PDA

View Full Version : Throttle mapping: anything you want? Really?



teddillard
11 September 2013, 0244
I've seen two guys post in the last few days that you can make the throttle on an electric drive train do "anything you want". Correct me if I'm wrong, but my understanding is that the throttle controls the PWM (pulse width modulation, or the duration of the pulse which determines the speed of the motor) of the controller. Sure, you can control the curve of the throttle output and thus determine how much RPM you're going to get at any given throttle position, which determines your torque at that throttle position, but that's a far cry from "anything you want".

to wit:

The thing is, though, can't you simply program the throttle to act any way you want?


Don't forget we can control the power, oops, torque delivery however we want. I'm particularly interested in working on the throttle mapping based on road speed, throttle opening and the rate of change for the throttle. It shouldn't be too hard to do with a fully programmable ECU, it might just be hard to find the right configuration so it doesn't feel too alien to the gas bike riders.

("ECU" for the record, refers to Engine Control Unit" or "Electronic", depending on your preference, and is pretty much specific to ICE: http://en.wikipedia.org/wiki/Engine_control_unit. It can control the mixture, the valve timing, the ignition timing and idle. Not stuff we have on an electric drivetrain, obviously.)

In the first place, we're talking throttle position, not motor RPM, bike speed, or anything else. Now a more sophisticated controller than mine, that has regen for instance, you can control the application of regen with some throttle configurations, can't you?

But there's a huge difference between what "throttle position" means on an electric bike and a gas bike. Accelerating out of a corner, for instance, with a gas bike the throttle is going to have immediate response because the motor RPM is pulled down to the bike speed at that gear. On my bike, for example, without regen, the motor is still overspinning the throttle position if I'm off the throttle in the deceleration, and so there's a considerable amount of twist I have to apply to the throttle to get it up to giving the motor juice again. Regardless of my curve on the throttle, the throttle position has to match the motor RPM to have any effect.

Not to mention the type of programming you can do with fuel injection... do we really have that degree of control?

I've only had personal experience with PMDC motors and controllers, so maybe there's some magic sauce with an AC controller? I don't see how that's possible though.

Tell me what I'm missing here.

Mike Edwards
11 September 2013, 0630
What's the actual question?

On an ICE SuperBike, just as on an electric SuperBike the link between the throttle and the engine is the ECU. The two do not have to have a linear relationship, hell, on a V4 ICE you have two sets of throttle bodies that can be doing totally different things, e.g. one can be controlling the engine braking while the other is completely closed, or one can be implementing the first stages of traction control whilst the other remains at the original position.

If all your throttle does is give a basic input in to a basic controller then the above doesn't hold true. It doesn't mean it couldn't, it just means it currently doesn't.

teddillard
11 September 2013, 0640
The comparison between an ICE drivetrain and an electric drivetrain breaks down because of the lack of engine braking, transmission and clutch, as far as I can see.

ECU is a term I've only heard used in an ICE drivetrain, for the record. Is it commonly used for an electric drivetrain too?

My basic question is, can you in fact control "anything you want"? I think that's a pretty meaningless statement, since you obviously can't control some things. So, let's make it a more meaningful question. What, in practice, can you control with an electric drivetrain, and how?

Let's see some specifics.

Ken Will
11 September 2013, 0640
Most controllers today are pretty dumb!

What you need is a intelligent unit to control the motor controller. (Intelligent Controller Environment or "ICE")( a ICE bike you can enjoy)

You can get a 3D gyro/3D accelerometer board for $65 and an Arduino or Raspberry Pi for $50.
With a little programing magic you can make it do anything. You could even have a button for automatic wheelie mode!

frodus
11 September 2013, 0837
Let's maybe separate this a little from standard pwm series/pmdc motor controllers and ac/Bldc controllers.

DC controllers aren't that dumb but all they can really do is control voltage and current. Its not a 1 to 1 with throttle position and pwm percent.... I normally see them in torque mode... So they're controlling and regulating current output. Some have options to control some aspects of the control algorithm... Like throttle map, ramp rate, torque or speed control, etc.

Sevcon and Curtis ac controllers (which I've got more experience with over DC controllers) are very configurable. They can control:
Rpm
Current/torque
Voltage
Ramp up rate
Ramp down rate
Off throttle/neutral regen
Brake regen
Direction change
Hill hold (holds vehicle still while on hills)
Inch forward
Gear soften
Max rpm
Idle rpm

Also you can set up the throttle curves so it is sensitive at the beginning and trails off.... Or it starts soft and goes to sensitive.

So with a VCU (vehicle control unit.... Like an ecu) you can control exactly how a motor behaves.... And even implement traction control, regen based on steering/ lean angle, acceleration based on steering/ lean angle, wheely control.... Etc. So there's lots you can control with these motors. The sky is the limit really. You can control just about any aspect of their movement..... How far... How fast... Which direction... How long... How much torque... Etc.

podolefsky
11 September 2013, 0847
[edit: posted same time as Travis...so what he said plus....]

The quote I remember was "torque delivery any way we want." If you have a controller with true torque control, that's pretty accurate. Torque control uses a feedback loop so that the throttle controls current rather than voltage to the controller - and current = torque. Yes, it's all done with PWM, but with torque control the controller varies the PWM signal to get the torque commanded. It's different from a "dumb" controller where throttle just changes PWM duty cycle (read: voltage).

Sure - there are limits, but it's way more than what you can do with a dumb controller like Alltrax. ("dumb" being a technical term for pure PWM voltage control with no feedback loop - technically an Alltrax isn't even speed control, since it doesn't have a speed sensor)

So with torque control, you can set a custom torque map, make the map vary with speed, pitch and yaw if you have an accelerometer like Ken said. Switch the map on the fly. Pretty much anything you can program. My Curtis is basically a programmable computer attached to a set of switching transistors. With the right equipment, you can reprogram the software, add features and use the analog and digital inputs for all sorts of stuff. It has a lot of capabilities.

teddillard
11 September 2013, 0851
OK cool. Some questions, though.


DC controllers aren't that dumb but all they can really do is control voltage and current. Its not a 1 to 1 with throttle position and pwm percent.... I normally see them in torque mode... So they're controlling and regulating current output.

Sevcon and Curtis are far more configurable. They can control:
Rpm
Current/torque
Voltage

--How are these different from the "voltage and current" you mentioned above? Do you mean it can read the RPM and control that in a different way than simply the voltage?

Ramp up rate
Ramp down rate

--Isn't that the same as the throttle curve I can control with the AXE? Or do you mean something different?

Off throttle/neutral regen
Brake regen

--What's the difference between these two types of regen? I've never heard the term "neutral regen".

Gear soften

--What the hell is that? lol!


Also you can set up the throttle curves so it is sensitive at the beginning and trails off.... Or it starts soft and goes to sensitive.

--Again, isn't that just a curve, like on the AXE?


EDIT: I just re-read your post and got that you're saying "Some have options to control some aspects of the control algorithm... Like throttle map, ramp rate, torque or speed control, etc. ", so is that the answer to my questions above, that yes, they are similar?

On the subject of a VCU/ECU - keep in mind my bikes (and my brain) are in the vintage of 1980 or so - is that unit integrated in the bike's system already? Like, for example, an R1 - would that have all that? How do you interface your motor controller to it?

Noah, sorry, just saw your post crossing mine... on the "dumb" Alltrax - is that because it's a PMDC controller - it's just the nature of the beast, or is it a limitation of the model? Do the new Alltrax SPP controllers have more, uh, control? (ie, not as dumb)

frodus
11 September 2013, 0932
Sevcon and Curtis are far more configurable. They can control:
Rpm
Current/torque
Voltage

--How are these different from the "voltage and current" you mentioned above?* Do you mean it can read the RPM and control that in a different way than simply the voltage?*

Exactly, While AC controllers also control voltage and current, the algorithm to control them is different. You’re not just using PWM as a square wave as a means to control voltage like a DC controller, you’re using PFM, or pulse frequency modulation, to control not only voltage and current, but the RPM (the frequency portion of PFM). It’s all inter-twined, you’re controlling PWM of a waveform to control current and voltage, but also the Frequency to control the RPM…. Hence the use of PFM.



Ramp up rate
Ramp down rate

--Isn't that the same as the throttle curve I can control with the AXE?* Or do you mean something different?*

It’s different. Ramp up rate is the fastest rate of change you can go from one RPM to the next…. i.e. if your ramp rate is 300RPM/s, then that’s the fastest the motor RPM will change, regardless of throttle telling it what to do. Without it, you could tell a motor to go from 0 to 6000RPM, and you may go to into current limit trying to get to that RPM. So imagine it as an acceleration governor, rather than an RPM governor. The ramp down is the same, the max rate you can regen or change direction.



Off throttle/neutral regen
Brake regen

--What's the difference between these two types of regen?* I've never heard the term "neutral regen".

Off throttle is basically like engine braking. You let off the throttle, the control goes to Neutral (neither forward or reverse or braking) and the motor has a regen percentage associated that it will try and slow down without any throttle or braking input.

Braking regen is an actual brake input (on/off or an analog 0-5V signal) that will force the vehicle into a regen state (usually a higher percentage than neutral braking) so that it slows down. Brake overrides throttle as well as neutral braking.



Gear soften

--What the hell is that?* lol!

Basically it controls how hard the motor will try to turn at first, so that gear slack/chain slack is taken up, then the ramp rate controls the RPM acceleration. This helps keep things less jerky.



Also you can set up the throttle curves so it is sensitive at the beginning and trails off.... Or it starts soft and goes to sensitive.

--Again, isn't that just a curve, like on the AXE?*

Yes it is, I just didn’t want to disclude it from the discussion, since it can be adjusted. In either torque or speed mode, you can map the throttle how you wish. Just like in SOME DC controllers. Not all of them have this adjustment though, only programmable controllers.



On the subject of a VCU/ECU - keep in mind my bikes are in the vintage of 1980 or so - is that unit integrated in the bike's system already?* Like, for example, an R1 - would that have all that?* How do you interface your motor controller to it?

Neither of those bikes would have a VCU that you would be able to use for this application. Most bikes only have an ECU that takes in engine sensors and has some outputs like spark plugs, throttle, fuel pump, etc. The Curtis and Sevcon however, are controlled via CANbus. What you’d need for your bike, if you wanted advanced controls implemented (like traction control, wheelie control, etc), is to buy a dedicated VCU and program it yourself for your needs for anything outside of the standard programming that Alltrax/Curtis/sevcon have within them.

teddillard
11 September 2013, 1018
...and WHY have I never heard of PFM before this? :O Seriously.

There's not even a whole lot on it online - a pretty thin entry in the Wikipedia. It still uses a square wave, but uses it in bursts kind of - rather than one big square pulse, you're getting a burst of closely spaces thinner pulses - thus the "frequency" part... is that a fair layman's way to describe it? Is it called something else, perchance?

Where would a mere mortal buy such a VCU?

podolefsky
11 September 2013, 1044
Here's a nifty picture to show how a sine wave is made with PWM. Technically it's still PWM, not PFM, because the pulses are constant frequency (same time between pulses). It's just that the pulse width is modulated twice - once to set the peak amplitude, and again to set the shape (like a sine wave). That second modulation also sets the frequency of the sine wave, which is related to the RPM of the motor.

(It's different for PMAC (synchronous) and induction (asynchronous) motors - for synchronous, the sine wave frequency is the same as the motor RPM. For asynchronous, they're not the same. To produce torque, you need slip (http://en.wikipedia.org/wiki/Induction_motor#Slip) - the sine wave from the controller has to be faster than the motor RPM.)

http://www.pstracks.com/wp-content/uploads/2012/12/fig4.gif

frodus
11 September 2013, 1051
Its a mix of pwm plus a sort of carrier frequency if you will... But... AFAIK it's always called pfm.

As far as a ecu... I doubt many people have experimented with advanced control... Those that so would really want to implement would likely develop themselves.

Neweagle does sell them but they're blank... You'd have to either buy Dev Software or pay to have them implement something.

I know Brammo has a VCU to comment between. Bms, controller, charger and the display.

podolefsky
11 September 2013, 1100
Not to get too far off topic - but - I guess you could call it PFM. Technically, here is the difference between PWM vs PFM. I always thought AC motor controllers used PWM since they usually have a rated operating frequency (but I could be wrong).

PWM - pulse widths change, spacing of pulses (frequency) is the same
PFM - pulse widths are the same, spacing changes

http://electronicdesign.com/Content/15001/61115-fig-4.jpg

frodus
11 September 2013, 1108
I guess for this discussions sake it'd be a mix then. When I went to school for signals, systems and transforms it was pfm and pwm and still is AFAIK. But yes... Its all pwm With a carrier frequency. You can have both though... Varied pulse width and varied pulse frequency.

So... It can get pretty complicated as far as control goes.

podolefsky
11 September 2013, 1127
So... It can get pretty complicated as far as control goes.

That's for sure. Torque control on a variable frequency induction motor is black magic.

Nuts & Volts
11 September 2013, 1224
Our OSU team uses this with the Tritium (http://tritium.com.au/products/ev-driver-controls/). The Tritium only has a CAN input line. Thus we have to use this to converter our throttle input into CAN. But we can program this to do a ton of things. We could have a wheel speed sensor that is always checked. If speed is <50mph maybe we limit throttle to only 80% of max. We basically would just change the current setpoint (limit) to 80% of it's max so that 100% throttle is only 80A not 100A to the motor now. All kinds of things we can do really easily. Other controllers like a Sevcon, Rinehart, or Curtis you can just add in extra sensors and then change part of the code to use these inputs to control something. Only issue is that sometimes the control software is locked. A VCU is a neat idea, but I think it's better when you can just do it directly with you controller since it already has the hardware to do it. Thou if you have different components (ie different BMS or charger or controller) a central VCU is nice as it just needs slight code variations to handle any new components

Mike Edwards
11 September 2013, 1410
This is turning in to quite an interesting discussion. Clearly at the outset there were some assumptions as to quite what the implementation was looking like which may have skewed the understanding of what is possible. It's the equivalent of comparing the ECU on your Honda CBR600 with the one on a full factory SuperBike.

Our goal is to take the (very) sophisticated ECU from our British SuperBike, reprogram it to write new firmware to make it suitable for electric vehicles yet keep the tried and tested strategies for traction control, launch control, anti-wheelie, etc.

Whilst we may be using a very capable controller it will never be smart enough so it will be confined to being much dumber than it could be so that we can leverage the capabilities of the brain to understand and interpret all the various sensors, whether connected via CAN bus or via a direct connection, and give us the level of control we are looking for.

By the time you add in the ECU, the dash, the wiring harness, the controller and the motor you have to be looking at upwards of 40k USD, if not more. The original statement was about what could be achieved rather than what is achievable by reprogramming the set up of your Zero or Brammo. It's certainly going to be an interesting exercise putting it all in to practice.

teddillard
11 September 2013, 1427
capabilities of the brain

You mean by this the "brain" of the bike's ECU? Just trying to follow along...

Very cool stuff, Kyle, thanks for chiming in!

Richard230
11 September 2013, 1446
I just finished reading the October issue of Popular Science, which features a cover titled "Revolutions" and a photo of the all-electric Toyota TMG EV P002 Pikes Peak racer. This issue contains several articles discussing the future of the automobile and includes an article about the Toyota and its success at Pikes Peak. In the article they mention that the car has a torque vectoring system that can be electrically programed to vary the speeds of each wheel's electric motor in order to provide the best traction when cornering, braking or accelerating. They feel this is a superior system to the mechanical traction controls used by IC cars. The article says that the car came in third in the race and they expect electric race cars to beat IC cars in the race in the near future. Then they go on to mention that the Lightning already is the fastest motorcycle up the Mountain. :)

podolefsky
11 September 2013, 1652
Interesting, I didn't know the TMG did that. I saw that the Mercedes SLS electric uses independent "torque vectoring" on each wheel. In addition to precise traction control on each wheel, they can actually apply negative torque to the inside wheels around a turn (negative meaning opposite the direction the wheels are turning). It's like having just the inside wheels brake, which literally spins the car around turns. Apparently works so well some people consider it cheating.

Doesn't really apply to motorcycles though...

Mike Edwards
12 September 2013, 0029
Without a differential every electric car needs a degree of torque vectoring or it won't make it around the corners very well.

It's most common in the 'one motor per wheel' set up and can be very effective if you have the time, resources and inclination to take it further.

teddillard
12 September 2013, 0313
You mean by this the "brain" of the bike's ECU? Just trying to follow along...

Would it be too much to ask for you to take a minute and clarify this comment? Sorry if it seems just too dumb a question.

Mike Edwards
12 September 2013, 0348
Would it be too much to ask for you to take a minute and clarify this comment? Sorry if it seems just too dumb a question.

Yep, I'm talking about a stand alone ECU, in this case a Motec M170 that we used on our British SuperBike Suzuki GSXR1000.

ARC EV Racing
12 September 2013, 1457
Whilst we may be using a very capable controller it will never be smart enough so it will be confined to being much dumber than it could be so that we can leverage the capabilities of the brain to understand and interpret all the various sensors, whether connected via CAN bus or via a direct connection, and give us the level of control we are looking for.


Mike I think this assumption is worth a bit of discussion, what makes you say that the controller will never be smart enough? Do you mean the controller you're thinking of using, or that no controller is smart enough?

We've done a lot with Curtis VCL and before that I worked with instrumentation for ten years so I have some valid experience here. The Curtis platform is very flexible and you might be surprised at what it's capable of, I certainly was.

Sent from my AZ210A using Tapatalk 2

Mike Edwards
15 September 2013, 1234
Mike I think this assumption is worth a bit of discussion, what makes you say that the controller will never be smart enough? Do you mean the controller you're thinking of using, or that no controller is smart enough?

We've done a lot with Curtis VCL and before that I worked with instrumentation for ten years so I have some valid experience here. The Curtis platform is very flexible and you might be surprised at what it's capable of, I certainly was.

I agree, there are some quite impressive controllers but do you really want the controller to monitor the battery temperatures and voltages, map the throttle input to the torque requirements, take the inputs from all the sensors to provide the traction control and other strategies, drive the dash and any other displays and provide data logging for all of the above?

I guess I am just coming at it from the other side of things. Anything that needs to do anything intelligent should be something more standalone, particularly if it then has the potential to become a commercial product :-)

podolefsky
15 September 2013, 1329
I agree, there are some quite impressive controllers but do you really want the controller to monitor the battery temperatures and voltages, map the throttle input to the torque requirements, take the inputs from all the sensors to provide the traction control and other strategies, drive the dash and any other displays and provide data logging for all of the above?

I guess I am just coming at it from the other side of things. Anything that needs to do anything intelligent should be something more standalone, particularly if it then has the potential to become a commercial product :-)

I think the point was that you can do quite a lot with a Curtis controller. Not exactly what you would do if you could design everything from the ground up, but you rarely get that choice in the DIY market.

Otherwise, good point - it's really an engineering / design question. You could certainly build everything into one box and call it the "controller". Or make it all modular, so you have a variable frequency drive module, a BMS module, a VCU module, etc. If you're building an e-assist bicycle, you put it all in one box because it's easier to attach to the bike, and if something goes wrong the whole thing only costs $100 to replace. If you're building an $60k EV, it's all separate because each part is $1000s to replace (and probably made by different manufacturers down the supply chain). A $10-18k electric motorcycle is probably somewhere in between, even a commercial one.

ARC EV Racing
16 September 2013, 0417
I agree, there are some quite impressive controllers but do you really want the controller to monitor the battery temperatures and voltages, map the throttle input to the torque requirements, take the inputs from all the sensors to provide the traction control and other strategies, drive the dash and any other displays and provide data logging for all of the above?

I guess I am just coming at it from the other side of things. Anything that needs to do anything intelligent should be something more standalone, particularly if it then has the potential to become a commercial product :-)

This is Brendan - Not Matt who normally contributes on here.

Why wouldn’t you want the controller to do all that? If it’s capable of course. I wouldn’t want another box doing it just for the sake of keeping it separate, that s just extra weight and wiring looms and another load of risk. The side of the controller that's 'programmable' is essentially a separate thing to the base control in any case. The Curtis (and most other high end units) controller can do all the things you have listed there so why would I want to pay however many 1000 for a Motec unit? Also your BMS does the work of monitoring and tells the controller directly how to act, the controller can quite easily read the BMS boards also and act accordingly.
Motec and similar engine management systems came about because the factory one’s were built to a low spec so they offered an upgrade for racing. The controllers we see now have a functionality equivalent to this as a base and I’d argue that it would be of benefit to use this, the Motec unit cant directly control your motor so is just acting as a middle man.
In my opinion there is no commercial product from a standalone unit, because there are standalone units such as Motec already for sale. The only novel thing is the algorithms, which most people develop to suit.

Mike Edwards
25 September 2013, 0902
Why wouldn’t you want the controller to do all that? If it’s capable of course. I wouldn’t want another box doing it just for the sake of keeping it separate, that s just extra weight and wiring looms and another load of risk. The side of the controller that's 'programmable' is essentially a separate thing to the base control in any case. The Curtis (and most other high end units) controller can do all the things you have listed there so why would I want to pay however many 1000 for a Motec unit? Also your BMS does the work of monitoring and tells the controller directly how to act, the controller can quite easily read the BMS boards also and act accordingly.
In my opinion there is no commercial product from a standalone unit, because there are standalone units such as Motec already for sale. The only novel thing is the algorithms, which most people develop to suit.

Why would you want to programme a controller when you can have the algorithms in a separate box that be plugged in to almost anything?

How much time and money gets wasted with everyone developing their own controller code to do what they need it to do ever though other people have already done it elsewhere? How much did it cost to develop the skills and experience to program the controller? How much do you charge per day Brendan if someone wanted you to develop something for them?

However much the Motec, or equivalent, costs I would say it represents good value for money, particularly if it provides all the additional performance strategies, BMS monitoring, data logging, dash interfacing, etc., etc.

podolefsky
25 September 2013, 1455
As far as I know, there's nothing like the Motec for evs, so at this point controller programming is the only choice.

Also, isn't the Motec m170 like $5000? An ac-35 and Curtis controller kit is like $4000. Once someone has developed the firmware, it's just a matter of flashing the controller, and programming new maps and such is just hooking up a computer. If the controller will handle all that, why spend 2x the money and add another box to do the same thing?

Maybe I don't understand what you guys are arguing about...

Mike Edwards
26 September 2013, 0232
The original discussion was that anything is possible and that most controllers either cannot match the functionality or require too much work to bring it to the same level. It appears to have turned in to a discussion that you can make the Curtis controller do anything that a Motec could do if you have the time and the skills.

I can't help but think that the cost of time and effort in bringing a Curtis controller even half way to a Motec, if one were available, would result in the home brew version being more expensive in the long term. I'm guessing of course that nobody tends to share their Curtis development for the benefit of other Curtis controller users.

ARC EV Racing
26 September 2013, 0233
Exactly - the code has to be written somewhere, whether it's adapting a race ecu for an ev or adapting an ev controller to act as a race ecu. Once it's written the bulk of the work is done and it's 'just' a matter of setting parameters for the specific application.

I can see the benefit of Mike's approach if the idea is to have something universal that will plug into any elmoto, regardless of controller type. We've tried building universal ev accessories in the past and it's just not the same as with ic bikes. The best you can really do with an elmoto is mess with the throttle signal, or I guess you could manipulate sensor inputs but I'm not sure where that would get you.

Oh and since you asked Mike I think I charged Hugues about £50 do develop custom code for his controller, but that's because he's a nice guy with an interesting project.

Matt

Sent from my AZ210A using Tapatalk 2

ARC EV Racing
26 September 2013, 0244
The original discussion was that anything is possible and that most controllers either cannot match the functionality or require too much work to bring it to the same level. It appears to have turned in to a discussion that you can make the Curtis controller do anything that a Motec could do if you have the time and the skills.

I can't help but think that the cost of time and effort in bringing a Curtis controller even half way to a Motec, if one were available, would result in the home brew version being more expensive in the long term. I'm guessing of course that nobody tends to share their Curtis development for the benefit of other Curtis controller users.

Yep we've gone a bit off track. My point was really just that there's no reason why a controller couldn't have hardware capable of doing what an external race ecu could do. I didn't mean to suggest that we all go off and develop our own version of code though, any more than I think you would suggest people cracking open their motec boxes and flashing their own code directly onto the micro. If there's a market for the functionality then there should be a product, whether it's a separate unit or a software upgrade is another matter

Sent from my AZ210A using Tapatalk 2

Hugues
26 September 2013, 0420
..
Oh and since you asked Mike I think I charged Hugues about £50 do develop custom code for his controller, but that's because he's a nice guy with an interesting project.

Matt

Sent from my AZ210A using Tapatalk 2

LOL
and the nice guy is reall happy with the code, just managed to make my LVC interface work between Orion BMS and Curtis controller, and happy to share the wiring and instructions with whoever is interested, just let me know :)
the system would work with pretty much any BMS with an LVC signal out,

podolefsky
26 September 2013, 0550
LOL
and the nice guy is reall happy with the code, just managed to make my LVC interface work between Orion BMS and Curtis controller, and happy to share the wiring and instructions with whoever is interested, just let me know :)
the system would work with pretty much any BMS with an LVC signal out,

The newest Curtis firmware from HPEVS interfaces directly with the Orion over CAN, and has some other nifty things like eco mode switch input. Probably just a matter of time before they have other features. As far as I know the cost of their systems didn't go up.

ARC EV Racing
26 September 2013, 0640
The newest Curtis firmware from HPEVS interfaces directly with the Orion over CAN, and has some other nifty things like eco mode switch input. Probably just a matter of time before they have other features. As far as I know the cost of their systems didn't go up.

Hugues' code has two power modes operated by a switch, and I think a 'limp home' option for when the BMS LVC is triggered. Mostly that code was about using his Harley throttle and making it fail safe but I based it on some code I'd already written and left in the features. Interesting that HPEVS have that now.

Our next race bike has can interface to the Manzanita BMS (via our can - serial hardware bridge). It also runs an engage display over can and an energy management algorithm that I'm developing. Last year we had launch control; max drive current and auto shift between first and second gear for better race starts. That was fun.

Sent from my AZ210A using Tapatalk 2

Ken Will
26 September 2013, 0734
Single unit VS modules:

Back in the 70s several people I knew bought "entertainment centers". They were beautiful pieces of furniture that contained every thing you needed for spectacular music. Some of us on the other hand bought separate tuners, amps and turntables. When we wanted to upgrade we just replaced one component instead of the whole shebang. I was into the latest high-tech so I added an 8-track player!

I also at one time had a 76 El Camino, that I loved, but, it wasn't as good a truck as a pickup and wasn't as good a car as a Chevelle.

So, I feel a dumb power module and a separate brain is the way to go. You can upgrade each separately and when the power module gets hot it will not affect the electronics in the brain.

teddillard
26 September 2013, 1254
Yep we've gone a bit off track.

Well, I, for one, don't mind a bit. Great stuff... good information - take it wherever it goes. :cool:

frodus
27 September 2013, 2019
So if someone builds a VCU that could communicate (CAN or serial, analog outputs to control throttle, analog inputs from throttle/position sensors, etc).... and then put some starter code online (kind of like Arduino) and let people develop their on stuff.

They could do things like:

- have lean sensors that could measure lean angle and then set the max regen on the motorcycle a little lower to prevent loss of traction.
- Log Controller, BMS, etc to an SD card
- Connect wirelessly to the VCU and talk to a display for gauges
- Control lights
- Control the Eltek chargers
- Control the J1772 signals properly
- Act as an antitheft device that would disable a controller unless a smartphone device was present
- etc etc (basically a handful of IO, some CAN, RS232/485, Ethernet, SD card, Bluetooth/Wifi)

Would it be of any interest to people?

Like a hardware device that is open source so people can develop on it?

Mike Edwards
28 September 2013, 0251
FYI Motec do actually have a development kit to allow you to write your own firmware, or modify theirs...

teddillard
28 September 2013, 0309
Just so some of us can follow along - the MoTeC m170 site is here: http://www.racedatasystems.com/motec-m170-magnesium-ecu.html

MoTeC Australia has a really informative site with training videos: https://www.youtube.com/user/MoTeCAustralia

...and here's a vid showing use:


https://www.youtube.com/watch?v=KCbsdkaa2do

This ****'s crazy, man!

(FYI, I've asked that this thread be moved to "Controllers" to be easier to find, since it's such a great thread.)

ARC EV Racing
28 September 2013, 0609
So if someone builds a VCU that could communicate (CAN or serial, analog outputs to control throttle, analog inputs from throttle/position sensors, etc).... and then put some starter code online (kind of like Arduino) and let people develop their on stuff.


That would be cool, and compatibility issues could be solved by providing modules to interface the standard code to a particular controller type. Kind of like a software driver per controller.

Sent from my AZ210A using Tapatalk 2

ARC EV Racing
28 September 2013, 0613
FYI Motec do actually have a development kit to allow you to write your own firmware, or modify theirs...

That's interesting, who do they provide those to? And do they provide the source code of their firmware as a basis for modification?

Sent from my AZ210A using Tapatalk 2

Mike Edwards
28 September 2013, 1133
That's interesting, who do they provide those to? And do they provide the source code of their firmware as a basis for modification?

If you can afford an M170 and the extra for the toolkit then I guess anyone can have it.

I doubt they supply source code for everything, more likely to be libraries for the core functionality plus code for the basics although I'm guessing thought as the toolkit is only just being released.