Hybrid controls and the MIMA project

[Yves M.]

The MCM is the main hybrid control system:

The connectors and descriptions are in the Service manual at 12-9 to 12-13

The circuit diagrams are at 12-25 to 12-27

The Electrical Troubleshooting manual has the IMA circuits diagrams at 32 to 32-5


Some general info:
http://www.md92raid.100freemb.com/pub/i ... age92.html

The MCM section:
http://www.md92raid.100freemb.com/pub/i ... ge132.html

Some more info on hte MCM connections (and next page):
http://www.md92raid.100freemb.com/pub/i ... ge136.html

[Yves M.]

I have tried to measure the MCM A,B and C plugs. They are deep, small and a lot of pins.

Here are 2 pics:
http://yves.fungiart.com/images/misc/dscf0008.jpg
http://yves.fungiart.com/images/misc/dscf0010.jpg

I'll need to put the stuff in a readable format. But it does not look to be an easy task to replicate a serial plug for this. What do you think Mike.

[Mike Dabrowski 2000]

Hi Yves
The connector that I am proposing, would jack onto the back of the existing connectors where the wires come out when the connectors are inserted as normal. They will therefore be in parallel to the connector. There will be no need to unplug the connectors since I will be initially only monitoring the signals that are on the working system. Later once we see what the signals look like, we may need to get in series with one or more of the signals to inject our own control signal.
What I would need to make the test connector would be the back view of the connectors, with dimensions of the connector body and the pins from the wire side. Since all of the individual connectors jack into the large socket, it may be easier for me to just make one large connector that can connect to any or all of the 4 connectors.
Mike[/quote]

[Yves M.]

OK that makes sense. I'll get on it next week-end. This week-end, it was 3 degrees F while I worked on the car (cold).

The pins are all aligned and of identical separating distance. But the plugs and sockets are not symmetrical.
Some are 1, 2, 3, 4 in the upper/lower sections and 7, 10, 12,13 for the center "straight pins of lengths" Maybe bars of terminals for each possible length

I'll get back with the details next week

[Mike Dabrowski 2000]

Yves
One other thing to look at is where to route the wires from the connectors. If you look at the test adapter that I made for the battery connector, you will see that part of it is cut out to allow the wires to pass through it. We will need to do the same for this test connector. We should note which pins we will want to monitor, and which we will not. If need be, we can cut out the test connector and route the wires through areas where the pins are of no interest.
Mike

[Mike Dabrowski 2000]

Since looking at the possibility of getting the electric assist and regeneration under manual control, I have been doing mental simulations to see how I would use such a control, and it's likely effect on MPG.
The logic of using the gas motor for the torque of maintaining speed,(gas priority) means gas milage is always the first thing to suffer when needing a slight increase in torque to maintain speed. If the gas engine output was held at say 100 mpg, and slight throttle compensations instead invoked assist or regen, one could both maintain speed, and get high MPG.
Hopefully we may know the answer before too long.

[clett]

As for estimates of likely resulting mpg, I think we may actually have originally underestimated. The EnergyCS guys have finally ironed out the bugs in their Plug-in Prius (see here), and it looks like their electric-assist basically triples mileage to about 150mpg, so we could be looking at maybe 225mpg at the end of the day? But maybe we should push for 300!

I think the most interesting part of their conversion, however, is the choice of batteries. The Saphion LiIons are only ~100Wh/kg, but are inherently safe (no fires possible due to unique chemistry), long cycle life and can be easily wired up like traditional lead acid cells, so no fiddling about with loads of tiny 18650s (which would ultimately be much better, IMHO!). You can read about their safety and other benefits etc here.

It looks like the cells are nicely sized to make up a pack (see pic of their conversion below). That's a 9kWh pack in the picture.

http://www.evworld.com/images/energycs_pluginprius.jpg

The individual cells are packaged to be the same size as a lead acid 12volt, and wired up the same (see below):

http://www.ai-online.com/images/ace/...013_image1.jpg



[Moderator Note] Please use links for non Insight related images.

[Yves M.]

The terminals 4 and 12 on connector C and the terminal 3 and 4 of connector A was not conclusive

[Yves M.]

Hi guys,

Here are results from just putting the tip of a meter (has frequency and % duty cycle for pulse wave width) into the MCM "C" and the other tip to the ground. Those are terminals that are between the ECM and the MCM. As the manuals says the ECM does the request for IMA assistance/regen

Terminal 1, ENGTRQ Engine Torque seemed not to work best in %mode as it gave always about 10% duty cycle but in "frequency" mode it had something similar to:
Full assist ---------- 2 Hz
4 bars of assist --- 50 Hz
No assist ---------- 80 Hz


THE MOST IMPORTANT TERMINAL:
=====================
The terminal 2 named CMDPWR or command power had always this result

Full assist ----------- 90%+
4 bars of assist --- 70%
No assist ----------- 50%
4 bars of regen ---- 40%
All bars of regen --- 20%-


So the Terminal 2 of the MCM "C" named CMDPWR is the one that the ECM uses to
ask for more IMA
Without any IMA usage it was 50% always, with some assist the % was higher than 50% and with regen it was lower than 50%


Therefore I think that CMDPWR is one of the connections that makes the ECM call for the MCM assist/regen. It uses pulse width +- 50%

Maybe that using an oscillator (temporarely removing hte communication), varying the pulse width here could control manually the IMA to either Assist or Regen

[Yves M.]

The project is going to be called the MIMA as decided with Mike ( Manual IMA),
THANKS to him for the circuit I made a board with an oscillator to give a 2KHz wave frequency and a duty cycle of 36% to 76%. It is not the entire possible range but it was sufficient for the test. It actually worked down to 10% while regen but could not go higher than 76% as it stoped. It is only a prototype for proof of concept.

Here is the result AND IT IS GOOD, IT WORKS !!
I have built the first MIMA version 0.1. It has a switch to either get the normal IMA control or the manual MIMA control

Note that the runs where on local streets at about 50 to 60 KM/H:

1) When the engine runs on the MIMA alone at 50%, there is no assist and no
regen. Only the ICE, no error code.

3) For regen, wow! I could almost stop the car like I was using the brakes. I
was thrown into my seat belt when I used it the first time at full regen. The MIMA was able to go to 10%
and that "really" brakes. I had all bars of regen and more braking power.

4) For assist, it was better than normal but not as much a difference as with
regen. But the MIMA could not get higher than about 76% and would stop. With
the IMA normal, I could see that the duty cycle could rize to 90% so it is
why I could not get more than about half of the Assist. But it was at a time where I would have had none or not more than a few bars. A MIMA with a wave of
up to 90% could easily show better results as the IMA itself can easily go
above 76%


IMPORTANT:
==============
For total manual control, there is a need to control the Engine Torque Input
probably. I could not control the IMA assist without some pedal pressure on
the gas, probably to induce some torque input.

And the opposite is also true, I could not get Regen without removing some of the gas pedal pressure to remove some of the torque.
But I could get regen while driving at a point where I had to press the gas pedal not to slow down

Basically, I had to press the gas pedal a little before I could get Assist,
although more assist than the normal IMA.
I had to remove slitly my foot of the gas to induce some regen, but more regen
than the IMA alone.


So with a control that would shorten or stretch the wave, it would be just
perfect.
And for total control, the engine torque is needed.


It is really easy to tap into the ECM/MCM connections into
the "Manual Transmission Shift Console" All the wires are there and some
slack is present. It is also a best place to put a MIMA control with some
lever or butons.
Only 4 bolts and 2 pins to remove along with the Shifter knob

Here are the pictures.
Of the system at normal driving at 49.8% of duty cycle


And one of the board with the pot to allow from around 30% to 50% and about 75%


Here is MIMA v0.1 (cheap and quick)


So now, Mike will look into a controler that will have the input wave length and output the new wave length. To actually just improve regen or improve assist.
This would be MIMA version 0.2