[Mike Dabrowski 2000]

I wanted to bounce something off you guys:
Assembly languare programmers always look for the simplest way to solve a problem reliabily. To write a tight assembly program for throttle control, means trying get one more control multiplexed on top of the standard throttle = more power less power control.
The second control function is the electric / gas power mix proportion, or level of assist/regen .
This is the information I am generating from the throttl pedal
I take 1/8 second samples of the throttle A/D value which is being sampled several hundred times a second. I do a running average of 8 of those samples.(1 second) Any throttle change will yield 4 numbers.
1. The running average of the last second.
2. The last throttle reading
3. The rate of change (difference between the average and the last reading)
4.The throttle position with 0ne part in 1000 resolution.

If you press down slowly, the difference can be as little as 3-4 even though the throttle may have moved from full off to full on, you did it slowly, so the average was almost able to keep up with your rate of change.
If you press down fast, you can generate almost 255 of difference,with just a short motion, since the value is rate of change rather than positional based. As soon as you stop moving the throttle, it settles back to zero in a sec or so.
One system of using these 3 numbers to control MIMA by throttle could go as follows:
1. cruising at 55, @70MPG+, in lean burn. See a hill coming up, want to climb it mostly MIMA powered. I slowly back off the throttle (there will be a speed of throttle rate of change that will not activate MIMA call it the dead band of throttle rate of change.) when I want to activate MIMA, I step down quickly (full assist) not so quick ( 3/4 assist) medium (1/2 assist) slow (1/4 assist). A quick slight backing off of the throttle and a return to the same lean burn point would turn MIMA assist off. Regen would work in a similar way as you lift the throttle, the speed that your foot moves at determines magnitude of regen.
Change your foot position slowley, and you can go from idle to full throttle, and never trigger assist, and the same in reverse as far as regen.
System tuning variables for throttle will be The assist dead band, assist (Magnify), the regen dead band and regen (Magnify)are settable with the left right joystick, and will be retained until you change them again.
That is how it is looking right now. Any ideas suggestions? Comments?

 

[Mike Dabrowski 2000]

The throttle based mode that I have described is only one possibility. A copy of a PM that I recieved about this :

There are three things that concern me regarding this mode.

1. It is very difficult for most people to have a sensitive and accurate foot. I try to slide my foot to the edge of the pedal and lock it into the carpet allowing the resistance of the carpet against my shoe to create a pivot point so that I can rotate my ankle slightly to gain tiny incremental changes in throttle position. Without doing this I find it virtually impossible to maintain a steady position of the throttle. Brian uses a throttle lock to achieve the same result.

2. Backing the throttle off when you are wanting assist seems counter intuitive and in fact the car will begin to slow, further convincing the mind that it is doing the wrong thing. It is difficult enough to master the art of blipping the throttle positively when down shifting to enable regeneration.

3. The G force in the car will be steady state then negative then positive. This might prove uncomfortable to a passenger. Accelerating before decelerating would give some passengers a real scare too.

4.In an emergency the car might react as if it had a mind of it's own, which of course it does in this case.



If I had to do it I would merely make the IMA a lot more sensitive to throttle action. I'm assuming that the Insight is slowing down on account of a hill. Move the pedal towards the floor and the IMA responds dramatically. The IMA would drop out again if the torque of the electric motor stopped the deceleration. (The opposite of this already happens when you touch the brakes.)

Second opinion:
> Mike, it's simple. Hold the throttle position steady with your foot, engage the Mima mode, if the car slows by a threshold amount down add assist, if the car speeds up by a threshold amount add regen.

The mode you are describing sounds more like the cruise control mode.



> Moving the throttle would add assist or regen while in transition and for a brief time after transition. In this mode the engine would act like it hqd a governor on it and changes to the throttle would merely establish a new set point. This is the way small engine governors are set up on a snow blower for example.

HAving assist only during the transition would not maintain assist for the duration of a hill climb, which is what needs to happen.

>
> The advantage of this setup would be that every action would be intuitive and if the battery was being drained , backing off the throttle position would automatically reduce, halt, or reverse the tenancy.

I see where you are coming from, but think that you may need to play with MIMA a little to tune your Intuition? once I get things all connected and prove the signals, I can play with different approaches to see what works best.

>
> Increasing the threshold would increase the efficiency but allow a greater speed variation.
>
> Adding a throttle lock or cruise control to this would be a no-brainer.
>
> Hope this makes sense.

This is the time for any opinions,
Whenever you are driving your car, do a mental simulation of these techniques, and continue to give me your opinion.
Thanks

 

[chrs]

I can't get excited about any of the modes described here. It's hard enough to hold the throttle right at the 75, 100, or 125 mpg point, without trying to simulatanously send telegraph messages with my foot. I'd rather use my hands to push buttons to set modes, and use a manual lever for analog control.

I haven't gotten through all the prior MIMA messages, so appologies if I'm missing some context.

Charlie

 

[Mike Dabrowski 2000]

Remember that the present IMA control is through the throttle, and no one has much problem using it. What I am looking for here is opinions as to how one could improve the throttle/IMA interaction by allowing the IMA to be controlled with variable aggressiveness, so it can be made to be the first power source to come on line when acceleration is needed, and be the first to absorb kinetic energy when deceleration is needed, in an adjustable way.
This subject was discussed a while ago, but now is the time to offer opinions.

I installed a complete new MIMA harness into my car this afternoon, and am feeling better about the level of difficulty for the install. I also think it can be made to look stock except for an extra bump here and there in the already bumpy harness. I am taking my time and documenting as I go, and presently feel that with a few more installs to tune the procedure, that MIMA could be installed in 2-3 hours, from start to finish all covers replaced.

 

[Rick Reece]

Mike,
Thats a great install time. I use a idle adjustment knob for a constant throttle position so as a result the ability to add regen or assist with a joystick really meets my needs. My goal is to keep the engine in its most efficient load range. Have fun, Rick

 

[chrs]

I use a idle adjustment knob for a constant throttle position so as a result the ability to add regen or assist with a joystick really meets my needs.

Rick,

That sounds really useful. Are there details in the thread somewhere that I couldn't find?

Thanks

Charlie

 

[chrs]

Remember that the present IMA control is through the throttle, and no one has much problem using it. What I am looking for here is opinions as to how one could improve the throttle/IMA interaction by allowing the IMA to be controlled with variable aggressiveness, so it can be made to be the first power source to come on line when acceleration is needed, and be the first to absorb kinetic energy when deceleration is needed, in an adjustable way.

OK, got it. So here's what I'd want: I'd want the IMA torque to be a linear function of the throttle position, with gain set by a knob. And I'd want to have a button I could hit to tell it to have zero assist/zero regen at the present throttle position. So I get to cruising speed and throttle position, and hit that button. Then, depending on where the gain knob was set, I'd get either a lot of IMA kicking in with a little throttle change, or about the amount I get now, or less.

Optionally, there could also be a deadband, which is usually the case with the OEM controls. That is, there's a range of throttle that gives no assist or regen. That could be a second knob, or maybe just a selector switch (yes or no).

Charlie

 

[Calpod]

What I'd like to see on a throttle based MIMA is to start getting assist at 100MPG on the bar instead of 40-45mpg, and start getting regen at 150 MPG just before fuel cut mode, so you don't have to shut off the throttle, you would be able to keep the engine at it's most efficient for a longer period of time.

It should be adjustable so you would start getting regen sooner, maybe at 140, or to delay it.

 

[Mike Dabrowski 2000]

Chrs
"OK, got it. So here's what I'd want: I'd want the IMA torque to be a linear function of the throttle position, with gain set by a knob."
A linear function of throttle position? does that mean as soon as one leaves idle and starts pressing on the throttle that the assist would activate, or would one get to target speed then activate?

"And I'd want to have a button I could hit to tell it to have zero assist/zero regen at the present throttle position. So I get to cruising speed and throttle position, and hit that button."
The MIMA mode one (joystick control) will cut all assist and regen when activated. Assist and regen only activate when the stick is moved, so the MIMA standby performs this function.

"Then, depending on where the gain knob was set, I'd get either a lot of IMA kicking in with a little throttle change, or about the amount I get now, or less."
This sounds like the way my first proposal would work, but I could change the magnify just by the speed that I moved the throttle.
We have no gain knob, I was plannng to use the left and right motion of the joystick(2 axis) for that function. The front to back is the MIMA manual control.
Pressing the joystick down (central switch)will store the magnify value in flash memory, and will be available until a new value is entered.
Calpod
"What I'd like to see on a throttle based MIMA is to start getting assist at 100MPG on the bar instead of 40-45mpg, and start getting regen at 150 MPG just before fuel cut mode, so you don't have to shut off the throttle, you would be able to keep the engine at it's most efficient for a longer period of time.
It should be adjustable so you would start getting regen sooner, maybe at 140, or to delay it."
Thats the way I hope it will work, but we do not directly have the MPG information.
I hope to approximate the MPG by looking at the MAP signal, but it will not tell me the actual MPG, only the manifold pressure. Rick Reese and the turbo guys have vacuume guages on their cars, which should be close to the MAP in function. I remember Rick indicating that the vac approximately reflected the MPG bargraph in many conditions.
I was not planning on using the MAP input (it will be connected and available), for this first pass at throttle based control, so you as the operator will be responsible for keeping the MPG where you want it.
The variable aggressiveness of the throttle based control as I see it would allow assist at any MPG, the trick is how to let it know with the throttle when and how strongly to come on line.
Rick
" the ability to add regen or assist with a joystick really meets my needs. "

Armin said the same thing, just give me the Joystick, thats all I need.

The throttle based MIMA can and will work, but Honda did not develop their control in a couple of weeks, and we may not be able to improve on it without some hands on MIMA experience from several users to help tweak the system into perfection.
Keep the discussion going, and we will zero in on a good way for the first throttle based MIMA to operate, and then we improve it through use and feedback.

 

[james]

I'm having a bit of trouble imagining just how the throttle-based control would feel - the more so since I haven't driven in a couple of months - but I think I'm on the skeptical side. I think of it more from a system/psychology perspective: MIMA seems to be sort of like getting a bunch of extra gears, and it's natural (ok, learned habit) to shift gears with your hand. So the joystick/mode buttons on the gearshift just make it a more complicated shift, they don't really require a new behavior.

I see I haven't really expressed it well, but I can't think of a better way and, as I said, I'm not at all sure myself.

 

[Mike Dabrowski 2000]

After driving with MIMA, it is still difficult to imagine this even for me, that I why I am looking for the feedback, to refine the imaginary throttle based control.
I may be able to actually turn on the MIMA_C in the car today, and watch the throttle signal with the MIMA. I have set up a test program where the difference between the average throttle signal and the current throttle signal is displayed on the battery leds in binary so that I can watch the detection programs output. That should let me have a better idea of if that approach would be able to work. A suggestion was made by PM to me that it may be better to combine the VSS information (cruise control) and the throttle, into the smart cruise control system that we have looked at as an option. That way the throttle would set the nominal MPG and speed, turn on the MIMA in this smart cruise mode, and the system turns on assist to maintain speed at a hill, and turns on regen to maintain speed down hill. maybe a 2-4 MPH deviation is allowed. The aggressiveness of the IMA actions could be tuned.
That sounds pretty transparent, and would be bretty natural.

 

[Rick Reece]

Chrs
Here is a link for the engine idle adjustment knob.
http://www.insightcentral.net/forum/viewtopic.php?t=591&highlight=idle+knob

 

[Mike Dabrowski 2000]

I got the TPS and MAP final wiring to the ECM finished and documented, and managed to make the additional wires hide in the existing harness pretty well.
I took a ride to watch the signal off the MAP sensor, and found that it was a very sensitive indicator of engine load. Decelerating with engine braking gave .65V, 125 MPG at steady speed and slope gave about .85V. the interesting thing though was that if I came to a small hill, maintained the throttle setting, the MAP would raise to 1V or more, but the MPG stayed the same. Set the MPG at 100MPG constant speed flat surface, and the MAP was about 1V. WOT MAp voltage was about 2.5V, and autostop MAP sat at 2.65.
Remember the A/D in the micro can measure changes of as little as .005V, so the voltage changes I am seeing on the MAP sensor are big signals, and may be able to indicate lean burn or not lean burn quite well. If one wanted to make a lean burn indicator, the MAP signal would be a good one to tap into.
I think that as we refine the future MIMA modes, we will find that John's suggestion of measuring the MAP signal was a good one.
Got the temp probes made and installed on the Battery and MPI, so tomorrow will be the big day, the first firing up of MIMA_C in the car.

 

[Insightful Trekker]

I think that as we refine the future MIMA modes, we will find that John's suggestion of measuring the MAP signal was a good one.

Since I am being quoted in such glowing terms ...

While I an not in the market for MIMA until its next generation allowing "better" batteries, I think the very complex issues in regard to the assist and charge algorithms are now clearly apparent.

And each individuals driving circumstances will dictate a somewhat different algorithm e.g. midwest, _flat_ vs. the hiLLy northeast. I would suggest that some type of adjustable bias controls be designed that will allow users to tweak the parameters to their needs is what will solve all the issues. Yes, MIMA by the "gas" pedal is the way to go, but we are discussing replacing Honda's preprogrammed algorithms with another. Until there are enough miles logged under various conditions I don't know what would be considered "better". And like several earlier posts in this thread its _very_ abstract to do simply by number. You need some time behind the wheel to "feel" what the numbers mean.

You'd therefore need adjustments for both charge and assist's start points, rate (probably linear in-between) and end point. With biases for throttle angle sensor input (TA), speed (VSS) and MAP. Once the initial MIMA equipped Insights log enough miles this data then can be gathered and subsequent MIMA boards _could_ have the proven limits for the parameters preprogrammed.

I think I made the same suggestion much earlier on, and AFAI remember there is some type of "technical" limitation as to the circuitry's design that makes offering this very difficult. Even if its only DIP switched in 10% increments I think this added flexibility will be proven invaluable.

Sorry if I am repeating an unhelpful suggestion.

Anyway, HTH!

 

[e*clipse]

Throttle Based MIMA Control

I would agree with John's suggestion about leaving some tuning parameters open to driving conditions/style.

I have a CVT, and it's response/needs are different than those of the manual. Unless there is a manual/CVT MIMA this may need to be addressed.

Primarily, I am looking for a boost in performance for hilly areas. To be able to adjust a background increase in (separately) boost/regen with a boost override for would be ideal. (lead finger - I guess) ;-)

Another thing to consider is the possibility of someone else driving a MIMA equipped Insight. IMHO it should still perform somewhat like a normal vehicle.

By the way, if you want to measure things more accurately with the 10bit A/D on the 18F452, you can use a different reference voltage than the supply voltage for the uP. For example, if you use 2.65V as your reference voltage, you would be looking at minimum step sizes of 0.0026V. Additionally, if consistant performance becomes an issue when looking at something as sensitive as the MAP data, Microchip makes some nice reference voltage chips. :wink:

 

[highwater]

e*clipse wrote:

I have a CVT, and it's response/needs are different than those of the manual. Unless there is a manual/CVT MIMA this may need to be addressed

The CVT question was brought up in the main MIMA thread. (Its down a few spaces in this forum header, hasn't been hit in a coulpe of days as we seem to have split into 3or4 related topics.) Mike had said the CVT had a 90% chance of working but I could not find where anyone had ever sent him any technical copies from the CVT manuals, so looks like this remains unanswered.

e*clipse wrote:

Another thing to consider is the possibility of someone else driving a MIMA equipped Insight. IMHO it should still perform somewhat like a normal vehicle

This question has surfaced several times. Answer is still: There is a switch to turn MIMA off so everything goes back to stock configuration. (Rule #1 when flying another mans aircraft: Don't mess with red covered switches :shock: )

 

[Mike Dabrowski 2000]

A CVT insight owner is waiting for his service manual, and will drive up from PA so we can try MIMA on his car. After that we will know what the isues are and if it can work.
MIMA switches back to standard IMA at a touch of a sw, or on cutting power to the device.

 

[e*clipse]

Thank you for the clarifications.

My statement for "normal driving" MIMA needs to be clarified. The comment came from a situation where my wife broke her arm in horse race. While she was hospitalized, someone else drove the car home.

She probably would have had the "red switch" engaged upon shutdown. Trying to explain disabling MIMA might be a little much for someone who's never driven a hybrid, even if its a "red switch." :?

Suggestion: In the CVT, on the steering wheel, there is a pair of switches for "sport" and "economy" mode. They are momentary switches, and as far as know, the car defaults to "economy" mode. Why not make the "MIMA" switch momentary contact, like those switches? It would allow a IMA default on startup and there are already status indicator lights. 8)

OK, now more on topic:
In rolling hills going into mountains, we've been able to <almost> maximize the capabilities of the battery pack. It would be very nice to be able to adjust in more electric boost <as necessary> via joystick or whatever for the climb to be able to maintain whatever speed or mileage you are trying to attain. On the CVT if you slow down for the hill, often you get no electric boost. :x For the downgrade, to decision would be based on state of charge of the pack and length of grade. You <may> want to do a little regen even with the throttle engaged just so your pack can be ready for the next hill. This all seems independent of motor performance alone.

 

[Mike Dabrowski 2000]

E*clipse

By the way, if you want to measure things more accurately with the 10bit A/D on the 18F452, you can use a different reference voltage than the supply voltage for the uP. For example, if you use 2.65V as your reference voltage, you would be looking at minimum step sizes of 0.0026V. Additionally, if consistant performance becomes an issue when looking at something as sensitive as the MAP data, Microchip makes some nice reference voltage chips.

This approach would work if all of the analog I/O were 0-2.5V, but the TPS, bat amps,Aux Joystick, ambient light, are all 0-5V signals. The MAP, and two temperature signals still have more than enough resolution for this application.

Why not make the "MIMA" switch momentary contact, like those switches? It would allow a IMA default on startup and there are already status indicator lights.

The switches are momentary contact, and do default to normal IMA mode

Still to early to know if MIMA is even going to work with the CVT.

 

[Jackmpg]

MIMA

Hi Mike,
I have seen your work in action, your not your ordinary engineer, you are many steps above. Just do what it is you do and it will be fine, you do not need to defend yourself, again, I have seen your work in action. There is no opportunity for advancement without some risk, more for non-engineers like me, much less advanced engineers like you. I trust people I can do business with on a hand shake, I've met you, I know I will buy your product, Jack

ps I know someone who is one of the smartest engineers on the planet who never spent a day in college, self taught. I have seen smart, you are one of them. If CVT's can be done, you'll find a way.