Leaf EM57 estimated output at different voltage/current

Nissan Leaf/e-NV200 drive stack topics
Post Reply
Zieg
Posts: 128
Joined: Mon Apr 25, 2022 3:31 am
Has thanked: 47 times
Been thanked: 45 times

Leaf EM57 estimated output at different voltage/current

Post by Zieg »

Hi everyone,

I have a gen 2 EM57 and inverter, and I have bought an Openinverter board but have not installed anything into the car yet. I am trying to calculate the expected performance for a given battery cell count, and thought I should ask for some help. The reason I'm asking this now is because I am considering using fewer cells from my Chevy Volt battery, and want to make sure I fully understand the impact that will have on how the car drives. The main reasons for this are for space and weight savings, as the car will be used mainly for Autocross.

We've probably all seen this graph before:
Powertrain-efficiency-of-a-Nissan-Leaf-motor-14.png
Looking at the peak power, it seems to be 275 N-m at 2750 RPM which of course equals 80 kW.

Then there's this chart for the 160 kW Leaf:
wykres_power.jpg
Here, the motor makes constant torque all the way out to 4600 RPM.

I know the 160 kW Leaf has a higher capacity battery, but I thought the voltage was the same. My understanding was that in order to maintain a given torque into a higher RPM, the voltage input had to be increased to overcome field weakening, so I would have expected the 160 kW graph to look like the 80 kW graph but scaled vertically by a factor of 2 (at least, approximately, and assuming the battery voltage is the same). Similarly, I thought that if you wanted to increase the speed at which field weakening takes over, the battery voltage had to be increased.

So, if I wanted to set up my Openinverter to increase the max current by 50% (120 kW), what would my graph look like? And if I wanted to decrease the battery voltage by say 12%, then what?

Any input is appreciated!
arber333
Posts: 3241
Joined: Mon Dec 24, 2018 1:37 pm
Location: Slovenia
Has thanked: 74 times
Been thanked: 223 times
Contact:

Re: Leaf EM57 estimated output at different voltage/current

Post by arber333 »

Zieg wrote: Thu Feb 09, 2023 11:54 pm ....
Any input is appreciated!
That is in EM57 just to protect the battery. If you use more powerfull inverter it doesnt have that limit.
Can you draw a graph for 300hp? That is what Arlin made on a dyno with EM61 and DIY inverter based on Lebowski brain.
Pete9008
Posts: 1801
Joined: Sun Apr 03, 2022 1:57 pm
Has thanked: 102 times
Been thanked: 347 times

Re: Leaf EM57 estimated output at different voltage/current

Post by Pete9008 »

Your understanding is pretty good. The horizontal line at the top of your first graph is set by the current limit set for the motor, the curved section is set either by the onset of field weakening or the DC limit of the battery (whichever happens sooner) and the diagonal section at the bottom right is the RPM limit for the inverter controller/motor (not too sure why it's diagonal rather than vertical though?). In your first plot I'm guessing it is the battery current that is limiting the curved section while the transition to it in the second plot is probably due to the transition into field weakening.

BTW - field weakening isn't quite the nice clean transition point usually described. There are components to the motor voltage due to the back emf but also due to stator currents. The more current being driven into the motor the earlier you run out of volts and the earlier the onset of FW.

It would be worth trying the Simulator to answer this type of question, I've used it for exactly that on the Outlander motor.

The thread is here: viewtopic.php?t=2611
Instructions are here: https://openinverter.org/wiki/IPM_Motor_Simulator
Example of it used on the Leaf motor is here: viewtopic.php?p=47770#p47770

Reasonable parameters to use for the Leaf motor seem to be Lq=0.7mH, Ld=0.2mH and flux linkage=72mWb (could do with someone doing a log and then using the calculator to verify these though).
Zieg
Posts: 128
Joined: Mon Apr 25, 2022 3:31 am
Has thanked: 47 times
Been thanked: 45 times

Re: Leaf EM57 estimated output at different voltage/current

Post by Zieg »

Pete9008 wrote: Fri Feb 10, 2023 9:46 am snip
Awesome, thank you so much!

I suppose I should next ask if anyone has some Inverter settings that work well with this motor. And the latest firmware version. I started off with the Leaf motor example and GT86 parameters from the wiki, but it looks like some of the parameter fields have changed.

I can get it to spit this out, but I can't really figure out why the torque is dropping off at such a low RPM.
350v.JPG
Pete9008
Posts: 1801
Joined: Sun Apr 03, 2022 1:57 pm
Has thanked: 102 times
Been thanked: 347 times

Re: Leaf EM57 estimated output at different voltage/current

Post by Pete9008 »

Glad you got it working :)

Try more FW Current Max, it should be set to around -flux linkage/ld so around -500.

Because of the low value set you are running out of -Id at higher speed and the control algorithm is having to back off the power instead to maintain control over the motor.
Dusty4
Posts: 83
Joined: Tue Dec 25, 2018 12:41 pm
Location: Western Australia
Has thanked: 11 times
Been thanked: 7 times

Re: Leaf EM57 estimated output at different voltage/current

Post by Dusty4 »

download/file.php?id=19700

The above link are the parameters Johannes posted of his Touran in the simulator thread, I am also running these parameters in my leaf motor and have been happy with it.
Zieg
Posts: 128
Joined: Mon Apr 25, 2022 3:31 am
Has thanked: 47 times
Been thanked: 45 times

Re: Leaf EM57 estimated output at different voltage/current

Post by Zieg »

Okay, I think I am making progress! Thank you! I plugged in those settings and got some pretty weird graphs, but once I started adjusting the Current Kp and Ki settings, it started to look better:
PI tuned.JPG
It still seems down on peak power and the torque is still falling off at a lower RPM than I would have expected. Increasing the FW Current Max beyond 600 started to produce some weird results into the higher RPM range. Peak power is still only around 88 kW.

I also had to keep the vehicle weight pretty high, I think if you lower the weight too much, it starts doing some weird stuff (graph turns into a spiral). In any case, my vehicle should come in closer to 800 kg and probably won't be able to use much more than about 300 N-m of torque, which is why I'm so focused on where it starts to drop off (or, more to the point, where the torque begins to drop below ~300 N-m).

I'm reading up on FOC theory so I can better understand what's going on, and hopefully I can work out how the later models are holding more torque into higher RPM. I see from that 160kW graph I posted earlier that they are still getting >250 N-m at 6000 RPM. Even if I could hit 200 at that speed, I'd be thrilled!
Pete9008
Posts: 1801
Joined: Sun Apr 03, 2022 1:57 pm
Has thanked: 102 times
Been thanked: 347 times

Re: Leaf EM57 estimated output at different voltage/current

Post by Pete9008 »

Unfortunately I've not used the simulator much with the latest OI firmware, or with Leaf parameters, but had a quick go with it yesterday. It's definitely not as stable as I remember and the power does seem a bit down. When I have time I'll try to have another look. Johannes probably has most experience using it with Leaf settings - @johu - any suggestions?

Regarding power there are a few things that could affect it:
  1. Wrong parameters - flux linkage in particular has a big effect on torque. If anyone has a running Leaf/OpenInverter setup with a web interface some csv logs would be very useful to verify we have the right settings - see viewtopic.php?t=3047
  2. Mtpv - I'm sure there is more power to be had in FW but haven't been able to progress it recently (see viewtopic.php?t=3080 for more details).
  3. There is no point going above -flux linkage/Ld for FW current. The flux linkage/Ld is the point that gives minimal back emf, so maximum field weakening, moving away from it in either direction just makes things worse. The spirals you mention are usually when the control loops are trying to find the minimum but instead of hitting it are circling around it.
  4. Unchecking the extra cycle delay box will improve stability. It means that the simulator doesn't fully represent the way the PWM values are updated in the OI firmware but that shouldn't matter for what you are doing.
  5. Until we know we have the right parameters for the motor, and the Mtpv stuff sorted, it may be worth not focusing too much on the absolute power levels and just use the simulator to evaluate the relative change produced by different battery voltage and throttle current settings. The other option is to tweak flux linkage (probably the parameter with least certainty at the moment) till you get values that you believe.
Pete9008
Posts: 1801
Joined: Sun Apr 03, 2022 1:57 pm
Has thanked: 102 times
Been thanked: 347 times

Re: Leaf EM57 estimated output at different voltage/current

Post by Pete9008 »

Two things. First Dusty4 very kindly took some logs from his Leaf motor (see viewtopic.php?p=52875#p52875) so we have some better motor parameters to try :)

Also I've had a look at the simulator to see what's causing the stability problems and have found a fairly obvious bug :oops: . If the motor current goes above around 400A using the standard param_prj.h file the adc saturates and the stability issues start. Essentially this is simulating the behaviour of the inverter correctly but the default file is for a Prius inverter which has a lower range on the CTs, not normally a problem on the Prius but if the same file is used on a Leaf setup without revisions you get adc saturation and stability problems. Sorry about that.

This is probably why you found the need for the very low kp and ki gains.

To fix change the lines in the param_prj.h file (the one in the same directory as the IPMMotorSim.pro file) from

Code: Select all

    PARAM_ENTRY(CAT_INVERTER,il1gain,     "dig/A",   -100,   100,    4.7,    27  ) \
    PARAM_ENTRY(CAT_INVERTER,il2gain,     "dig/A",   -100,   100,    4.7,    28  ) \
to

Code: Select all

    PARAM_ENTRY(CAT_INVERTER,il1gain,     "dig/A",   -100,   100,    1,    27  ) \
    PARAM_ENTRY(CAT_INVERTER,il2gain,     "dig/A",   -100,   100,    1,    28  ) \
and rebuild the project. This should fix it in the simulator.

Here are a couple of plots using Dusty4's parameters and the fixed simulator code. First with the standard code:
Std_700A_380v.png
And next with the pseudo Mtpv derivative I'm working on:
Mtpv2_700A_380v.png
Unfortunately the later isn't ready to be released yet. (Edit - but it is getting a lot closer to your 250Nm @ 6krpm)
Zieg
Posts: 128
Joined: Mon Apr 25, 2022 3:31 am
Has thanked: 47 times
Been thanked: 45 times

Re: Leaf EM57 estimated output at different voltage/current

Post by Zieg »

Oh wow, that's awesome! Thank you so much! I was just typing out a reply to thank you for your last post and when I hit submit, it told me there was a new post, haha. I'll make that change shortly and retest. I really appreciate all the help. Again, THANK YOU!
Pete9008
Posts: 1801
Joined: Sun Apr 03, 2022 1:57 pm
Has thanked: 102 times
Been thanked: 347 times

Re: Leaf EM57 estimated output at different voltage/current

Post by Pete9008 »

No problem, thanks for identifying the issue :)

Have a feeling it's been causing me the occasional issue for a while but never twigged what the cause was, nice to have found it. Hoping some of the Mtpv algorithms might start to behave a bit better now too!
User avatar
rstevens81
Posts: 349
Joined: Sun Dec 22, 2019 10:36 am
Location: Bristol, UK
Has thanked: 21 times
Been thanked: 91 times

Re: Leaf EM57 estimated output at different voltage/current

Post by rstevens81 »

am i reading the graph's correctly that you estimate that you could get 600Nm of torque out of the leaf? and then constant power from 2000 rpm as if this is remotely true it opens up direct drive possibilities (no gearbox)
Rule 1 of EV Club is don't buy a rust bucket....
Which rule does everyone forget 🤪
Pete9008
Posts: 1801
Joined: Sun Apr 03, 2022 1:57 pm
Has thanked: 102 times
Been thanked: 347 times

Re: Leaf EM57 estimated output at different voltage/current

Post by Pete9008 »

In theory it should be possible.

In practise something may well melt or break. In Zieg's plot Nissan seem to limit it at 350Nm, possibly to protect the gearbox?

The only way to find out would be to try it!
Zieg
Posts: 128
Joined: Mon Apr 25, 2022 3:31 am
Has thanked: 47 times
Been thanked: 45 times

Re: Leaf EM57 estimated output at different voltage/current

Post by Zieg »

Well 350Nm x 8.2 (reducer ratio) is 2870 Nm. Divide that by the 0.335m radius tire and you have some 8567 N of possible forward thrust. Google says a 2019 model has 916 kg of mass on the front wheels (8986 N force) and of course under acceleration that'll be reduced due to rearward weight transfer. Unless we're running tires with a friction coefficient greater than 1, it actually seems like a surprisingly good match!
glink
Posts: 18
Joined: Mon Mar 15, 2021 7:02 pm
Been thanked: 3 times

Re: Leaf EM57 estimated output at different voltage/current

Post by glink »

Hi, the settings here has helped a lot. I have set up a "dyno" rig to try to find out how to configure my setup for my boat using only 48V. I'm using Auris inverter (about the same spec as Prius inverter I guess) and Nissan EM57 motor. The KV of the EM57 is 18.46 according to sources, and correlates very well IRL. I'm able to get around 850 RPM on my EM57. Question is, is it possible "to learn this setup new tricks" with field weakening? I find no setting in the latest v5.24.R FOC to tweak so that I could get more RPM. Problem is, I do not want to go beond 48V on the battery side and the boost converter on the Toyota inverter seems to be very limited, only able to squeeze 3kw out of it before the heat sinks scream for help. Is it possible to trade torque for RPM in my setup? I would only need to go to 1500 RPM.
Pete9008
Posts: 1801
Joined: Sun Apr 03, 2022 1:57 pm
Has thanked: 102 times
Been thanked: 347 times

Re: Leaf EM57 estimated output at different voltage/current

Post by Pete9008 »

Was curious what might be possible here so did a couple of simulations. I've used 40V to allow for the voltage drop across the IGBTs and diode's. First with the standard parameters:
Leaf_40V_Std.png
There is a bit of oscillation coming in at higher speeds so another with the PI gains increased to compensate for the lower voltage (the actual gain seen by the motor is the product of the gains and the battery voltage):
Leaf_40V_Mod.png
So the limit is about 8kW up to the speeds you need. The motor is actually in field weakening above around 350rpm.

The boost converter is limited to around 100A so at 48V this is around 4.8kW so it's not too suprising that you struggle to get more than around 3kW before things start to get hot.

I think you will be better off bypassing the boost converter (to remove the series resistance) and driving the DC bus directly from the battery. You should then be able to get around 8kW. If you need more then I'm afraid that you will need a higher battery voltage or a different motor. It might also be possible to rewind the Leaf motor to suit lower voltage operation (see posts by Romale).

Edit - thinking about it 40V might be a little pessimistic (some of the volt drop is already included in Rs) so another plot at the full 48V to give the optimistic view on what could be expected:
Leaf_48V_Mod.png
So 10kW - the real figure will hopefully be somewhere between 8 and 10kW.
Post Reply