DCFC - Slower than expected - 2018 SV

I have reasons to believe that DCFC is not occuring at the maximum possible speed.

I have had the software update regarding rapidgate (NTB19-056) applied this summer. (How can I confirm this?) I have not had the chance to comprehensively test the update ever since (i.e. trace a propre full charge curve with ideal starting conditions).

Different chargers were used (Quebec region):
- Electric Circuit 50kW and higher (AddÉnergie),
- ChargePoint 125kW,
- Electric Circuit 175kW CHAdeMO (300kW CCS, ABB)

When multiples chargers were present, I am only considering charging sessions where no one else were using the other units simultaneously (i.e. if power sharing scheme were applied, they wouldn't be in effect). I've sometimes had erratic charging speeds (random variations vetween 20kW and 26kW) which I've attributed to power sharing schemes; I am disregarding these experiences for this investigation.

Pack temperature were between 0-15C (32-59F), at start (multiple different attempts). (Fall/Winter in Quebec). I am only considering single charges on a day with little to no prior driving. I am not considering multiple successive charges.

A typical charge with the above criterias would go as follow: initial ramp up to 125A (maximum speed). At 38-40%, charging speed started throttling. Throttling would continue slowly with a decay shape until 50% is reached. Then the throttling is linear and decreases steadily to 20kW.

I am knowledgeable in CC-CV charging curves for lithium packs and electricity notions (electrical engineer). I have compared my experiences with typical Leaf packs (such as this one for instance). I am expecting the knee point to be reached slightly before 60% for a typical, throttle-free charge. In my case I observe this happening more around 40%.

I am running out of ideas and am wondering if it would be worth it involving the dealer. At this point I am questioning if the dealer's intervention (NTB19-056) has anything to do wth it (i.e. the parameters were set correctly in the charging system, if that's how it's done in the NTB...)).

Any insight would be appreciated. Thanks.

Picture is a one of the charge where charging was slow. Disregard gaps as I walked away from the car or forgot to keep the car on.

Pack temperature effects I presume, although I don't have any insight into how battery degradation might also be in play.

Nothing to do repair wise.

What are the amperage and voltage readings as displayed by the Chademo units while charging--that is data you also need to record to know the situation.

Also can you really make any conclusions with connector issues and dropouts in the data, plus who knows which mode the car is getting with the key switch being operated while doing a DCQC. What is the effect of re-starting an interrupted session, might that cause a less-than-maximum condition?

Can a EE really trust data which uses an imaginary made up unit of measure known as gids? The dealer's Consult Tool probably doesn't measure such a unit.

nlspace said:

What are the amperage and voltage readings as displayed by the Chademo units while charging--that is data you also need to record to know the situation.

Click to expand...

Here's a graph of the charger's values (below). The numbers correspond except when accessories were used (last part, annotated).

nlspace said:

Also can you really make any conclusions with connector issues and dropouts in the data, plus who knows which mode the car is getting with the key switch being operated while doing a DCQC.

Click to expand...

Indeed, I figured toying with the car while charging wouldn't allow meaningful conclusions. For this charge, the car was left unpowered and without any accessory, minus the last part. About the last part where heaters were activated: I had already done most of the charge by then and the car had not been touched, saved for physically opening the door and entering it. At this point my test was over; I just carried on recording and activated the heaters for my own sake. Also of interest is that the charger's output was unaffected.

nlspace said:

What is the effect of re-starting an interrupted session, might that cause a less-than-maximum condition?

Click to expand...

Whenever I re-started a session, regardless of the cause of interruption, the car always seems to go and fetch the highest power possible (as expected), but quickly drops down (decaying curve - like at the 2nd restart mark). (As long as it is in the constant-current phase, obviously.)

I'm aware that there is a lot of other factors being included and I'm doing efforts in ruling them out. I had planned this session specifically for recording it properly since I knew otherwise the results would be biased. In this case, I was not expecting session disconnects and I don't know how to keep LeafSpy alive and recording without intervening. I just open the unlock the door which keeps it alive in an attempt to minimize interaction with the car.

My next step would be to remove the OBDII dongle altogether but then I would have no metrics except from the charger's graph.

nlspace said:

Can a EE really trust data which uses an imaginary made up unit of measure known as gids? The dealer's Consult Tool probably doesn't measure such a unit.

Click to expand...

I am not but I find interesting the efforts of community to reverse engineer their cars, and how Gids came up as a unit.

Graph:

SageBrush said:

Pack temperature effects I presume, although I don't have any insight into how battery degradation might also be in play.

Click to expand...

In the summer I have experience better charges at temperature levels under rapidgate - in this case would it be because of too low temperatures then?

Now that's a nice looking graph. The charging started at nearly 50kW, which may be the maximum of the Chademo station. The power was immediately pulled back at a fairly steep slope and gradually flattened until about 38% and then it went to a nearly linear rampdown.

Cold temperatures are a first order factor to determine charge rate. Chemical reaction rates are slower.

What do your cell voltages look like when doing a DCQC--are they evenly balanced or do you have any jumping out above or below the average? That could also be a major factor in charge rate.

Noticed some relatively weak charging curves lately during DCFC sessions on my 40KWh. Today was one of those sessions and had a similar charge curve as the last few times.

Basically ramps up quickly & peaks at roughly 47KW, holds that plateau until reaching @ 30% SOC, then tapers quickly. By 50% SOC it's tapered to just under 30KW, then steadily tapers to @ 16KW by 80% SOC. So basically peaks as expected, then begins the taper at a relatively low SOC. Today's session was plugged in at 12% SOC and unplugged at 92% SOC @ 8KW power. Total session was 29KWh dispensed in 60 minutes (80% of pack capacity). Ambient temps have been in the 40'sF so the pack isn't excessively cold and far from overheated (5th gauge segment to 7th segment when charged).

Starting to wonder if this is the programming of car, or maybe the local ChargePoint station I frequent when needing to DCFC.

My 2013 SV charges even slower than your 2018 SV at this time of year. I'm no expert, but I've always assumed that internal resistance increases as the pack ages and that's the main driver of sustainable charging rate (other than ambient temperature).

rogersleaf said:

Starting to wonder if this is the programming of car, or maybe the local ChargePoint station I frequent when needing to DCFC.

Click to expand...

I can say for certain that not all DCFC are created equal. I can take daily identical battery conditions to pit several network DCFC against each other (Blink, EA, Nissan HQ) in the same area only a few miles apart and some are more aggressive than others when charging given identical battery SOC, temperature, readings, etc. In my area, EA has the most aggressive DCFC. When Blink, Nissan HQ get past the 80% SOC mark and really ramp down the charging power, EA will keep it high all the way to 100%. So imagine that at 95% SOC, most DCFC are probably down to the 6kW to 8kW of charging power, the EA station will still be pushing 20kW to 30wK all the way to 100% SOC. I've done multiple experiments with my 2013, 2018, 2020 Leaf, the results are always the same in my area, the EA chargers are way more aggressive in power.
I know technically, it shouldn't be that way as the Leaf is suppose to be "telling" the DCFC how much power to send, but I don't think the Leaf has any safety protocols against too much power, seems to be more of a "at least this much power" protocol.

I've always thought of L3 charging as having the charger outside of the car and the charger being directly connected to the battery terminals. As such, the car can report voltage, battery temperature, etc but the charger is in control of the voltage and current applied and it's outside the car.

I suppose at some point the car could open a relay to stop the process but I don't know if that is feasible when the CHAdeMO is trying to send a large amount of DC current into the battery.

nlspace said:

Can a EE really trust data which uses an imaginary made up unit of measure known as gids? The dealer's Consult Tool probably doesn't measure such a unit.

Click to expand...

Silly isn't it, using the last name of one of the individuals who first accessed Leaf battery data, i.e the present state of battery energy (kWh),
from the Leaf's BMS. As if some new battery characteristic/parameter was discovered. Typically, battery capacity is represented by using Ah,
especially for Li Ion technology where the voltage delta is relatively flat as the battery discharges, and is the more common engineering term.

illidotane said:

I am not but I find interesting the efforts of community to reverse engineer their cars, and how Gids came up as a unit.

Click to expand...

The use of the Leaf term Gids surely provided a good laugh for both Elon & JB Straubel when they first heard it used, as if Nissan provided
some significant discovery/aspect to EV technology.

For me, the parameter of interest is kWh and honestly, I don't know how - or if - that is different than a 'GID'. While a Li cell does have a fairly flat voltage discharge profile, it is less so when approaching the lower limit. Ie, the 'fuel gauge' is most non-linear where you need it most - when getting close to 'empty'.

Does using GID's help with this? I'm not sure since I don't know how a GID is defined. If there was a good way to measure remaining kWh that would be my preference, even if Ah is the preferred engineering metric.

I bet Mr Gid also gets a kick out of folks who use the term. In my biz there have been a few terms that took on the names of their inventors. They were mostly just used within the company but it's somewhat a badge of engineering honor to have some effect, or curve, or blip named after you. Kalman filters are an example that has taken on widespread usage but there are other related things that are more localized.

alozzy said:

My 2013 SV charges even slower than your 2018 SV at this time of year. I'm no expert, but I've always assumed that internal resistance increases as the pack ages

Click to expand...

That's correct.

alozzy said:

and that's the main driver of sustainable charging rate (other than ambient temperature).

Click to expand...

This provides a more comprehensive discussion, besides just internal resistance, of the actual heat sources during charging/discharging;
https://mynissanleaf.com/viewtopic.php?f=34&t=27600&hilit=battery+research&start=390#p585497

goldbrick said:

For me, the parameter of interest is kWh and honestly, I don't know how - or if - that is different than a 'GID'. While a Li cell does have a fairly flat voltage discharge profile, it is less so when approaching the lower limit. Ie, the 'fuel gauge' is most non-linear where you need it most - when getting close to 'empty'.

Does using GID's help with this? I'm not sure since I don't know how a GID is defined. If there was a good way to measure remaining kWh that would be my preference, even if Ah is the preferred engineering metric.

I bet Mr Gid also gets a kick out of folks who use the term. In my biz there have been a few terms that took on the names of their inventors. They were mostly just used within the company but it's somewhat a badge of engineering honor to have some effect, or curve, or blip named after you. Kalman filters are an example that has taken on widespread usage but there are other related things that are more localized.

Click to expand...

Here's some history; https://mynissanleaf.com/viewtopic.php?t=19144#p412068

Gid is just a Leaf term for battery energy used continually by a few who still live in the Leaf intro era of 2010/2011.

goldbrick said:

For me, the parameter of interest is kWh and honestly, I don't know how - or if - that is different than a 'GID'. While a Li cell does have a fairly flat voltage discharge profile, it is less so when approaching the lower limit. Ie, the 'fuel gauge' is most non-linear where you need it most - when getting close to 'empty'.

Does using GID's help with this? I'm not sure since I don't know how a GID is defined. If there was a good way to measure remaining kWh that would be my preference, even if Ah is the preferred engineering metric.

Click to expand...

I have not proven to myself that the following approach is valid, but it is what I use. I suppose the best I can say is that it works for me:
I know my Ahr total usable capacity. I multiply it by 0.36 kV for usable kWh in full pack. Then I use SoC that either the car or LeafSpy provides.
E.g, the usable in my 2013 LEAF is 17 kWh. Each 10% SoC is 1.7 kWh

Since I live on a hill, I have never been tempted to validate that last 50 - 100 Wh of energy.

knightmb said:

I can say for certain that not all DCFC are created equal. I can take daily identical battery conditions to pit several network DCFC against each other (Blink, EA, Nissan HQ) in the same area only a few miles apart and some are more aggressive than others when charging given identical battery SOC, temperature, readings, etc.

Click to expand...

Interesting. I was under the impression that the car was actually requesting the current. The AddEnergie chargers (used by Electric Circuit) have an LCD that displays both the charger's possible current and car's requested current. Whenever throttling occurs, the displayed requested current matches the throttled current i.e. it appears as if throttling was because the car asked for it.

That being said, AFAIK we have no way to determine what the car is actually requesting, so I have no way of knowing if what's displayed on the LCD is truly what the car's requesting, or something the charger decided through a fancy algorithm.

knightmb said:

I know technically, it shouldn't be that way as the Leaf is suppose to be "telling" the DCFC how much power to send, but I don't think the Leaf has any safety protocols against too much power, seems to be more of a "at least this much power" protocol.

Click to expand...

Perhaps in your example the folks programming these chargers at EA have a x1.5 factor applied on the requested current or something... We have Electrify Canada chargers around - I will give them a go.

nlspace said:

Now that's a nice looking graph. The charging started at nearly 50kW, which may be the maximum of the Chademo station. The power was immediately pulled back at a fairly steep slope and gradually flattened until about 38% and then it went to a nearly linear rampdown.

Cold temperatures are a first order factor to determine charge rate. Chemical reaction rates are slower.

What do your cell voltages look like when doing a DCQC--are they evenly balanced or do you have any jumping out above or below the average? That could also be a major factor in charge rate.

Click to expand...

By looking quickly at the raw data, I see a 30mV max diff when charging at 125A. Max CP diff mV peaked at 50 when charging at 90A then tapered down as power went down. Max CP diff was about 17mV at rest after charging. I will try to work out a proper graph and will attempt a new DCFC with the right conditions this week.

Is there a way to keep LeafSpy recording throughout the whole charge without interacting with the car? It appears the CAN bus at the OBD port goes to sleep after 30-60 seconds with no user interaction. I found that unlocking or opening a door restores it temporarily, which is what I've been doing to minimize interaction with the car. Having the car on (or in ACC mode) would work but I am trying to avoid car interaction to eliminate variables.

knightmb said:

rogersleaf said:

Starting to wonder if this is the programming of car, or maybe the local ChargePoint station I frequent when needing to DCFC.

Click to expand...

I can say for certain that not all DCFC are created equal. I can take daily identical battery conditions to pit several network DCFC against each other (Blink, EA, Nissan HQ) in the same area only a few miles apart and some are more aggressive than others when charging given identical battery SOC, temperature, readings, etc. In my area, EA has the most aggressive DCFC. When Blink, Nissan HQ get past the 80% SOC mark and really ramp down the charging power, EA will keep it high all the way to 100%. So imagine that at 95% SOC, most DCFC are probably down to the 6kW to 8kW of charging power, the EA station will still be pushing 20kW to 30wK all the way to 100% SOC. I've done multiple experiments with my 2013, 2018, 2020 Leaf, the results are always the same in my area, the EA chargers are way more aggressive in power.
I know technically, it shouldn't be that way as the Leaf is suppose to be "telling" the DCFC how much power to send, but I don't think the Leaf has any safety protocols against too much power, seems to be more of a "at least this much power" protocol.

Click to expand...

Too bad there isn't much in the way of DCFC in this area to compare. I consider myself really fortunate to have this one relatively close, and it's still free. The specific charger has been broke down more than working over the past year (especially CCS side) so have a gut hunch they might be cutting power output back to compensate. However, my friend has a '20 LEAF+ SL that he said runs the 47KWh output max all the way to @ 65% SOC before descending. So mine is definitely behind his curve. Oh well, at least I wasn't being charged a fee, or charged by the minute.

rogersleaf said:

Too bad there isn't much in the way of DCFC in this area to compare. I consider myself really fortunate to have this one relatively close, and it's still free. The specific charger has been broke down more than working over the past year (especially CCS side) so have a gut hunch they might be cutting power output back to compensate. However, my friend has a '20 LEAF+ SL that he said runs the 47KWh output max all the way to @ 65% SOC before descending. So mine is definitely behind his curve. Oh well, at least I wasn't being charged a fee, or charged by the minute.

Click to expand...

I actually have 4 in my area, I forgot to list EVgo DCFC in my previous post, they sit along with Blink at being the "least aggressive" fast chargers. Since they charge by time, I don't want to put on my tinfoil hat, but it seems they do it that way on purpose. :?

Nominal 50 kW EvGo yesterday maintained 44 kW all the way from 10% to 80%, and cost $21.50 or about $.50 per kWh. A ripoff as far as cost goes but what are the options on the road.....