that it has to go into some sort of power saving mode, where it minimises draw from the relatively small 12 V battery, so most probably shuts down as much stuff as possible. I'd assumed that this is the "deep sleep" mode, but I could well be wrong.

If someone knows for sure how the car powers down, and when each power down stage happens, that could be useful. At some point the car clearly powers down the charge port controller, and stops it responding to charge point requests. It may be that if I knew how long that period of time is, I could adapt my charge point code so that when it's in timed charge mode it just periodically starts a charge for a minute or two then stops it, just to keep the charge port controller awake.

 

[M3BlueGeorgia]

Just going on what's been reported/mentioned in various posts on different forums, some of which mention a deeper sleep mode that the car enters after a period of time (~2 hours has been mentioned). This could be just inaccurate observation, or speculation, though.

All I've noticed is that if sentry mode is off, the car seems to disconnect the main battery contactor a short time after it's locked (it's possible to hear the contactor opening with a fairly loud clunk). With the HV battery isolated, then presumably the car relies on the 12 V battery for comms etc, as it seems it can still be partially awake with the contactor open.

At some point I'd guess that it has to go into some sort of power saving mode, where it minimises draw from the relatively small 12 V battery, so most probably shuts down as much stuff as possible. I'd assumed that this is the "deep sleep" mode, but I could well be wrong.

If someone knows for sure how the car powers down, and when each power down stage happens, that could be useful. At some point the car clearly powers down the charge port controller, and stops it responding to charge point requests. It may be that if I knew how long that period of time is, I could adapt my charge point code so that when it's in timed charge mode it just periodically starts a charge for a minute or two then stops it, just to keep the charge port controller awake.

Seems like the bigger issue is you have two smart devices not cooperating because each expects a dumb connection on the other end.
Can you set the Chargepoint to be dumb? Just simply provide power all the time and let the car have control.
Otherwise, sell the Chargepoint and replace with a dumb connection...

 

[Cwmwd]

mention a deeper sleep mode that the car enters after a period of time (~2 hours has been mentioned)

I think this might be "Idle" and then "Sleeping" after a while? The TeslaFi Knowledge Base has an article about how polling the car can be done without waking it but you only get a limited set of data back:

Only a few data points are returned from this call. The most useful point is the online state of the vehicle such as online/offline/asleep/waking/unknown. The important part of this call is that it can be called without affecting the vehicle’s state. This means it can be called and it won't prevent the vehicle from sleeping and it won't wake the vehicle if it's asleep.

Enabling sleep settings to limit vampire loss. / Knowledge base / TeslaFi

 

[Jason71]

Just going on what's been reported/mentioned in various posts on different forums, some of which mention a deeper sleep mode that the car enters after a period of time (~2 hours has been mentioned). This could be just inaccurate observation, or speculation, though.

All I've noticed is that if sentry mode is off, the car seems to disconnect the main battery contactor a short time after it's locked (it's possible to hear the contactor opening with a fairly loud clunk). With the HV battery isolated, then presumably the car relies on the 12 V battery for comms etc, as it seems it can still be partially awake with the contactor open.

At some point I'd guess that it has to go into some sort of power saving mode, where it minimises draw from the relatively small 12 V battery, so most probably shuts down as much stuff as possible. I'd assumed that this is the "deep sleep" mode, but I could well be wrong.

If someone knows for sure how the car powers down, and when each power down stage happens, that could be useful. At some point the car clearly powers down the charge port controller, and stops it responding to charge point requests. It may be that if I knew how long that period of time is, I could adapt my charge point code so that when it's in timed charge mode it just periodically starts a charge for a minute or two then stops it, just to keep the charge port controller awake.

When not in use mine goes to sleep typically after about 15 minutes. after that all you can get from the api is "asleep" there is a more detailed api but to use that you have to wake it so it's a bit of a schrodinger's cat situation. Not saying there is not a deeper sleep but you clearly can't find anything out about it via the API's which is why I was not aware of even the rumour of its existence.

 

[Glan gluaisne]

Seems like the bigger issue is you have two smart devices not cooperating because each expects a dumb connection on the other end.
Can you set the Chargepoint to be dumb? Just simply provide power all the time and let the car have control.
Otherwise, sell the Chargepoint and replace with a dumb connection...

For scheduled charging with a dumb charge point, only during the 7 hour, overnight, off-peak electricity period then the in-car system fails. A charge start time can be set, using the scheduled charging option, but a charge stop time cannot be also set. This means the car can start charging when the off-peak period starts (midnight during winter here), but the car will carrying on charging until it's fully charged, and cannot be set to switch off at 07:00, the end of the off-peak electricity period.

An alternative is to set the car to a departure time, which will stop it charging at the end of the off-peak period, but it may well start charging way earlier than the midnight start of the off-peak rate.

Ideally, I'd like either the car to do as my charge point does, and accept a start charging time and a stop charging time, so it only ever charges during the off-peak period, or I'd like the car to just follow the internationally agreed protocol and let the charge point control when the car charges and how much charge current it uses.

The latter is a mandatory requirement here for all charge points installed with a government grant, so the fact that the car refuses to comply with the way these smart charge points are supposed to work negates the reason that the government have mandated they be installed, part of which is to allow charge points to be remotely controlled so as to balance the grid, if needed. If this starts happening and cars refuse to accept charge because Tesla have screwed up the way that the car's charge port controller works, then I can see people being a bit unhappy (and there are already a few people that are unhappy about the car refusing to accept a charge point scheduled charge).

Tesla know of the problem, have known about it since the end of 2018, and have apparently said there will be a fix, but it's now a year on and no fix seems to be forthcoming.

 

[Glan gluaisne]

When not in use mine goes to sleep typically after about 15 minutes. after that all you can get from the api is "asleep" there is a more detailed api but to use that you have to wake it so it's a bit of a schrodinger's cat situation. Not saying there is not a deeper sleep but you clearly can't find anything out about it via the API's which is why I was not aware of even the rumour of its existence.

One experiment I could try is to start a charge, lock the car, then stop the charge from the charge point. If I then try to restart the charge at varying time periods after the charge has stopped I should be able to narrow down the time taken for the car to sleep, or at least enter whatever sleep mode disables the charge port controller.

Bit of a PITA to do, but might be useful, if only to get a firm idea as to how the car shuts systems down.

Doing this experiment in conjunction with the info in this post, might be even more useful:

I think this might be "Idle" and then "Sleeping" after a while? The TeslaFi Knowledge Base has an article about how polling the car can be done without waking it but you only get a limited set of data back:



Enabling sleep settings to limit vampire loss. / Knowledge base / TeslaFi

 

[M3BlueGeorgia]

For scheduled charging with a dumb charge point, only during the 7 hour, overnight, off-peak electricity period then the in-car system fails. A charge start time can be set, using the scheduled charging option, but a charge stop time cannot be also set. This means the car can start charging when the off-peak period starts (midnight during winter here), but the car will carrying on charging until it's fully charged, and cannot be set to switch off at 07:00, the end of the off-peak electricity period.

An alternative is to set the car to a departure time, which will stop it charging at the end of the off-peak period, but it may well start charging way earlier than the midnight start of the off-peak rate.

Ideally, I'd like either the car to do as my charge point does, and accept a start charging time and a stop charging time, so it only ever charges during the off-peak period, or I'd like the car to just follow the internationally agreed protocol and let the charge point control when the car charges and how much charge current it uses.

The latter is a mandatory requirement here for all charge points installed with a government grant, so the fact that the car refuses to comply with the way these smart charge points are supposed to work negates the reason that the government have mandated they be installed, part of which is to allow charge points to be remotely controlled so as to balance the grid, if needed. If this starts happening and cars refuse to accept charge because Tesla have screwed up the way that the car's charge port controller works, then I can see people being a bit unhappy (and there are already a few people that are unhappy about the car refusing to accept a charge point scheduled charge).

Tesla know of the problem, have known about it since the end of 2018, and have apparently said there will be a fix, but it's now a year on and no fix seems to be forthcoming.

Wow, does your car normally charge for over 7 hours a night?
I'm only charging on a 240v/30A socket, allowing 24A continuous for a charge rate of 22mph, but it is extremely rare to run outside my optimum charge window of midnight to 6am. If I think it might, but don't need to fill the battery that night, I'll set the target percent to a lower number and complete the charge the next night.

Could you set the Chargepoint to provide power from (say) noon to 7am and then use the car to start charging at the time you tell it, but the Chargepoint to stop it at 7am?

 

[Glan gluaisne]

Wow, does your car normally charge for over 7 hours a night?
I'm only charging on a 240v/30A socket, allowing 24A continuous for a charge rate of 22mph, but it is extremely rare to run outside my optimum charge window of midnight to 6am. If I think it might, but don't need to fill the battery that night, I'll set the target percent to a lower number and complete the charge the next night.

Could you set the Chargepoint to provide power from (say) noon to 7am and then use the car to start charging at the time you tell it, but the Chargepoint to stop it at 7am?

It can do, if I've done a long trip. Happened once so far, where charging over ran for two hours into the peak electricity price period.

I can do the work around where I use the car to schedule the start of charge and the charge point to schedule the end, and that works for the winter, when almost all charging will be overnight at the cheap rate.

What I want to be able to do is get back to what I've been doing for the past few years with other EVs/plug-in hybrids, and have total control of charging via the charge point. If I can crack this, then it will allow me to use flexible charging during periods when we are generating enough electricity from the solar panels, something that probably accounted for around half my annual driving this past year (in a little BMW i3). To do that means getting the car to always be subservient to the charge point, within the hierarchy of control. This is supposed to be something that car has to do to be compliant with the standard, anyway, and Tesla know of the non-compliance and have supposedly been working on a fix for over a year now.

One reason it's needed is because, sooner or later, all the smart charge points that are being installed here will be able to control grid demand by throttling back, or switching off, EV charging during peak demand periods, so as to keep the national grid balanced. A couple of days ago one energy company started paying consumers to use electricity, believe it or not, as the national grid was awash with wind generation and needed consumers to increase demand. Getting EV owners to charge their cars was singled out for thanks by a manager at the national grid, in this message:

 

[pdk42]

This is a very interesting thread. I've learned a lot. I've been struggling with problems on my home charger (unrelated to this thread), but have also had the problem where occasionally the car (using scheduled start) will charge at a much lower current than available - I think this is what @Jeremy Harris is referring to when he says:

In addition, there seem to be reports that if the car is used to schedule charging, it fails to reliably charge at the current being advertised by the charge point. This may well be related to the failure to wake up and charge problem, as it seems as if it may be connected to the way the car responds to the advertised maximum available current if it is woken up specifically to charge.

I'd also like to get to the point where I can control the charging entirely from the charger. The charger I've got is pretty crap (an EVBox - ignorance when first looked at chargers!), but they have recently added the capability of controlling start/end times.

 

[Val33]

I have an Ohme charger cable. This works seamlessly with Octopus Agile and the M3. The other night it charged the car in three separate time blocks to make use of the cheapest available tariff. I'm not sure how it overcomes the problem you are experiencing, but you do need to link it to your Tesla account so I guess it may be using the API to wake the car?

 

[Glan gluaisne]

I have an Ohme charger cable. This works seamlessly with Octopus Agile and the M3. The other night it charged the car in three separate time blocks to make use of the cheapest available tariff. I'm not sure how it overcomes the problem you are experiencing, but you do need to link it to your Tesla account so I guess it may be using the API to wake the car?

It must be waking the car up, as it seems pretty clear that the car refuses to accept any charge power available advertisement from the charge point if it happens to be in sleep mode.

The protocol requires the car to start charging when the charge point starts to advertise that charge power is available, with no maximum time limit between the charge cable being plugged in and the charge point initiating a charge.

 

[MidhurstNigel]

It must be waking the car up, as it seems pretty clear that the car refuses to accept any charge power available advertisement from the charge point if it happens to be in sleep mode.

The problem seems to be something other than the awake/asleep status when charging starts. I have for a while been using Teslafi to wake my M3 up at 0025 and setting my Zappi 2 to start charging at 0030 (and stopping at 0430). This has worked fine with a full 32A charge throughout the charging period (or until limit of SOC) until a couple of nights ago (the last charge done) when it only charged at 16A throughout the Zappi on period. I think this was the night after the last software upgrade, so maybe that upset it. I'll now go back to having Teslafi stop charging at 0035 and restarting at 0040 (as well as the 0025 wake-up) to see if that kicks it back upto 32A

 

[Glan gluaisne]

The problem seems to be something other than the awake/asleep status when charging starts. I have for a while been using Teslafi to wake my M3 up at 0025 and setting my Zappi 2 to start charging at 0030 (and stopping at 0430). This has worked fine with a full 32A charge throughout the charging period (or until limit of SOC) until a couple of nights ago (the last charge done) when it only charged at 16A throughout the Zappi on period. I think this was the night after the last software upgrade, so maybe that upset it. I'll now go back to having Teslafi stop charging at 0035 and restarting at 0040 (as well as the 0025 wake-up) to see if that kicks it back upto 32A

I've been wondering if the reports of charge current dropping when using scheduled charging are related to the refusal to start charging when the car is asleep or not. From your description, it seems as if the charge current dropping issue may be another software bug, perhaps unrelated to the refusal to charge bug.

With more and more people charging using smart chargers, and also using smart electricity tariffs, this is something that Tesla really need to pull their finger out and fix, IMHO, as it seems likely that there will be more owners affected by these problems with every passing day, as taking advantage of time of use (ToU) tariffs becomes more popular (I've been charging during the E7 off-peak period for 6 years now, with no problems at all until I bought the M3).

I don't know of any other EV that has such flaky charge performance. I know that ToU tariffs are commonplace in the US, that some areas there have similar smart metering systems to the UK, and that smart charge points, that can control charging, are not just a UK thing, so, with the number of cars sold so far I'm really surprised that there aren't far more complaints about this issue. The earliest complaints about this I've found date back to around November 2018, with Tesla stating that they were aware of the problem and working on a fix. A year later and nothing at all seems to have been done, and anecdotally it seems that some things (like the reduced charge rate problem) may even have been made worse.

 

[arg]

It's possible that this is contained within the current version of IEC61851, but I don't have the current version, and it costs ~£300 to buy, as I no longer have free access to standards (maybe someone here does? @arg perhaps?).

That "wake up sequence" isn't mentioned in IEC61851-1:2010/EN61851-1:2011 (which is current), nor in SAE J1772:OCT2012 (I don't have the 2017 update of J1772), however it is of course perfectly legal behaviour within the standard's rules: the chargepoint is simply going to state F ("EVSE not available") for three seconds, then declaring itself available again.

I'm guessing this is just a workaround for some manufacturer's bug. I seem to recall the original GM Volt having issues around the order in which you plugged things in, which could be the source for the Phoenix Contact note.

The Model 3 failing to wake up is undoubtedly a bug - Models S/X have always implemented the proper behaviour.

I suspect that the 16A/32A behaviour is a separate bug, somewhat specific to the UK (as we have the three-phase capable inlets, unlike the US, but have common use of high-current single phase, unlike most other parts of Europe).