Lebowski's motor controller IC, schematic & setup, new v2.A1

Arlo1 said:
Uhm actually the measurement I got was all over the place but almost never that low.

Amber DID NOT CALCULATE THE DEAD TIME CORRECTLY. This will cause miss readings in the resistance and all kinds of other problems. The dead time is determined by math based on all the worst case scenarios this is for prevention of a shoot through due to one transistor turning on before the other is fully off. If you are switching them to close together it will cause weird measurements. Also if you switch them to fast. Amber maybe its wise for you to start your own thread and we can try to debug your system.

Hm... i will create my own thread for sure. Now that you present argument like that... i will take a look at my readings.
I dont have a diff scope. I can only scope board signals. Or i can scope phases directly without GND.

tdead = [(t D _ OFF − t D _ ON )+ (tPDD _ MAX − tPDD _ MIN )]×1.2

I calculated deadtime with Infineon equasion from IGBT datasheet and made asumption about driver optos having the 0.25us delay. I dont know which optos Hitachi used, so this is educated guess.
On closer inspection i see i should be using toff max and ton min. Together with my assumption i arrive at 2.1uS for deadtime then.

I could try with 2us deadtime like you did. In my car i have Skyper 32pro drivers which demand 3us deadtime. But that is a different matter.

A
 

Attachments

  • mbb600tv6a.pdf
    844 KB · Views: 108
arber333 said:
Arlo1 said:
I would not put more then 10nf on the current signals it changes the timing. From signal to gnd

Maybe scope the hall signals vs the phase signals.

I dont mean to decouple current signals. They already have their caps. I am talking about hall sensors. They have 4K7 pullups, but i am wondering if it wouldnt be good to use 1nF cap on each signal towards GND?

A

I had my encoder wire fail and when I fixed it operation in sensored mode was rough and worse then every. I eventually come to the conclusion to add 10ohms (I think it was) in series between the encoder and controller GND.
 
arber333 said:
kiwifiat said:
I think that there is something fundamentally wrong with your setup if you are getting 6.66 mOhms for motor resistance. The resistance posted by Arlo for the Gen1 motor is over 20mOhms which agrees with the figure I get for the gen2 motor. Also I think you are asking for trouble setting 500ns for deadtime on those Chevy Volt inverter IGBT's if you haven't scoped the gate signals to see how fast they are switching. Better to err on the side of caution.

No! Deadtime is set at 1.65us. You have misinterpreted the data. It was calculated and tried with another brain. I succesfully drove my ACIM car with this inverter. So i am not worried about deadtime. Well as long as 1650ns really mean 1.65us.

I am worried about low resistance. This is a parameter that is good for Emrax motor not Leaf. But strange is i also get very consistend measurement with only a 2uH difference and almost no resistance difference...

1.65uS or 1650nS is likely ok but make sure its calculated and I did play with this and is changed the FOC measurement. I only played with it trying to debug things.

Also I always did my FOC measurements at a lower current. For this you need to go through all the menus first. Then do the FOC measurement. Then reset you current to the level you want to run.

I did my FOC at 200 phase amps (my system will be different then yours)
Also with the FOC measurement at to low of a value or to high of a value it will be incorrect.
 
Arlo1 said:
Uhm actually the measurement I got was all over the place but almost never that low.

Amber DID NOT CALCULATE THE DEAD TIME CORRECTLY. This will cause miss readings in the resistance and all kinds of other problems. The dead time is determined by math based on all the worst case scenarios this is for prevention of a shoot through due to one transistor turning on before the other is fully off. If you are switching them to close together it will cause weird measurements. Also if you switch them to fast. Amber maybe its wise for you to start your own thread and we can try to debug your system.

To quote you with the latest Lebowski firmware that Arber is also running:

" I also tried using the inductance and resistance measuring option in the controller to let it try to measure the motor and it give me a number of 150 uH and 25 mOhm (the numbers I have had working for the 1/4 mile runs" .
You have posted other data indicating resistance in the 20's for the leaf motor.

The point is that 6.66 mOhm is plainly indicating an issue that needs to be resolved.
 
I don't remember how the controller reacted to resistance measurement changes but the math for current raise time is not effected by resistance unless it's insanely high like a few ohms
 
Hi Lebowski

What kind of circuit would you use with analog voltage signal 0-5V for values from 0-400Vdc?
You say your circuit only sees signal from 2.5V to 5V. That would mean i can only use signal from 200V to 400V! Could i translate 0-5V signal to 2.5-5V effectively? Suggestion?
Should i use one ATtiny chip, reading the value on one analog pin and transmitting PWM on another? Would that work?

tnx

Arber
 
I use just a resistor, but Arlo1 haf a good isolating opamp in his high power design.

Yes it is programmed only for a range of 2, as that should cover a battery between empty and full ...? It needs to do a 1/x internally, the factor 2 is chosen to limit the output range of the 1/x calculation...
 
Yeah I used a 100:1 voltage divider to bring it down to 0-5v 5v = 500v
Then a isolation chip this one with a op amp and brought it back up to 0-5Vdc isolated on the output.
https://docs.broadcom.com/docs/AV02-3563EN


I have 2 verions. This one I used to run dual sign/co-sign size 6 controllers for a project for Zero. <a href="https://oshpark.com/shared_projects/fxERmxzo"><img src="https://oshpark.com/assets/badge-5b7ec47045b78aef6eb9d83b3bac6b1920de805e9a0c227658eac6e19a045b9c.png" alt="Order from OSH Park"></img></a>

And this I used in the car for precharge and the signal to the brain board.

<a href="https://oshpark.com/shared_projects/VeA4yYcf"><img src="https://oshpark.com/assets/badge-5b7ec47045b78aef6eb9d83b3bac6b1920de805e9a0c227658eac6e19a045b9c.png" alt="Order from OSH Park"></img></a>
 

Attachments

  • C87.JPG
    C87.JPG
    37.2 KB · Views: 3,750
Ah and you use resistors to set the division for your viable input range. I get it! No need to worry about 2.5V range if you are never there.
Thank you Arlo.

However i am working now on Prius inverter and it has nice 0 - 5V voltage detection output which describes 0V - 650Vdc on the HV side and 0V - 400Vdc on battery side. Since signal is linear with HV i cant simply use resistor divider. Does anyone have any idea how to change 0 - 5V signal to 2 - 5V signal?
 
Arber, I think you misunderstand.

When you want to deal with a 400 V battery you need to divide this voltage down by a factor 80.
The controller IC will then work correctly for battery voltages between 200 and 400V, as these map
in the 2.5 - 5 V range.

You should not map 0 - 400V on 2.5 - 5V ! You're supposed to use a simple divider. If your Prius dectection
circuit linearly maps 0-650V on 0-5V, you can use this signal directly. The controller will then work correctly for batteries between 325 and 650V, as this range is mapped to 2.5 - 5V

Outside of the 2.5 - 5 V range the controller IC wills till work, just the FOC accuracy will go down a bit...
 
Lebowski said:
Arber, I think you misunderstand.

When you want to deal with a 400 V battery you need to divide this voltage down by a factor 80.
The controller IC will then work correctly for battery voltages between 200 and 400V, as these map
in the 2.5 - 5 V range.

You should not map 0 - 400V on 2.5 - 5V ! You're supposed to use a simple divider. If your Prius dectection
circuit linearly maps 0-650V on 0-5V, you can use this signal directly. The controller will then work correctly for batteries between 325 and 650V, as this range is mapped to 2.5 - 5V

Outside of the 2.5 - 5 V range the controller IC wills till work, just the FOC accuracy will go down a bit...

Ah, ok!
650Vdc sensing is done on the side of the HV capacitor. While the battery voltage is sensed one the battery capacitor side. This side mesures up to 400V with 0V - 5V signal. Between those is a push pull transistor pair and inductor. We are now experimenting with this inverter and it seems it is surprisingly tough.
Well if this will work regardless i am not going to fuss about it and i will just use it.
 
Amber this needs to be isolated!

That's why I posted the circuit!
 
Arlo1 said:
Amber this needs to be isolated!

That's why I posted the circuit!

Depends on how far reaching you want your power stage blowup to be :wink:

Arber, i dont really understand the distinction beween the 400 and 650, but the controller ic needs to see the power stage supply voltage.
 
Lebowski said:
Arlo1 said:
Amber this needs to be isolated!

That's why I posted the circuit!

Depends on how far reaching you want your power stage blowup to be :wink:

Arber, i dont really understand the distinction beween the 400 and 650, but the controller ic needs to see the power stage supply voltage.

And if you want your car chassis live and likely to kill you or a first responder or a innocent passer by.
 
Arlo1 said:
Amber this needs to be isolated!

That's why I posted the circuit!

Nope, Prius driver stage takes care of this for me. I just reap the fruits of its labour :). It is similar to Volt driver stage, even better protected.
 
Lebowski said:
Arlo1 said:
Amber this needs to be isolated!

That's why I posted the circuit!

Depends on how far reaching you want your power stage blowup to be :wink:

Arber, i dont really understand the distinction beween the 400 and 650, but the controller ic needs to see the power stage supply voltage.

:?: I guess my comment was not clear enough. I will use Toyota Prius Gen 3 power stage. It has dual IGBT bridges and DCDC buck-boost stage between battery (0V - 400Vdc) and motor capacitor (0V - 650Vdc). Each of those caps has voltage measurement attached that produces proportional 0 - 5V ISOLATED signal.
My idea is to use MG2 output as drive for Leaf motor and use MG1 outputs and DCDC section as buck charger from 580Vdc rectified 3phase to 400Vdc battery. Sadly 13V regulator which is also inside doesnt work above 300Vdc.

It is one very compact and affordable unit.

Some drive analysis
https://techno-fandom.org/~hobbit/cars/ginv/VH.html
 

Attachments

  • ckt.gif
    ckt.gif
    10.6 KB · Views: 3,666
Lebowski said:
Why build this boost converter and not just use field weakening ?

Arber, you want to measure after the boost converter, so at the power stages.

Yup boost converter = fail. This is only added losses!

Just because someone else did it doesn't meen its the right way to achieve a goal.
 
Haha guys slow down please... i want to ABuse Prius inverter.

I will not use it as it was intended. Instead i want to keep top transistor open to conduct in both directions and use Lebowski brain with FW to power the motor.
BUT while car will be charging i want to use this halfh bridge to charge the battery from high source - 3phase rectified. I will switch top transistor and use the diode from the lower one as flyback. Inductor is already inline and liquid cooled. Perfect.
I can use one arduino nano for this... i will have to observe both voltages as well as another current sensor though.
Otherwise the MG2 output will work as every other inverter. This is why i think i will use only voltage sensor close to battery.
Everything in there was already built, i just have to use it in a convenient way...
 
i've been wanting to get the lebowski controller setup for 2wd on my sb cruiser trike for quite a long time. part of my problem is a powerstage--i don't have the skill to just design and build one "from scratch", but i'm going to need one (two, really) that can handle 3-4kw+ each to each run one of the hubmotors in the sb cruiser's rear wheels. i have a pair of the pcbs (from bobc?) to use the integrated powerchips, but i haven't found a version of them that will handle what i want out of them.


so...a question (or more likely a series of them):

once i have a pair of brain boards working, i would like to adapt an existing powerstage from another controller, such as a bricked kelly or similar, to be run by one of the brain boards. (i figure there are enough people here on es with bricked kellys laying around that i might be able to buy a couple identical or at least close-enough for reasonably cheap, and disconnect the mcu/etc inside so the lebowski brain can run everything instead).


i'm not really expecting it to be plug-and-play, but it should be (haha) easier for me than designing and building something from scratch, since i don't want to spend months (or years more likely) and tons of money tuning a powerstage design--i would like something that will work with the least experimenting and kentucky-fried-fets. ;)

(at one time i would have relished the challenge...nowadays i just don't have enough time and energy for that; if there was a place to buy a a complete assembled and working lebowski controller that just worked, and i could afford it, i'd do that. )


anyway...i know that the powerstage would need to have (or have installed) compatible current sensors for each phase, and would need compatible gate drivers, or to have the existing ones modified to work compatibly (which i would need help figuring out).


so...is this possible?
 
Yes, totally possible. Probably what I would do. I don’t have any Kelly controllers to look at but if you post good enough pictures we can figure it out. The stock gate drivers should be fine. You want to carefully check out all the FETs on the bricked ones to see if any are blown.
 
thanks--i don't have any yet to open up; it's just a hypothetical. had to verify the potential for the idea first, then i have to find a pair someone (or a couple of someones) is willing to let go for next to nothing (so i can afford the project).

(there's a pair in the for sale section i'd actually asked about and then somehow forgot, almost a year ago, apparently. i can't actually remember why i didn't get them at the time :( :oops: ).
 
I would not use a Kelly controller for this. Their powerstages are designed like crap and very prone to failing because the MOSFETs don't share current well.
 
well, they're the only one i know of that can likely be had cheap with a working powerstage but dead brain (because it's easy to brick some of the kellys just by spinning the motor while it's hooked up to a programmer, among other various ways to do it). I've seen some as low as $50 for two (last year), which is kind of my price range. ;)

i've got a fair few 12fet generic controllers that work but could be used for this, but i expect their powerstages are worse than the kellys.

if you have any other ideas where i might get a pair of similar powerstages for a 2wd (both rear wheels) delta trike, for next to nothing, relatively speaking, i'm all ears. ;)


i have a (working, afaik; can't test it myself) sevcon off a zero mc, don't remember the model number, if it's powerstage could be adapted (or if it could be traded for a couple of powerstages that *would* work). but adapting that still only takes care of one; i'd need another one for the other controller.

i have a box of assorted sets of fets, and some dip-style gate driver ics, and various other parts leftover from my brief but dramatic-failure-filled controller design foray. :oops: but anything i build out of them is likely to suffer the same fates my previous attempts did back then, since it would pretty much have to be handwired on perfboard, right now.
 
Back
Top