Paul and Sabrina's Cheap 3 Phase Inverter (AC Controller) with Field Oriented Control

[e*clipse]

Paul, thingstodo -

These are some great videos and data of the debugging process - thank you very much for posting all of it.

I'm a bit confused about the regen mode, could you clarify some things?

Is the DC motor driving the rotation of both motors?

It seems that when the rotation gets to zero, it then reverses - is this true?

Is the net current between the AC motor and the battery pack flowing into the battery at all times during regen? If not, it's more of "reverse torque" than pure regen - does it switch modes?

If so, (if it does reverse) what happens with the current near 0 rpm?

- Thanks,
E*clipse

[thingstodo]

Originally Posted by e*clipse

Paul, thingstodo -

These are some great videos and data of the debugging process - thank you very much for posting all of it.

I'm a bit confused about the regen mode, could you clarify some things?

Is the DC motor driving the rotation of both motors?

Yes. The motors are coupled together with a joy coupler

It seems that when the rotation gets to zero, it then reverses - is this true?

Yes. The negative torque command (throttle or potentiometer) slows down rotation using the AC motor, then the direction reverses and the AC motor continues to 'accelerate' in the negative direction

Is the net current between the AC motor and the battery pack flowing into the battery at all times during regen? If not, it's more of "reverse torque" than pure regen - does it switch modes?

In the first part of the test, the 24V pack is putting power into the DC motor, which drives both motors. Using the throttle, I command negative torque on the AC controller, which is done by lowering the frequency put out by the AC motor. That puts the motor into a 'generator' mode. That energy output from the generator flows back into the AC controller, becomes DC, charges up the capacitors to a higher voltags, and supplies the power used by the electronics on the AC controller. Since the voltage is now above the voltage of the battery pack, the surplus power flows back into the high voltage pack.

When the 24V pack is turned off and the inertia of the motors turning is overcome by the negative torque on the AC controller .. there is no more energy to feed 'back' into the AC controller. The DC bus voltage drops back to the battery pack voltage. The AC controller starts to consume energy from the battery pack again.

That likely does not make much sense .. but that's how I understand it.

If so, (if it does reverse) what happens with the current near 0 rpm?

The DC current into the battery pack drops until it is near zero, then the battery pack starts to supply current to the AC controller.

I don't have a way to correlate at what rpm that happens ... I suspect somewhere around 300 rpm but that's just a wild guess.

Perhaps I can rig up some cameras to synchronize measurements for the next test so that everyone can see what is going on.

[e*clipse]

Originally Posted by thingstodo

Yes. The negative torque command (throttle or potentiometer) slows down rotation using the AC motor, then the direction reverses and the AC motor continues to 'accelerate' in the negative direction

In the first part of the test, the 24V pack is putting power into the DC motor, which drives both motors. Using the throttle, I command negative torque on the AC controller, which is done by lowering the frequency put out by the AC motor. That puts the motor into a 'generator' mode. That energy output from the generator flows back into the AC controller, becomes DC, charges up the capacitors to a higher voltags, and supplies the power used by the electronics on the AC controller. Since the voltage is now above the voltage of the battery pack, the surplus power flows back into the high voltage pack.

When the 24V pack is turned off and the inertia of the motors turning is overcome by the negative torque on the AC controller .. there is no more energy to feed 'back' into the AC controller. The DC bus voltage drops back to the battery pack voltage. The AC controller starts to consume energy from the battery pack again.

That likely does not make much sense .. but that's how I understand it.

Ok, it seems to be doing what I thought it was - thank you.

What's interesting is that when commanded negative torque when the motor is turning a particular speed can result in energy being fed back to the AC controller's supply. So it appears that the system doesn't really "care" if there's a "regen braking" command or if the controller's basically given a "reverse" command.

So it seems that somewhere around 300rpm, the BEMF produced by the motor reduces to some insignificant amount and the controller has to supply power in order to maintain the negative torque command, thus "regen" turns into "reverse" at that point. If the system is purely regen ( power fed back to the supply ) then that would be helpful for extending range. However, if battery power is used to stop the vehicle, then it seems more efficient to use friction brakes at that point.

It seems a set-up like this would be nice:
When torque commanded is equal to or greater than the torque required to maintain speed, the controller would consume power from the supply. When the torque commanded is less than the torque required to maintain speed, then then the controller would "regen" power back to the supply. This is essentially what is happenning with a "positive" throttle. It would also "feel" similar to an IC car when downshifting.

If more negative torque is needed, then the driver can hit the brakes. It may be possible to put a sensor on the brakes - either a pressure sensor or a brake pedal position sensor. When the brake pedal is pushed, then the regen could be increased. This reverse torque would only be available when energy flows to the supply. Perhaps the controller could monitor the current flow and basically "ignore" the regen command when it results in power flowing from the supply.

However - and this is where things get tough from a user-interface perspective - the regen torque will be proportional to speed. As this decreases due to the reducing speed, the braking torque from the friction brakes would have to be increased. Most drivers want very predictable brakes, where the negative torque is proportional to brake pedal pressure. Balancing/tuning this will be a challenge, but very rewarding.

- E*clipse

[thingstodo]

Originally Posted by e*clipse

So it seems that somewhere around 300rpm, the BEMF produced by the motor reduces to some insignificant amount and the controller has to supply power in order to maintain the negative torque command, thus "regen" turns into "reverse" at that point. If the system is purely regen ( power fed back to the supply ) then that would be helpful for extending range. However, if battery power is used to stop the vehicle, then it seems more efficient to use friction brakes at that point.

I think that is what EVTV has done with the brake hydraulic pressure sensor. Pressing on the brake does regen braking proportional to the pressure, before the friction brakes engage. If the regen is not 'enough' to stop you, you put a bit more 'hip into it' and use the friction brakes.

It seems a set-up like this would be nice:
When torque commanded is equal to or greater than the torque required to maintain speed, the controller would consume power from the supply. When the torque commanded is less than the torque required to maintain speed, then then the controller would "regen" power back to the supply. This is essentially what is happenning with a "positive" throttle. It would also "feel" similar to an IC car when downshifting.

If more negative torque is needed, then the driver can hit the brakes. It may be possible to put a sensor on the brakes - either a pressure sensor or a brake pedal position sensor. When the brake pedal is pushed, then the regen could be increased. This reverse torque would only be available when energy flows to the supply. Perhaps the controller could monitor the current flow and basically "ignore" the regen command when it results in power flowing from the supply.

However - and this is where things get tough from a user-interface perspective - the regen torque will be proportional to speed. As this decreases due to the reducing speed, the braking torque from the friction brakes would have to be increased. Most drivers want very predictable brakes, where the negative torque is proportional to brake pedal pressure. Balancing/tuning this will be a challenge, but very rewarding.

At least - that's what I have read. I am not in a position to confirm it ... since my car does not run as yet (and it's DC not AC). Maybe my next project!

[MPaulHolmes]

Look what the swedish people are doing!:

ELBIL :: Lser mne - Byggtrd Paul&Sabrina 200kw ac-controller

I sent them 3 of those control/driver boards and now they are starting the soldering.

[Astro]

Originally Posted by thingstodo

I think that is what EVTV has done with the brake hydraulic pressure sensor. Pressing on the brake does regen braking proportional to the pressure, before the friction brakes engage. If the regen is not 'enough' to stop you, you put a bit more 'hip into it' and use the friction brakes.
At least - that's what I have read. I am not in a position to confirm it ... since my car does not run as yet (and it's DC not AC). Maybe my next project!

That is the set up i am hoping to implement.
I want it to mimic a gas engine as much as possible. The way i look at it, i have been driving a gas vehicle for decades so the way they behave has been integrated into my driving habits. So why change habits when i can change the cars design.
My thoughts are to have the throttle pedal rest position cause slight regen. That way lifting the gas pedal will cause a slight amount of regen, mimicking the drag of gas engine at zero throttle. Once that is set up and running then a pressure sensor is added to the brake line which pulls the throttle voltage lower as the brake pedal is pressed thereby increasing the amount of regen. If more aggressive braking is required then pressing even harder on the brake pedal will engage the friction braking in addition to maximum regen.

[thingstodo]

Originally Posted by Astro

My thoughts are to have the throttle pedal rest position cause slight regen. That way lifting the gas pedal will cause a slight amount of regen, mimicking the drag of gas engine at zero throttle.

Is this a programmable setting or are you using 'solder' to adjust the regen?

Once that is set up and running then a pressure sensor is added to the brake line which pulls the throttle voltage lower as the brake pedal is pressed thereby increasing the amount of regen.

So you want to use a single input to the controller, which varies from small regen to full acceleration. And the pedal rest position must be a high enough signal to allow for another circuit to 'pull the signal down' for more aggressive regen. Did I get that right?

If more aggressive braking is required then pressing even harder on the brake pedal will engage the friction braking in addition to maximum regen.

What happens if you have a Toyota throttle problem - unintended signal for large acceleration - as well as brake signal? Do they cancel each other out?

I like the idea for a very simple interface for throttle and brake on one wire.

I'm having trouble visualizing how a hall effect throttle can give decent performance for you unless you have access to the firmware of the controller to adjust regen slopes or lookup tables.

[thingstodo]

My updates are *BADLY* behind - this one started Dec 6, with more video taken Dec 11 and 12

Setting up liquid cooling for the Siemens AC motor.

Very simple system with a small flow rate.

https://youtu.be/O4Hy9f3hqL0

[thingstodo]

As I said - my updates are *BADLY* behind - this one ALSO started Dec 6, with more video taken Dec 11 and 12

I decided to see what else I could use as thermal paste to transfer heat from the heat spreader on the AC Controller to the large aluminum plate that was originally part of a Better Place battery pack. This is what I came up with.
https://youtu.be/VZzPtBo4i1k

links

Nutella as thermal paste

https://www.youtube.com/watch?v=EwBfet9_v4A

toothpaste as thermal paste - why not?
Thermal transfer compound comparison

The next set was taped on Dec 21. I skipped shooting more video this week and am editing instead

[e*clipse]

Originally Posted by thingstodo

Is this a programmable setting or are you using 'solder' to adjust the regen?

Do you want to use a single input to the controller, which varies from small regen to full acceleration. And the pedal rest position must be a high enough signal to allow for another circuit to 'pull the signal down' for more aggressive regen. Did I get that right?

What happens if you have a Toyota throttle problem - unintended signal for large acceleration - as well as brake signal? Do they cancel each other out?

I like the idea for a very simple interface for throttle and brake on one wire.

I'm having trouble visualizing how a hall effect throttle can give decent performance for you unless you have access to the firmware of the controller to adjust regen slopes or lookup tables.

I can't answer for Astro, but I think these questions bring up some very valuable points.

I think it would be best to have the user input completely software adjustable. This would allow easy adjustment for all throttle and brake behavior. I saw a review of the Lightening motorcycle where they spoke of the "accessable" power of the bike. The torque didn't come on immediately with a tiny twist of the throttle. This would make the bike very difficult to ride. I've also seen non-linear throttle curve parameters on controllers. Again, this will make the power and torque more accessable.

Regarding the Toyota "unintended throttle" problem and Hall-effect sensors. I think hall-effect sensors are more reliable than standard potentiometers. All standard potentiomters have a finite life span due to the wiper contacting the resistor. I just replaced the potentiometers on my stove because they produced unreliable output at about 25%. Hall-effect sensors do not have this issue. Also, hall-effect sensors generally operate between 0.5V and 4.5V, out of a 5V range. This gives the opportunity for software to test for open or short circuits if the voltage is out of range.

In addition to that, the Toyota throttle pedal has two hall-effect sensors. The outputs are arranged physically so that they look like a hystresis curve. This allows the code to continually check between the two to make sure there is no issue with the hall-effect output. I like redundant systems with easy error checking.

A similar setup with redundant sensors could be implemented on the brake pedal. At the very least, it should have an output range like the hall-effect sensor to detect correct operation vs a short or short circuit.

- E*clipse