Hmmm... and you will be needing a guinea pig for this?



Probably 10 times in the last week I was driving and something happened that made me think "wish I had MIMA."

 

I'll be testing it on my bench setup first, then the UK car when I borrow it in a few weeks, then you may be a US guinea pig.

 

You may need 2 guinea pigs to test out any differences

 

Well I can get the OBDIIC&C gauge to squirt through the IMA data reqd between the HCH1 ecm & mcm with the addition of one resistor, it's currently working on the bench, but at the moment I can't get it to do that and the gauge function at the same time. It was quite a big ask. Two seperate lots of serial data to deal with and only one usart. It might still be possible but only with a lot of work I will look at it.

So we have a couple of options.

1. Have a completely seperate IMAC&C device with it's own 16x2 serial enabled display. This would only require a 12F1822 8 pin pic and be cheap as chips as far as the hardware is concerned. 4 connections to the car reqd. +5v, Gnd, MotA ECM, MotA MCM

2. Use the OBDIIC&C gauge with two modes OBDIIC&C or IMAC&C until I get it sorted. Hmm

 

Wait, so the second board would need more than one harness splice?

Obviously having the software all work on the OBDIIC&C is the best solution... but that said:

I'd say being able to see the actual IMA data at the same time as controlling it is pretty much a requirement... otherwise all you have to go on is the dash gauge, and that won't show you things like voltage sag and temp and true SOC. Which makes the ability to control the IMA only half as useful.

I vote second board... however, does the second board even need a screen at all? Or are you saying the main box would be blank or say something like "IMAC&C mode" and only the secondary small board's LCD would be active?

If I understand correctly, the second board won't be plugged into the OBDII port, so you could have the OBDIIC&C plugged in and showing the IMA state and just hide the IMAC&C board completely?

Sam

 

An hch1 IMAC&C board would be a completley independent device with 16x2 serial lcd display for configuring it and operating data etc . It would have the standard 5 way switch button.

Lots to think about though, incorporating it into the OBDIIC&C would be good as it would have access to the brake/clutch/neutral switch and oll the other obd data without having to monitor those lines independently.

I'll keep working on the combined solution for now.

 

Yeah, I was just thinking of LCD screen propagation

I was also just thinking that pulling out that power window lockout switch from the driver's door armrest and letting the joystick stick up through that hole:

2004 Honda Civic Hybrid Sedan Door Panel Photos | GTcarlot.com

 

IMAC&C Data Analysis Openday

I have now processed and condensed the data I captured recently from the UK HCH1 Civic.

This data is the info being sent from the ECM > MCM over the MotA Line at 9600,8,E,1 when the car is on.

It contains the IMA Control data and is made up of one repeating 9 byte data packet sent every 30ms. The packet starts with the ident byte B3 and ends in a checksum byte, that means we have only seven data bytes to decode. .

The spreadsheet contains a number of short data captures covering a variety of driving and motor operation scenarios. We now need to look at the data and from the info make some inspired deductions about which bytes are controlling which IMA functions.

I making the data available here for the benefit of the community. It has taken many hours of work to obtain. I welcome ideas and analysis of the data. Feel free to modify and improve the spreadsheet I am no expert.

I also have the MCM > ECM MotB Line data but that is not processed yet and comprises four different 9 byte packets types, at least one of which will be devoted to driving the Oem dash gauge in a similar way to the METSCI line on the Insight.

www.solarvan.co.uk/Civic/CombinedFiles.zip

Enjoy

 

Mot A comms via OBDIIC&C working

Now got my 18F2620 running at 32mhz and it is keeping up nicely with the interrupt driven comms on the MotA line and the OBDII port at the same time

Stacks or work to do and I need your help with that data anlaysis.
This work benefits all IMA car owners as it increases our understanding of the cars systems. So you Insight gurus please chip in.

Next stage is to get the basic Calpod/Sam HCH1 OBDIIC&C working with this data passthru MotA comms in the background then I can start manipulating the new data.

Hope my guinea pigs are ready LOL!!!

Note to enable the OBDIIC&C to pass thru the MotA data on the logging port I changed R15 to 330R & added a 2k2 resistor between the data connector side of R14 and +5v on the board. I used pin1 of J4 for the 5v in my test setup. But you could use any nearby solder pad that is connected to the +5v supply.

Youtube video from the Perkins Labs.

 

ill take a look at it shortly... currently out of town with the wife. back monday. sam

 

Thanks Sam

I've modified the Calpod version of the hch1 OBDIIC&C software to do the passthru MotA stuff and will test that tomorrow. Calpod's responds the same as my bench UK ecm so i can test it with that.

You will both need to upgrade to an 18F2620 chip to use the next versions of the software I will release as these will include the MotA passthru tests.

 

Test 2

MotA passthru working with normal OBDIIC&C data and functions on display. Soc buttons and ecm/mcm version display also work fine

Just need to modify that B3 packet for IMA control

Has anyone apart from Sam actually downloaded the spreadsheet of data I posted and looked at it? I welcome any ideas.
I'm going to start studying it tonight with a bit of luck.

Youtube video from the Perkins Labs.

 

Does the brand on the 18F2620 matter? Is this the same model that will be in the I1's OBDIIC&C this summer, so should I order two?

Sam

 

Brand? Should all be made by Microchip ASFAIK :?
Yes it's the upgrade for the summer chip. Needs to be 0.3" width version

 

Thanks for the info, I'll get mine ordered, will this chip also work with the current Insight version?

 

It's the chip that will be reqd for the summer so get two if you have two boards.

You need new software for the Insight for it to work

For the Civic HCH1 it will also req new software with new pic.

 

Sorry for my ignorance, but which one of these, specifically, is the right one? I'm guessing it's the -E with the PDIP package? Is PDIP compatible with DIP?

Sample	Buy	Part Number	Leads	Package Type	Temp Range	Packing	1-25	26-99	100+	1000-4999	5000+
	 	PIC18F2620-E/SO	28	SOIC 300mil	-40C to +125C	TUBE 	5.36	5.23	5.10	4.70 *	4.47 *
	 	PIC18F2620-E/SP	28	PDIP 300mil	-40C to +125C	TUBE 	5.53	5.40	5.26	4.84 *	4.61 *
 	 	PIC18F2620-I/SO	28	SOIC 300mil	-40C to +85C	TUBE 	4.87	4.76	4.64	4.27 *	4.06 *
 	 	PIC18F2620-I/SP	28	PDIP 300mil	-40C to +85C	TUBE 	5.04	4.92	4.80	4.42 *	4.20 *
	 	PIC18F2620T-I/SO	28	SOIC 300mil	-40C to +85C	T/R 	4.91	4.80	4.68	4.31 *	4.10 *
 	 	PIC18LF2620-I/SO	28	SOIC 300mil	-40C to +85C	TUBE 	5.12	4.99	4.87	4.48 *	4.26 *
 	 	PIC18LF2620-I/SP	28	PDIP 300mil	-40C to +85C	TUBE 	5.28	5.16	5.03	4.63 *	4.40 *
	 	PIC18LF2620T-I/SO	28	SOIC 300mil	-40C to +85C	T/R 	5.16	5.03	4.91	4.52 *	4.30 *

Sam

 

PIC18F2620-I/SP 28 PDIP 300mil -40C to +85C TUBE 5.04 4.92 4.80 4.42 * 4.20 *

Looks right to me.

 

Ok, ordered two.

$11.00 for the parts, $6 for shipping... definitely why I wanted to get two at once. The shipping was the same for one.

Sam

 

The more I look at it, the more I think that setting that 5th bit to 1 will force regen mode, and the 2nd bit controls how much on a scale of 127->0.

Can you inject a single value into a packet like that, or does it need to replace ALL the data in that chunk each time?

 

The packet is 9 bytes inc the address and chksum, we can change what we like, the chksum will be calculated on the fly by my software. Keep the ideas coming, can you post the updated spreadsheet (link) with some colour coding of your ideas/column titles? Notes in the SS
?

 

Oh! I just realized that the sixth number is speed! Although not in km/hr or anything... 1 seems to be around 4-5 mph?

Look at the log titled "accelerate with assist." I'm going to refer to the columns 1-7 as "C1-C7" for brevity.

You can watch the rowing of gears and speed incrementing perfectly.

First, you're sitting at non-autostop idle (C1 = 64, C4=20), so it's calling for no power from the IMA (C2=0), C2 seems to mean the amount of juice wanted in/out of the IMA. and the IMA is in non-assist mode (C5=33). C7 and C3 are a mystery. Although C3 may be the autostart number.

Then C1 drops to 16 calling for assist at line 54 for a brief burst and you can see C2 start to increment up but stop suddenly... probably a bit of low speed manuevering.

There's a 1 second pause, then you start accelerating in earnest at line 96, you spend about a second in 1st gear (C1=16-17, C2 goes 1->22 as you press the gas.

Then there's about a half second while you're changing into 2nd starting at 124. C1 changes to 36, C2 to 0, C6 seems to keep incrementing upward a bit, but looking at the data, it seems to lag slightly... probably just an artifact of the low-res way speed is measured.

At line 143, you're now accelerating in 2nd, C1=16, C2 starts incrementing up again (for the logic, you might want to have it "ease in" like that, I suspect that's to make it smoother and reduce shearing from instant electric torque), C6 is still increasing. If I'm right and C3 is the autostop number, it goes to 16 at C6=5 on line 163 meaning autstop would no longer be available (maybe this is about 20mi/hr). You're in second for maybe 3 seconds.

Following the same pattern, 3 gear accelleration starts at line 280, and for some reason C3 goes to 32 here? You let off the gas enough at line ~488 for assist to stop, but still in gear, and you can see C2 feathers out over a period of about 1/3 of a second.

You can apply this same pattern to the "Accelerate with full assist" log and it makes sense. You managed to end up in autostop and the IMA restarts it twice before you start taking off, then you can see C6 go up much more quickly, and C2 maxes out at 70. (4 bars seems to be 25).

In regen, C2 is completely different, maybe scale reversing when C5=1? 127 is no regen with maybe 1/3-1/2 second of "ease in" as it falls to the 90s and 80s for coasting gradually working its way back to 110 or so as you lose speed. Where with braking it gets to a maximum (min) of 18. Yes, that fits perfectly.

Sam

 

Hm, don't know enough about your testing, but that C7 could be the engine temp (once it falls below 20 or so, autostop is available?

 

The tetsting was simply driving up and down and airfield after a few miles warm up.

I'll have a look at your data/ideas when I get back home.

 

https://docs.google.com/spreadsheet/ccc?key=0AsCIWnPHFF4KdFdkSks3QVM3eWxtQTNZX1BoSEtMQ0E

There it is with the color formatting... or at least, the two I was talking about in my last post (doing the full sheet causes google docs to barf).

 

Testing time

OK I have drafted some code to try our first Hch1 IMA assist overide experiment. It's quite exciting

Gleaning info from the spreadsheet I have tried toggling four bars of assist when the center button is pressed.

I will need my guinea pigs to shout up when they are ready and I will post it after a bit of code bench testing later this coming week.

To test it you need to have spliced into the MotA Line and brought it out to the 3.5mm stereo jack plug, you also need to have added the 2k2 resistor to the obdiic&c pcb as described earlier and changed one resistor for a 330R variant.

I'll post the test code and pics of my modified board late tuesday with luck.