Francisco's Electric Pajero

Post up a thread for your EV. Progress pics, description and assorted alliteration
antiscab
Senior Member
Posts: 2809
Joined: Mon, 26 Nov 2007, 05:39
Real Name: Matthew Lacey
Location: Perth, WA

Re: Francisco's Electric Pajero

Post by antiscab »

francisco.shi wrote: Fri, 16 Apr 2021, 21:01 I have been thinking making it V2H capable.
The plan is to connect an inverter via the DC charge port.
I am going to make an isolated bidirectional DC link and feed the output to a solar inverter via the solar panel inputs and the solar panels will go to my converter. So that way I can feed the house from the solar or from the car.
You could use a separate inverter for the V2H - an export limiting inverter would probably be most cost effective (fronius, goodwe, etc)
the upgrade for a fronius symo to do export limiting is only $500 on 3-phase
Matt
2017 Renault zoe - 25'000km
2007 vectrix - 156'000km
1998 prius - needs Batt
1999 Prius - needs batt
2000 prius - has 200 x headway 38120 cells
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

coulomb wrote: Fri, 16 Apr 2021, 21:36
Neat! Presumably, you'll have some sort of adjustable sag that the inverter sees, so it knows when to stop drawing power (when it sees what it thinks is the maximum power point).
That is one way. So the add on box will have a CT to decide how much power to give the inverter and make the bus sag until the MPPT draws the right power.
The main disadvantage is that I could not charge from the grid using this setup.
In my opinion any car should be able to share the battery thru the DC charge port. This way it would not need any hardware changes in the car. All it would need is a software update to stop the car disconnecting the battery if it sees the charger trying to discharge the battery.
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

I did a bit more work on the inverter software today.
I have been having some problems with the control loop. As the speed goes up it gets to a situation where the control loop seems to be unable to control the current.
Here are the readings from the Q axis.
Q axis.
Q axis.
20210501_220428.jpg (1.38 MiB) Viewed 2278 times
The feedback goes down even as the output goes up and also the current goes up. Eventually it trips on overcurrent.
Here is the D axis sowing the same problem.
D axis
D axis
20210501_220504.jpg (1.37 MiB) Viewed 2278 times
After some thinking and doing a few tests I came to the conclusion that the problem is due to the measurement and the update of the output taking place too far apart. After the measurement is taken the control loop does the calculations and it then updates the PWM values which get updated on the next PWM cycle. The end result is that as the speed goes up the rotor angle changes enough that when the update happens it is already too far off. I had already added some compensation for the rotor speed. The problem was that there are 2 PWM cycles between the measurement and the update but I was only allowing for 1 cycle.
Here is with the fix implemented.

The motor got to about 8krpm (about 100kph)
I didn't get a plot of the currents. I will do that later and will post them.
Tomorrow I will do some work on the field weakening.
antiscab
Senior Member
Posts: 2809
Joined: Mon, 26 Nov 2007, 05:39
Real Name: Matthew Lacey
Location: Perth, WA

Re: Francisco's Electric Pajero

Post by antiscab »

I look forward to seeing your field weakening results

Are you planning on doing field strengthening too? (more low end torque and more effective low speed regen)
Matt
2017 Renault zoe - 25'000km
2007 vectrix - 156'000km
1998 prius - needs Batt
1999 Prius - needs batt
2000 prius - has 200 x headway 38120 cells
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

From the maths it seems field strengthening should allow for more torque at low speeds but I am not sure if it works. I am not sure if something in the magnetic path may saturate or if the windings will be able to take enough current.
I have done some testing on field weakening before but it appears very inefficient. So far to get the motor to spin at twice the speed I have had to apply the rated current to the D axis which leaves nothing for the Q axis. If I did field strengthening I would expect a similar problem. I didn't test the field weakening under load so I can't tell how well it will work and when I tried to do the field strengthening I did not seem to get more torque. I didn't measure the torque accurately because I have no way of measuring the torque on the motor especially when it is moving at speed.
The only way I can apply a load to the motor is to drive the car.
The other problem I have testing field weakening is that my driveway is not long enough to get to speed and maintain the speed and I don't want o drive the wreck on the road. It is a bit too obvious.
The plan for field weakening is to come up with a method of calculating the current that I need to apply to the D axis. I need to measure something and then increase the current as the speed increases.
One idea is to look at the output voltage and increase the D current as the output voltage gets close to saturation and run a control loop that keeps the output voltage from reaching saturation.
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

I didn't get any work done today. Some friends came over and ended up just playing with the car.
But I did get some videos:

This is a demonstration of the one pedal braking accuracy.
I would challenge any other EV driver to do the same without touching the brake.
This one I just wanted to do an acceleration test to compare to before when it had only one motor but didn't take it from the right angle.
fffffred
Groupie
Posts: 31
Joined: Wed, 09 Jan 2019, 10:16

Re: Francisco's Electric Pajero

Post by fffffred »

amazing work @francisco.shi
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

Haven't had much progress lately.
I was doing some tests a few days ago and decided to check the current out of the battery. I got 300A out peak and about 200A in during regen.
So I need to start building the battery pack. The leaf cells I am using are not going to like to be treated like this.
So today I decided to finish assembling the battery assembly jig.
Battery assembly jig
Battery assembly jig
20210515_184915.jpg (1.62 MiB) Viewed 2045 times
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

I have doing some work on the inverters and found there is still a problem with the control loops not working correctly as the speed goes up despite adding the advance for the resolver measurement.
After lots of work, analysis and hair pulling I came to the conclusion that the problem must be with the timing.
So I took the motor out of the car and put it on the bench and started analysing the sampling and timing and came to the conclusion that the timing was the problem. So there several problems.
The first problem was that the sampling of the resolver was at a different frequency as the control loop and the sampling of the currents so the time between the resolver measurement and the current measuremrnt kept changing. That is why the advance did not work correctly.
So the fix was to synchronise the resolver excitation signal with the PWM So that the time between the resolver sample and the current measurement sample was fixed. I the used the oscilloscope to estimate the time by looking at when the ADC interrupts happen.
I also used the same method to determine the time between the resolver reading and the PWM update.
I then applied a compensation to the resolver angle used in the Parke transform of the currents and the inverse Parke transform for the applied phase voltages based on motor speed.
This seems to have solved the problem. The control loops now seem to behave correctly.
There is still a problem with the resolver reading. There is some noise and non linearity which is causing some distortion in the transformed signals. Unfortunately this needs some hardware modifications so it will take a bit of time. In the meantime I will try to get the field weakening working.
User avatar
jonescg
Senior Member
Posts: 3953
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

Re: Francisco's Electric Pajero

Post by jonescg »

I've always found it interesting that high-end AC synchronous motors use resolvers, and lower-end motors use encoders. Never worked out why.
AEVA National Secretary, WA branch chair.
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

Resolves give you an absolute position, are a bit trickier to read and generally have been expensive (but they are really cheap to make)
Encoders generally give you an incremental position which means you need to find the origin.
Encoders are easier to read.
User avatar
OzSpider
Groupie
Posts: 52
Joined: Mon, 14 Dec 2020, 10:02
Real Name: Derek
Location: Brisbane, Australia

Re: Francisco's Electric Pajero

Post by OzSpider »

Hi @francisco.shi did you retain the two gear ratio's of the toyota transfer case gearset? Was there an option to use both ratio's for a two speed option? I know a std transfer case can't be shifted from low to high on the fly. But I'm curious...
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

I didn't because the ratio is 1:1 which was not useful or 2.56:1.
Surprisingly the transfer case has synchros so it could be made to shift on the fly.
If I wanted to have two ratios I would have had to put two reducers. That would have made it perfect for offroad but it would have made the motor assembly too long to fit batteries in the middle.
Caco 4x4
Noobie
Posts: 3
Joined: Thu, 24 Jun 2021, 10:22
Real Name: Juan Carlos Puccio
Location: Saniago, Chile

Re: Francisco's Electric Pajero

Post by Caco 4x4 »

Hi Francisco, I want to convert my 1990 montero manual 4x4.
I drive in mountains and would like to start with 50 km range.
I want to start with final motor and controller and upgrade battery pack and other things once its running
and I have the bucks.
I read the whole thread, you do incredible work, way above anything I can do.
Why not put 1 powerful motor or 2 hooked up to original transmission and transfer.
Is it weight?
Juan Puccio
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

There are a few issues.
1) a more powerful motor is bigger and heavier. The main reason is that most manufacturers are making motors with about 100kw and 300Nm torque for small cars. If you want more power you need a bigger motor which is generally for a bus or a truck and they are much heavier. So to get twice the power it will weigh more than twice the weight. Mainly because they spin slower and there isn't anything in between (not at the moment anyway)
2) if I put two motors I can remove the gearbox and transfer case which uses a lot of space under the body. If I put one motor and keep the transfer case I would not have as much space for the batteries. I would end up with almost bald the battery capacity.
The 1990 also has live axle rear which means you can not put batteries in the middle.

My main objective was to have as big a battery as possible (range anxiety) but not have batteries spread out everywhere and good acceleration.
I did not consider offroad ability high on the list.
As it is the car will be able to go offroad but the gearing is too tall and would be very inefficient on low speed rock crawling. It really would need a low range gear. I could make a two speed gearbox still in the one unit but it would make the motor longer and I would loose the batteries in between the motors.
Another possibility would be to put one motor for each wheel or a transaxle for each axle (like a manual Camry gearbox) but that is for the next life.

Where are you?
Caco 4x4
Noobie
Posts: 3
Joined: Thu, 24 Jun 2021, 10:22
Real Name: Juan Carlos Puccio
Location: Saniago, Chile

Re: Francisco's Electric Pajero

Post by Caco 4x4 »

Thanks for your quick reply, I´m in Santiago, Chile the other down under. haha

1. Off road ability is not high on my list either. Some sand for fishing and a little snow on mountain roads.

2. Initial range is not a priority (50 k) I want to start with the best motor and controller etc. that I can get on a usd10k budget. And at some point in the future upgrade to Lithium.

3. I have a 4km steep climb to get home and the curb weight is 2050kgs.(is there a way to use 10 minutes of down hill in the morning?)

4. I don´t know if its worth going AC or not.

5. Should I start with just rear wheel drive, with original trans and later add motor for the front?

I would definitely need to buy the controller and BMS, I´m a builder and continuity is about my level of electric know how.
Pretty much every thing will need to be imported.

Caco
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

It will really depend on how much power you want to have. If you keep the manual gearbox you can get away with less power.
Roughly my car with one motor and single speed feels like driving in second gear so not very responsive. With two motors it is about 30% stronger than the original in 1st gear.
The main issue is take off speed.
So you will need at least the same power as the original motor.
Off the shelf you can get a Hyper9 but that is 35kw continuous and 100kw peak.
Another thing you need to consider is waterproofing. The Hyper9 is not very waterproof.
The easiest would be to use a motor out of a wrecked EV but most are still only about 120kw. Except if you get a Tesla motor.
I would go AC. It is a much better solution and it will do regen braking.
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

I have been a bit busy lately and have not done much work on the conversion.
I have been having some problems with the inverters.
The problem is that I am getting some DC thru the motor when the motor runs. The DC is proportional to the current flowing thru the stator. I have checked the modulation and gate drive signals and all seems fine. So far I can be confident it is not a measurement error or the logic board applying the wrong PWM signals.
So far I have no explanation for the behavior and it also doesn't make sense.
To get to the bottom of the problem I need to run the motor under load with good access to the power electronics.
So here is a jig to run one motor against the other.
Test rig.
Test rig.
20210919_115407.jpg (1.56 MiB) Viewed 1356 times
I hope to do some testing in the next week or so.
User avatar
jonescg
Senior Member
Posts: 3953
Joined: Thu, 21 Jan 2010, 23:05
Real Name: Chris Jones
Location: Perth, WA.

Re: Francisco's Electric Pajero

Post by jonescg »

Short the windings through a heater element and use it as a dyno?
AEVA National Secretary, WA branch chair.
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

The first test i want to do is short the terminals of the test motor (with a bus bar) and use the other motor to drive. Then check that there is no DC (I am pretty confident there won't be any) and check that the D and Q currents are smooth.
After that I will connect the motor terminals to the power stage and use the power stage to apply a short circuit to the motor by applying 50% duty cycle to all 3 outputs and go from there.
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

I have been doing a bit of testing over the weekend.
I replaced the resolver exciter circuit with a slightly better design.
I made a resonant circuit that runs from the logic output and then I used an audio amplifier to drive the exciter coil. This arrangement works better from the point of view that it has less effect on the exciter signal when the rotor turns.
I did a few measurements and all seems fine.
There appears to be no DC offset any more.
Here is a video of the test.

I was running 450A RMS into the motor. You can see the cables moving from the current flow.
The inverter under test has now decided to behave as it should.
The DC offset has miraculously disappeared.
This is the current measurement using a bus bar to short the terminals.
Perfect sine waves and no DC offset.
Inverter under test with motor shorted with bus bar.
Inverter under test with motor shorted with bus bar.
20211010_212134.jpg (1.89 MiB) Viewed 746 times
This is the current measurement as above but using the output stage to short the motor but only half the windings shorted to avoid damage to the output stage.
Inverter under test with motor shorted thru output stage.
Inverter under test with motor shorted thru output stage.
20211010_212156.jpg (1.82 MiB) Viewed 746 times
This is the current measurement with the inverter driving with 40v bus voltage.
No DC offset and output current reasonably sinusoidal. The distortion is due to the voltage drop from the diodes which I will attempt to fix at some point and may not be necessary because once working at higher voltage the distortion will be reduced.
Inverter under test.
Inverter under test.
20211010_205328.jpg (1.22 MiB) Viewed 746 times
This is the other inverter with no modifications.
You can see the current waveforms are very distorted and there is a DC offset.
Driving inverter.
Driving inverter.
20211010_205406.jpg (2.02 MiB) Viewed 746 times
I really hate these solutions when the problem just disappears without any apparent reason.

Tomorrow I will make the resolver mods to the other inverter and see what happens.
I do not see how fixing the resolver can fix the DC offset. From my understanding it is not possible for the FOC output to produce or correct a DC offset.
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

I replaced the exciter circuit on both motors and now the DC problem seems to have miraculously disappeared.
I do not understand why this fixed the problem.
The good thing is that I should be able to run the motors at full power without tripping and the rumbling noise has disappeared.
I will try to run the motors at 200v instead of 40v and see if the problem has been fixed.
I am hoping this will make the project progress quicker.
I don't want to drive an ICE car anymore. I have also tested the one pedal driving mode on an Imiev and I can't wait to try it on a car with more power and AWD.
necrogt4
Groupie
Posts: 89
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Francisco's Electric Pajero

Post by necrogt4 »

What's the max voltage and amps you plan to run through these motors when everything is running properly?
francisco.shi
Senior Member
Posts: 626
Joined: Mon, 17 Sep 2018, 16:30
Real Name: Francisco
Location: Brisbane, Strathpine

Re: Francisco's Electric Pajero

Post by francisco.shi »

The current inverter can do 800A 500v.
But I can only get 420 from the battery (not enough space for more cells).
I have plans to change the transistor packs which will allow 1200A and up to 900v. But I would have to change the bus capacitor.
The motor in its original winding will take about 400A maximum and in the rewound form it will take 1000A.
To get the full benefit of the rewound motor and higher current I need to change the gear ratio of the planetary gearbox (or diff) but I am not sure if I will go to that trouble.
necrogt4
Groupie
Posts: 89
Joined: Sun, 28 Oct 2018, 16:49
Real Name: Luke

Re: Francisco's Electric Pajero

Post by necrogt4 »

How many cells did you manage to fit? I'm running 128 in my MX-5 but keeping voltage to 132v max so they'll never be 100% charged.

I imagine 420v @ 800A would provide plenty of forward momentum!
Post Reply