Speak EV - Electric Car Forums banner

On-board charger limits power after about an hour

5K views 25 replies 9 participants last post by  KiwiME 
#1 ·
I've noticed this issue since day one of purchasing the Kona several months ago but am ready to investigate it now at a technical level and would like to hear from those owners who are lucky enough to have 32 amps at home, or otherwise fully utilise the OBC's capacity. I'm in NZ and it's been summer-like weather the entire 4 months I've owned the car. I thought this might also be relevant as the northern hemisphere moves into summer.

In summary, the OBC drops the charge power in numerous steps all the way down to 0.5 kW (then cycles up and down) after roughly an hour at 7.4 kW, 32 amps (on a 43 kW ABB EVSE.) This happens at all state-of-charge levels that I typically find myself at, between 50 and 90%. I often charge only for an hour at a time anyway (public charger near home) so it doesn't often affect me but now that it's quite warm it can be an issue in less than an hour, or if I return later for a second short session. This is the only 32+ amp Type 2 station I have good access to and that cable is rarely used by other EVs but seems to be in top working order. I can't easily test this issue on other EVSEs.

Taking temperatures around the front bay in an ambient of 25-28°C gives me 52°C on the top surface of the OBC, the fan is not running and if the water pump is running it's too quiet to hear. The hoses to the radiator are 45 and 40°C. Nothing seems abnormal but why would it throttle itself? What temperature gradient might there be between the power semiconductors and the coolant and why would the fan not start if heat was the issue? I doubt that would be more than 20 C; 75 C at the devices should be fine.

Based on a diagram of the cooling and HVAC system, the OBC is plumbed in series with the motor, drive electronics (which of course provides thermal mass,) water pump and radiator.

Yesterday I had an idea to run the heater during this adverse event, as it would use the heat pump and pull cold air through the evaporator and radiator. That didn't work because the max heater setting is 27, below the ambient.
So I tried the AC, which did run the fan but nothing obvious changed, noting of course that using AC means hot air is being pulled through the radiator, not what I want. Using a fan or spraying water over the radiator is not practical, people would gather around and ask silly questions :(

Generally if I restart the charging session the power will go back to 7.4 kW for perhaps 10 minutes before dithering down to 6.0, 3.8, 1.8, 0.5 over some minutes, then cycling up and down again.

I'm of the opinion that the OBC should be able to operate continuously and there is no reason why the it should back off regarding the Kona's battery in the ambient and SoC conditions I'm in. The EVSE is offering 63 amps and has no reason to limit that. If the OBC is overheating then the fan and pump should be working to prevent that before power throttling starts.

A message to Hyundai NZ did not unearth an explanation, responding: "It doesn't seem if anything is out of the ordinary but please keep us posted with your monitoring."
I'm not going to try the dealer unless there is evidence that my car is faulty.

One owner in the UK has offered info on FB that his has charged fine for a number of hours, but in 7°C ambient.

So, owners using full Type 2, are you able to charge for hours at a time under full power, and what is your ambient temp? To be clear, fast DC charging works fine and I can use it ... this is the AC side only in question.
 
See less See more
#2 ·
No problems here mine will pull 7.2kw on a 32amp charger rom empty to 95%

The only time I see it drop to 0. 5kw is if I open the drivers door while charging which is its programmed behaviour

You mention you are using a 43kw rapid charger to do your 7kw charging have you considered this may be the issue and try a proper 7kw home charger instead

Rapid chargers get temperamental when hot and as outputting 7ke is not the primary design of those units that would be my first investigative point.
 
#16 · (Edited)
The pilot pin on the plug is shorter than the AC pins, for the reason that if you start to unplug it without warning when it's unlocked the charge current will be shut off before the power pins break contact, so no arcing... The clever people that came up with the Type 1 and Type 2 charging standards thought of this for you already. ;)
I can't think of a good reason the charge rate should slow down to a trickle but keep going when the car is unlocked. Does any other EV do this ?
I can't think of a good reason either but I'm sure that information will eventually turn up! The connector would surely still be locked.

I've stumbled on the Type 2 connector standards as well while studying up on the J1772 protocol. What you say about the control pilot disconnecting first is true of course, however when I bench-characterised the Ratio trickle EVSE that was supplied with my Kona (for the 6/8 amp NZ market) I did notice that the response of the contactor to a loss of CP "loading" (i.e. "EV charging" state) is not quite instant, and I doubt it could drop AC power in adequate time to avoid a potential arc at the AC contacts. Obviously other EVSEs may respond faster but I believe the design intent is purely for safety, not to avoid a "hot" disconnect.

The circumstances of finding yourself in this mode of operation would be limited to cases where the "emergency" release is used on the connector lock in the midst of a session, as otherwise it's always locked when the CP state is "EV charging." I cringed when Bjorn Nyland did that on the Audi to terminate a DC session because he seemed unwilling to check the manual first (ironically, info missing in the Kona manual.)

You would have noted the Type 2 connector minimum test specifications under the category of hot unplugging are very low, 50 or 100 events if I recall compared to 50,000 plug cycles for cold, so it's not a normal mode of operation by any stretch.
 
#20 ·
I've stumbled on the Type 2 connector standards as well while studying up on the J1772 protocol. What you say about the control pilot disconnecting first is true of course, however when I bench-characterised the Ratio trickle EVSE that was supplied with my Kona (for the 6/8 amp NZ market) I did notice that the response of the contactor to a loss of CP "loading" (i.e. "EV charging" state) is not quite instant, and I doubt it could drop AC power in adequate time to avoid a potential arc at the AC contacts. Obviously other EVSEs may respond faster but I believe the design intent is purely for safety, not to avoid a "hot" disconnect.
Sorry I actually meant the proximity pilot signal not the control pilot signal. I always mix these two up unless I look them up first to remind myself.

The proximity pilot signal also has a short pin on the plug, as soon as this signal is disconnected the car will detect this and shut down it's on-board charger. As this shutdown in load is electronic it's more than fast enough to shut down before the AC pins are physically withdrawn by a slow human hand.

As you point out the contactors in the EVSE would have a slower reaction time and may not be fast enough to avoid arcing at the plug, and even if they were you would not want the contactors opening under full load and arcing themselves - they're not rated to do that on a regular day to day basis only on an emergency basis in the event of the on-board charger failing.

The loss of the proximity pilot signal will signal the car to stop charging, the car will then stop signalling the EVSE over the control pilot line which would cause the EVSE to open the contactor or the disconnecting of the control pilot line would also cause the EVSE to open the contactor.

While both control pilot and proximity pilot pins are short, I'm not sure if they're the same length or one is even shorter than the other.
The circumstances of finding yourself in this mode of operation would be limited to cases where the "emergency" release is used on the connector lock in the midst of a session, as otherwise it's always locked when the CP state is "EV charging."
Remember though that the J-1772 protocol itself was designed for a connector that didn't support electronic locking, as most early cars with Type 1 like my Ion can't lock the connector. Part of the solution for that is a trigger switch which signals to the car to stop charging, and must be pressed before the plug can be withdrawn.

Interestingly when you press this trigger, which is connected to a resistor on the proximity pilot line the car immediately reduces its AC current draw to near zero, however it keeps the control pilot line active so the EVSE contactors stay closed. On my car it also keeps the charging light on the dashboard even though charge current has essentially dropped to zero. So going by observable behaviour you would think pressing the trigger did nothing because the EVSE remained on and the car charging light is still on, but the current has actually shut off.

Then when you withdraw the plug what I described above with the two pilot lines disconnecting occurs. Of course a Type 2 connector does not have this extra resistor and trigger switch so that in the event of the plug being manually unlocked and pulled out while charging it relies entirely on the proximity pilot line to tell the car to stop charging in time before the AC pins disconnect, and the loss of the control pilot will cause the EVSE contactors to open before the plug can be removed and the pins touched by the operator.

From what I can see both Type 1 and Type 2 charging plugs and the J-1772 protocol make for an extremely safe system even when being abused. I'm quite impressed by how well thought out it is.
You would have noted the Type 2 connector minimum test specifications under the category of hot unplugging are very low, 50 or 100 events if I recall compared to 50,000 plug cycles for cold, so it's not a normal mode of operation by any stretch.
I would assume that "hot unplugging" would be in the event that the proximity pilot failed to stop the on-board charger from charging in time during a hot unplug.

However assuming that the proximity pilot does work correctly and the on-board charger does stop instantly (electronically, rather than just opening an in-car contactor) then a hot unplug is not truly a hot unplug from the perspective of the load carrying AC pins.
 
#3 ·
I wonder if it could be the rapid as well. Last night I charged my Leaf on a rapid, using the chademo, and at 85% it was down to 6A and rediculously slow.
First thing I would try is a normal fast charger and see what you get, although I assume if you are using the AC on a rapid there isn't another option near by?
No charger at home?
 
#18 · (Edited)
...First thing I would try is a normal fast charger and see what you get, although I assume if you are using the AC on a rapid there isn't another option near by?
No charger at home?
No charger at home unfortunately as I'm in a flat on a busy high street. The Kona manual stresses in numerous places to prioritise AC as preference over DC when practical and since I paid 75,000 hard-earned dollars for this car I'm going to abide by that advise, despite understanding that many owners don't :)
It's worth pointing out that the AC part of a public charger has minimal relationship to the DC side, as long as the total power draw is under the supply, 100kW in this case. I'm nearly the only user of the Type 2 on this pay-for site since there are two free ones just a half km away and all three are used more by spiders for building webs than for charging EVs!
But, as you suggested, I might get the on-board charger good and hot and then continue charging on a free Type 2 to see what happens.
 
#24 ·
I'm pretty sure my home charger is 7kw/32a and my charger shows 27kw pulled over 4 hours.
I wonder if fiddling with the AC power draw settings would help
Yes, did try that while it was charging. Noted the curiosity that one or two YouTubers commented on, the remaining charge time does not update while throttling until you change this setting. So with 3.8 kW at the Max Current settling the time still assumed it was 7.4. After changing to Reduced Current it went to 3.6 kW and updated the time accordingly. Once changed back to Max Current, the time estimate matched the throttled 3.8 kW.
 
#25 ·
Here is what I get from my charging history. I have an Andersen A2 home charger.
Thanks for posting that list. If that's a spreadsheet you can edit, a useful column to add would be total kWh divided by total time in hours, giving you the average charge rate per session in kW.

I checked the longest charge you have as an example and it's 7.04 kW, right on the money.

In my case that number is significantly lower due to this throttling issue. It's one way I can detect it after the session has completed, without having to watch the rate manually, assuming I stay under about 95% SoC.
 
#12 · (Edited)
And if you're paying a connection fee or minimum fee for charging, and returning to the car to get your wallet causes you to have to fiddle with an app to start a new charge (ecotricity etc) and/or have to pay a second connection fee, how is that good design ?

I'm happy my Ion will continue charging when the doors are opened/closed/locked/unlocked... I'd rather have a charge cancel button which stops the charge and unlocks the plug than have it tied into opening and closing of doors.
 
#26 · (Edited)
I charged for an hour on my regular ABB Terra 54 (the Type 2 cable rated 43 kW 3-ph) to warm it up and verify that the issue continues.

81% starting SoC, 22°C.
Started at 7.4 kW as it should, monitoring on the app ...
... after 31 minutes dipped down to 6.5 for a minute, back up to 7.0 to 7.2 kW
... 39 min, 4.5
... 42 min, 4.8 - 4.9
... 46 min to 60 min, back up to 7.0-7.1
Stopped at 60 min, average rate 7.0, not horrible but there was no reason for it to be dithering around.
Noting on most other occasions it would dither between 1.8 and 3.8, once up to 55°C.

Drove over to the free Type 2 EVSE (rated 22 kW 3-ph) right outside the big box.
Charged for 40 minutes, checked the dash few times (no app of course) and it was rock-solid at 7.1 kW every time.
I also noted that the drop to 0.5 thing did not seem to happen when I opened the door.

I'm going to gather more data before I contact ABB and/or Hyundai (again) because I would like to know what's happening, even if it's not fixable. This is my closest charger, 1 minute walk from home, the car is safe there and I don't mind the cost, so I would like it to work properly.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top