Ampera/Volt inverter with Lebowski brain

Well i tied to run Mitsubishi Outlander rear motor with Lebowski brain and Volt inverter. Results are satisfying, but motor still conks out at start.

Parameters:
L = 210uH
R = 180mOhm estimated
 
Well Bas if you will make a hex file for us with 45deg force start, i suggest it would be best to post it on source code thread.
Do you think 45deg start would be best used with hall sensor assisted sensorless? My car will weigh 1.5t so motor will require some torque to move. I like how that sounds "assisted sensorless":).

thanks

Arber
 
arber333 said:
Well Bas if you will make a hex file for us with 45deg force start, i suggest it would be best to post it on source code thread.
Do you think 45deg start would be best used with hall sensor assisted sensorless? My car will weigh 1.5t so motor will require some torque to move. I like how that sounds "assisted sensorless":).

thanks

Arber

The 45 deg has to do with how the control loops operate, but not with what the eventual phase current angle will be. It is still FOC... i will pm you and kiwifiat the hex when I have had time to look at it, and only post it in the original thread when it is confirmed to work. Again, i lost my test setup as now I use my hobbydesk for home office...
 
I can see also that this compressor would be usefull as a motor on a fast bike or scooter. You would need to spin it at 4000rpm and use at least 160Vdc but it would produce some 2 - 3kW. I cant see any other motor at $100 value :).
 
Hi Bas

Can you tell me what CAN transciever have you had in mind when you built the code? I have tried with TJA1050 and it doesnt work. Actually i get problem with TX RX serial lines. I cant receive anything. If i disconnect power to the module serial starts to work normally. Strange... any clues?

I should check with MPC2562 chip also...
 
Well Bas i have tried your new 45deg starting logic and it is a success! I can power the motor in any case from any throttle position. It is just at start you have to be gentle with throttle, but whan the motor is spinning you can vary throttle as you like.

This brings me to a question. How does that tie to hall assisted sensorless? Would that be a good idea to use or not? On my Leaf i have RLS commutation sensor and i can use pure sensor mode. I just want to know if sensorless can bring any benefit here.


tnx again it works really nice on AC compressor motor. I will try on Outlander motor again and then on my Leaf motor inside my Peugeot.
 
The new 45 degree thing has nothing to do with motor starting or sensored versus sensorless. Independent of sensored/sensorless the algorithm uses control loops to control the throttle and fieldweakening current. The 45 degree thing is a change in the operation of the control loops. It affects how good the control loops function. It does not change the throttle and fieldweak currents achieved, as these will still be what the user requests. For more information have a look at the thread where I uploaded the code, the 45 degree thing is explained there and in the list of improvement recommendations.

Hall assisted sensorless again is something independent of this. Motor start can be sensored or sensorless (as specified in the main menu) but at higher speed the motor will always run sensorless. In this sensorless operation it keeps track of motor speed and position. When you have a very sudden change in motor speed (like from full speed to low speed in less than a few 0.1 seconds or something like that, really fast speed change) the part of the control loop that keeps track of motor speed cannot keep up. This will then lead to conkout. So what it does in hall assisted sensorless is that it no longer keeps track of motor speed (in a sensorless way) but takes this information from the hall sensors. Motor POSITION however is still controlled sensorless, so FOC correctness is still excellent as this is still sensorless.
 
Ok i managed to get PUTTY terminal to copy and paste HEX settings! Now i can test various motors and quickly recover settings if something goes wrong. :idea:
I noticed though there is no menu to enable 45deg option. What happens if i now choose hall sensors will they be effective at all upwards of mode 2?

Here is the run up of the motor...
https://www.youtube.com/watch?v=r0wZcXttehc
I had to reduce eRPM since the according to him motor is 3pole with 6 magnets in rotor.
https://www.diyelectriccar.com/forums/showthread.php?t=140346&highlight=prius+compressor+inverter
 
arber333 said:
Ok i managed to get PUTTY terminal to copy and paste HEX settings! Now i can test various motors and quickly recover settings if something goes wrong. :idea:
I noticed though there is no menu to enable 45deg option. What happens if i now choose hall sensors will they be effective at all upwards of mode 2?

Here is the run up of the motor...
https://www.youtube.com/watch?v=r0wZcXttehc
I had to reduce eRPM since the according to him motor is 3pole with 6 magnets in rotor.
https://www.diyelectriccar.com/forums/showthread.php?t=140346&highlight=prius+compressor+inverter

The 45 degree things has nothing to do with halls, sensorless etc etc. It is something in the control loops to make it more stable and remove the sensitivity to haveing the correct resistor. Remember how you had to change the resistor by hand to get the motor to not conk out ? That should now be gone. There is no setting for this.
 
I am trying to get CAN bus working. And whenever i turn CAN transciever on i cant talk to Serial port. Is that normal?

Also how is it with CAN rx/tx lines? Do they have to be swapped the same way as UART needs to be? So chipRx to CANTx and chipTx to CANRx? Or like this?
scheme_can.gif


tnx

A
 
The CAN as implement now is a custom protocol only meant for talking to other controller ICs... it does not conform to any standard so wont work with normal automotive CAN stuff.. but you can always change the code and make this yourself.
 
Oh it seems i may have saturated the CAN pins as well as serial.
I will go and read about this in code. Maybe i can find some CAN library for Pic.
 
I tried running the Outlander motor with new firmware 2B0. I can see it is more stable just by running it without load.
But when starting from still motor wiggles anyways. I thought it would just turn into one direction.

I will try and connect inverter to Leaf EM61 motor in my car now. I will test it under load.
I will report back when i have results.
 
arber333 said:
I tried running the Outlander motor with new firmware 2B0. I can see it is more stable just by running it without load.
But when starting from still motor wiggles anyways. I thought it would just turn into one direction.

I will try and connect inverter to Leaf EM61 motor in my car now. I will test it under load.
I will report back when i have results.
You must not have the sensorless start tuned optimally. I find that with IMP motors like the Leaf and Hyundai I get perfect starts every time. I will post my sensorless settings tomorrow. I can't imagine the Outlander rear transaxle motor will be significantly different.
 
How did you tune up your motor for a good start. Which parameters are involved?

Tnx
 
Lebowski said:
Did you play with the wiggle angle and frequency in the misc menu ?

Yes quite some, but i couldnt eliminate the jumpiness of the motor at start. I can work it out, but i need a way how to approach this. Is there an approach, like "use 5Hz period and move in steps of 5deg" or something?
How would you call the minimum angular unit for 8 pole motor with 48 stator slots?

tnx
 
Lebowski said:
I'd try angle of 10, 20, 30 and 40 degrees, and 5, 10, 15 and 20 Hz for frequency. This makes 16 combinations to try...

That is precisely what I did except I went way higher on frequency. When I get back to my workshop later today I'll post my settings. With ebike hub motors the default settings work fine but IPM motors seem to respond better to larger angles and much higher frequency.
 
Below the relevant section for the Sonata HSG running on 48S ~ 200V on a Honda IMA inverter. The motor starts perfectly in sensorless.
With both the HSG and Leaf motor higher wiggle rate seemed to my real time statistics to work better.

sensorless start settings.PNG
 
Well i tested what you said and i used those values on my Prius compressor

f) wiggle range: 74 deg
g) wiggle rate: 119 Hz

I now get reliable start everytime with some ratcheting of course. But important is that motor wont just conk out if i apply too much throttle. Eventually it will start accelerating. This will be very important for my auto throttle on AC drive.

Thank you.
 
arber333 said:
Well i tested what you said and i used those values on my Prius compressor

f) wiggle range: 74 deg
g) wiggle rate: 119 Hz

I now get reliable start everytime with some ratcheting of course. But important is that motor wont just conk out if i apply too much throttle. Eventually it will start accelerating. This will be very important for my auto throttle on AC drive.

Thank you.

That is great news. It will interesting to see how it works out with your Outlander transaxle with its higher pole pair count compared to the Leaf and Sonata motors. I will be getting some PCB's made this week for a resolver to HALL converter for the Lebowski controller that uses a cheapo AVR ATMEGA328 processor. Usually takes about 2 weeks from OSH Park in the US.
 
Yes please let me know how it goes. I have great experience with simple Johannes resolver interface.
EDIT: I may put a dual Lebowski board in production. This one i intend to make with Prius resolver chip. @Kiwifiat, how did you find that chip operating in resolver 2 communtation mode on its own? I would provide oscilator and filters on circuit. Do you think it would be good to work with?

On the other hand i turned the wheels on my car with Leaf motor yesterday. It will spin!
But actually it didnt turn wheels smoothly in sensorless it kept trying to look backwards and then spun forwards. I will definitely need hall sensors to turn the wheels.
I also tried to use larger wiggle parameters and with wheels on i didnt get any different result. I will keep trying, but i will go and calibrate hall sensors now.

What i found from previous experience the PLL damping is great factor in smooth take off using hall sensors. So i keep mine above 0.85 with 100Hz.
 
Back
Top