150KW Lebowski Controller

I have been working with the Lebowski controllers ICs for a little while and its time to start scaling up my project. The end goal is to develop a energy recovery system to sit alongside a race engine and supplement its power when required. The current motors are 2 Honda IMA sandwiched between the engine and gearbox. These are capable of running at 20KW all day with no problem, I have manufactured a new housing that facilitates the addition of water cooling and we have been able to push these motors 60kw each for 2 miniutes at a 40% duty cycle, this is perfect for my application.

My controller has been based around a breadboard which is not ideal for the motorsport environment, its now time to get the car off the dyno and round the track. I have been researching existing automotive controllers to see if something could be repurposed to suit my application. I have found what i believe to be the perfect donor for a Lebowski controller power stage. The Mitsubishi Outlander PHEV is fitted with a 60KW generator off the engine, the generator is capable of starting the engine. A 60KW motor is fitted to the front axle. The inverter for these motors are fitted in one unit that i have just had delivered to my door from a 2014 PHEV for £150. That is less than the price of the IGBTS alone.

Below are some images of the teardown.


This is the control board, we are not really interested in this as it is manufacturer specific, it appears to multiplex all the signals then take them out of the inverter to another black box



This is the power stage you can see the current sensors for each inverter at the bottom of the image the battery connection is the centre two terminals. This is all mounted on a 450 volt 1600uf capacitor. The capacitor is potted and acts not only as the DC Link cap but the distribution to each IGBT module

Now the important bit the power stages. This inverter is fitted with two identical power stages. In the car they are designed to operate two 60kw motors. The IGBTs are rated at 600A dc and the DC link voltage is 300v nominal, I make that 180KW each power stage 360KW total. Obviously these are Chinese rating but my rough estimates each inverter is capable of 150KW Peak.



Separate and identical power stages



Each power stage view i will discuss components below.



Direct Water Cooling for the IGBT, for those without the space for water pumps and radiators, you could consider phase change cooling.

More to come....

The Power stages are based around a hitachi MBB600TV6A. this device is capable of 600v ABSMAX 600A DC. The IGBT is controlled by opto isolated gate drivers that monitor for shoot through and desaturation of the device ACPL-34JT. The DC Link voltage is monitiored again optoisolated by a ACPL-C87AT.

I am just working through the PCB to produce a wiring diagram but all the signal are brought out onto nice headers for us to connect the Lebowski controllers to.

I am going to have to implement some additional logic on to of the controllers as the board gives us lots of fault feedback and information that needs to be processed.

I will attach the circuit diagram shortly. Stop building power staged and get buying these

Thanks

James

This is pretty exciting, James.

I know this is a little off-topic, but do you mind sharing any information on your motor enclosures? I did most of the design on one awhile back, but never produced it.

Do you have logged data showing motor power vs. temp or the like? The ORNL report claims the motor starts saturation around 300A, but of course it doesn't fully saturate until much higher. Can you share what current and voltage you're running, along with how you implemented your control scheme to take advantage of the IPM's saliency?

Looking forward to seeing more of this, good luck!

-Coe

Nice work!!

Very impressive!

exciting stuff, Bobc and I a using lebowskis chip in a controller for a scooter only 10kw though on one ixys mti 6pack chip.

If you want any pcb (through hole) boards I sell them pretty much at cost, well £9 before postage (bought bulk to reduce my cost from pcbway). You might be designing your own surface mount board like animalLector but if you aren't bothered about that then this is Lebowskis own latest design and hes given his blessing for me to distribute these. PM me if you are interested.

Gonna subscribe, also I drive a Honda Insight mk1 so its interesting to see how you get on driving that ima with lebowskis chip.

Ok, time for an update. We are very busy at the workshop getting the car ready for next season so inverter development a little difficult at the moment. I have made some good progress. I have produced a schematic for the power stage and mapped all the connectors so I'm ready to connect to the bread board over the next few weeks.

I am going to order some of the latest chips from Lebowski. I am going to be using two ME0913 for testing out of the car.

I will be happy to share all the hardware designs next year once the car is complete , the design is subject to change at the moment and i don't want anyone getting a head start on us



Here is the link if the picture doesn't display correctly https://scontent-lhr3-1.xx.fbcdn.net/t31.0-8/14940018_10154517309286839_4952996924566000476_o.jpg

Now the important bit the power stages. This inverter is fitted with two identical power stages. In the car they are designed to operate two 60kw motors. The IGBTs are rated at 600A dc and the DC link voltage is 300v nominal, I make that 180KW each power stage 360KW total. Obviously these are Chinese rating but my rough estimates each inverter is capable of 150KW Peak.

Click to expand...

Wouldnt that be "Japanese rating". rather than Chinese ?
In which case i wouldnt under estimate them !

Since the Mitsubishi Inverter only uses two phase current sensors per IGBT set, what phase current sensors are you thinking of using?

Might try to get my hands on the same powerstage as it looks like alot of potential, atleast more then my current Honda IMA powerstage.

Only two current sensors are required i think. I will check though.

I should be running in a month or so, i will let you know how i get on.

I stepped away from 2 current sensors and went to 3 only. On a theoretical level only 2 sensors is possible, as with Ia and Ib known Ic can be calculated as being -Ia-Ib. But in practise 3 current sensors is much better than 2. With only 2 sensors all the noise and (common mode) disturbances are assigned to the 3rd channel, making for bad and noisy current measurement results. With 3 sensors you have the option to filter out common mode disturbances and noise, ending up with much cleaner signals. I know many controllers from known manufacturers work with only 2 sensors, for me it is not something that should be used in real-world electronics...

Anyway, the simple option would be to add this back to the controller IC, but I do not want to encourage using only 2 sensors (did I sat it was bad ? ) so am very, very hesitant about doing this. Even with the controller IC wanting 3 sensor signals and not having the option of 2 sensors in the menu structure, you still can fake the 3rd signal by using a simple Opamp circuit:

This I got to see completed.

Any update? I have the rear inverter which uses the same IGBT module, and i am also using the lebowski chip. I wish i had saved the pinouts, Can you re upload them?
Cheers