[WillS]

handy hands

Peter - if you're willing to have me I'd like to participate in the odb11c&c canbus upgrade testing.
To say I'm a bit embarrassed by my lack of understanding of so much of the programming discussion
is an understatement, especeially due to my interest in learning assembly lang.
for decades, yet still haven't buckled down and bulled ahead with it.
If you deem me worthy I'd prefer to purchase an assembled (calibrated with yours ) odb11c&c / canbus upgrade, 2 reduce possible
skewing of data from probable assembly differences.. I'm also up for giving it my best shot at assembling parts
if that fits the process more realistically. I have both, hch2 service manual and electrical manual on site.
I've heard some have the HIM/HDS but I wonder if they all can access IMA ....

I've got some cmdline / bash scripting and a bit of awk / gawk programming skill. I'm also currently looking
for an open ground on the IC board inside my ryobi 18v lithium pack. All 10 cells hold charge nice.
I've got other packs so I'm hashing it out one step at a time.

I'm still learning new tricks at balancing my hch2 packs. Most recently, I believe I've got my hch2 handling
optimal pack voltages. I remember MikeD telling me the first time we met where he would start his
(Tech Mode - Pulsed Discharge ) showing a spot on Labviews discharge Slope. Further, I think I finally have
the hch2 using the voltages that are within the flattest range of the slope, ie: giving the pack a fighting chance,
with the car calling 4 full assist after it's last professional service Nov 2013. Before that service the car would only use half assist according to the dash gauge.
2 years ago. I replaced all the IPU components, piece-meal, until the ECU accepted all the changes.
Than with some pack heating in winter, the pack I just pulled last week lasted for 2 years before it started recaling.
( More testing of this current pack still needed in the car to be sure ), but so far so good.
I also have a complete IPU for benching, but no oscilloscope or adjustable power supply. Only downside is I'm not sure I'll have time to dig into this project the way I'd like to, due to .....

Please give a thumbs up or down to this request and if a thumbs up the cost + shipping of materials.
cheers,
Will

 

[retepsnikrep]

Please e-mail me at my signature address.

 

[cactus]

Here's a link to my post in my other thread regarding successful connection to the CANBUS stick voltages in my HCH2

Post discussing HCH2 battery sticks

 

[retepsnikrep]

Quick HCH2 IMA Control thread resurrection.

I think it would now be possible for me to make man in the middle CAN device for the IMA control line. I did gather a lot of data a few years ago and now with some actual tweaking we might unlock manual control.

This would be two 64mhz 18F26K80 PICS joined together (Master/Slave) using the SPI data bus at 16mhz. Both with CAN boards attached.

The Master would also have a tellymate video chip on board and/or 20x4 LCD and button/joystick. The Slave would be faily dumb just acting like a buffer.

ECM <<<CAN>>> MASTER PIC <<<SPI>>> SLAVE PIC <<<CAN>>> MCM

The first stage would be build the device and see if the CAR will accept it passing thu all messages in both directions.

This work would also potentially benefit G2 Insight/CRZ owners etc.

Now i'm super busy with a million other projects,
so this might get done over the winter after all the UK LTO Group stuff.

 

[retepsnikrep]

Prototype Test.

2 x OBDIIC&C 500k CAN bus connected together using SPI BUS running at 8mhz (8 million bps)

 

[retepsnikrep]

After a tough couple of days and a steep learning curve.

LTO CAN Test board now split into two CAN busses and working with the prototype (MITM) man in the middle device.

CAN BUS 1 Right (Master OBDIIC&C $100 + LTO$201 + LTO$204 + LTO$207 + MITM Master)

CAN BUS 2 Left (Slave OBDIIC&C $150 + LTO$208 + LTO$20B + LTO$20C + Video $125 + MITM Slave)

The MITM Master & Slave are joined using the PIC Hardware SPI bus running at 8mhz.
This should allow us to filter and modify IMA control bytes on the CAN equipped cars like the HCH2, CRZ, and others.

That's the plan anyway..

My SPI bus snakes around a bit on the back of those prototype boards, 10cm or so.
When that gets built properly with just a few mm between the PIC's we should see better stability, less errors etc.

I need to add an SPI Bus checksum routine for now..

The 18F26K80 for the test pcb will see a doubling in PIC speed to 64mhz and SPI bus speed to 16mhz.

 

[retepsnikrep]

Now working correctly

Time to dig out my old HCH2 logic IMA CAN data files...

 

[retepsnikrep]

My feeble Excel skills have failed me again.

Can some open these files and sort them so that the CAN ID 111 data is on the left and the ID 115 on the right?
Not intermixed.... ?????

Thanks

Rename it from pdf to zip then extract.

My next test will be to see if my HCH2 car runs with the CAN MITM device installed.

 

[retepsnikrep]

Standalone minimal MITM unit for in car testing (HCH2 Civic IMA 2006-2011)

This one with the 2 x PIC18F26K80 devices runs at 64mhz with a 16mhz SPI Bus.
Double the speed of the prototype I was testing yesterday.
It's running quite happily on the CAN test bed with no missed packets etc.

The EUSART port is free on the chips and it can output serial debugging data to a Tellymate video chip if needed.

(In fact I could probably make it put the debugging info on the CAN Bus so it's picked up by the CAN Video board!!)

That's a good test i'll try it.

 

[retepsnikrep]

HCH2 Civic Hybrid In car first test.

The vehicle accepted the device shoehorned into the IMACAN bus behind the rear seat and started with the IMA.
It worked for about 30s then fell behind on frames and dropped enough to give a warning and then the IMA light came on.

I was only running it at 32mhz in this test, and the position for inserting it onto the BUS is not the best.
It probably needs to go inside the IPU on the BUS just next to the MCM.

It doesn't need to pass every packet on the BUS to the MCM as there are likely several nodes it doesn't care about.
I need to weed those out and see what else is connected to the BUS. Make a definitive list along with ID's.

At the moment the device only passes standard 11 bit ID 8 byte CAN frames, there may be some 29 bit extended ID frames it is overlooking, so I need to get my CAN analyser on the BUS and check. I can soon reconfigure it to pass everything.

Anyway this was a positive test and shows definite promise.

 

[retepsnikrep]

So I just captured a bit of live HCH2 IMACAN bus data and it's in the attached txt file.

It would be very useful if someone could import that into excel and somehow sort it into ID's with timestamps.

Then we can see how often each ID appears etc..
The screenshot shows there are quite a few ID's but no extended ones.

There are a few < 8 byte frames which I did not account for/remember.
So I need to adjust my MITM software accordingly as this may well have been an issue as it only sends 8 byte frames at the moment!

 

[retepsnikrep]

Just re-reading the PIC Datasheet I completely overlooked the TXB0DLC & RXB0DLC registers. Doh..
Tomorrow's job reprogram MITM to deal with variable length CAN frames.

RXB0DLC = %00000000

bit 7 Unimplemented: Read as ‘0’
bit 6 RXRTR: Receiver Remote Transmission Request
1 = Remote transfer request
0 = No remote transfer request
bit 5 RB1: Reserved bit 1
Reserved by CAN Spec and read as ‘0’.
bit 4 RB0: Reserved bit 0
Reserved by CAN Spec and read as ‘0’.
bit 3-0 DLC<3:0>: Data Length Code bits
1111 = Invalid
1110 = Invalid
1101 = Invalid
1100 = Invalid
1011 = Invalid
1010 = Invalid
1001 = Invalid
1000 = Data length = 8 bytes
0111 = Data length = 7 bytes
0110 = Data length = 6 bytes
0101 = Data length = 5 bytes
0100 = Data length = 4 bytes
0011 = Data length = 3 bytes
0010 = Data length = 2 bytes
0001 = Data length = 1 byte
0000 = Data length = 0 bytes


TXB0DLC = %00000000

bit 7 Unimplemented: Read as ‘0’
bit 6 TXRTR: Transmit Remote Frame Transmission Request bit
1 = Transmitted message will have the TXRTR bit set
0 = Transmitted message will have the TXRTR bit cleared
bit 5-4 Unimplemented: Read as ‘0’
bit 3-0 DLC<3:0>: Data Length Code bits
1111 = Reserved
1110 = Reserved
1101 = Reserved
1100 = Reserved
1011 = Reserved
1010 = Reserved
1001 = Reserved
1000 = Data length = 8 bytes
0111 = Data length = 7 bytes
0110 = Data length = 6 bytes
0101 = Data length = 5 bytes
0100 = Data length = 4 bytes
0011 = Data length = 3 bytes
0010 = Data length = 2 bytes
0001 = Data length = 1 bytes
0000 = Data length = 0 bytes

 

[retepsnikrep]

Sorted out the variable length frames issue and it's happily working but still missing a few packets setting the IMA light.

I need to speed it up and move it to the MCM so I can see what packets I need to pass..

 

[retepsnikrep]

My HCH2 has now been scrapped as it failed the MOT and was in very poor condition.
But this MITM project hasn't been..

I removed all the IMA stuff in the back so I can do some bench testing on ID's etc
I'll also do some internal research on the various modules, they look intriguing.

My next purchase shortly will likely be an early NIMH Honda CRZ which uses a very similar IMA setup.
It is a more sporty vehicle suitable for IMA tweaking mods and CAN research.

I suspect a NIMH CRZ with it's lower ~100V IMA voltage is also ripe for a two pack LTO swap..

This one looks interesting.. UK CRZ

 

[retepsnikrep]

These little $2 gadgets arrived today.

30.82US $ |10PCS MCP2515 CAN Bus Module TJA1050 receiver SPI For 51 MCU ARM controller|Integrated Circuits| - AliExpress

Smarter Shopping, Better Living! Aliexpress.com

www.aliexpress.com

It's basically a CAN line driver and a MCP 2515 CAN module chip on a nice little board with SPI interface to a PIC.

So two of these together and one cheaper PIC might be better than two $$ CAN enabled Pics

Anyway more toys to play with..

 

[*sean*]

These little $2 gadgets arrived today.

30.82US $ |10PCS MCP2515 CAN Bus Module TJA1050 receiver SPI For 51 MCU ARM controller|Integrated Circuits| - AliExpress

Smarter Shopping, Better Living! Aliexpress.com

www.aliexpress.com

It's basically a CAN line driver and a MCP 2515 CAN module chip on a nice little board with SPI interface to a PIC.

So two of these together and one cheaper PIC might be better than two $$ CAN enabled Pics

Anyway more toys to play with..

I use the MCP2515 modules with Arduino Nano boards. Seven wires and you're reading CAN bus. With a Raspberry Pi and seven wires you can read and send CAN bus data, without writing a single line of code. Just attach it to a CAN bus, and run the candump program and pipe it to a file and you can sniff until the SD card fills up.

 

[Techie007]

My HCH2 has now been scrapped as it failed the MOT and was in very poor condition.
But this MITM project hasn't been..

I removed all the IMA stuff in the back so I can do some bench testing on ID's etc
I'll also do some internal research on the various modules, they look intriguing.

My next purchase shortly will likely be an early NIMH Honda CRZ which uses a very similar IMA setup.
It is a more sporty vehicle suitable for IMA tweaking mods and CAN research.

Hey, I've been watching your recent YouTube videos with interest as I have a HCH2 that I've been wanting to upgrade to a high capacity LiFePo4 pack. Towards the end of your January 1 video, you noted that you have several volunteers in the UK for testing, and that it would be useful to get a couple from the US. I am in Ohio and am interested in testing, as long as it's not going to cost too much.
The biggest issue that has kept my upgrade project at bay is the lack of charge/discharge control. For example, if cold, I would need to limit the maximum charge current to the 8-24a range (depending on how many packs are in parallel), and I'd need to abort charging or discharging immediately if any of the individual cells went above 3.6v or below 2.8v, respectively. I have an active balancer that would gladly keep the LiFePo4 pack balanced and report all the cell voltages, which I can easily connect to an ESP32 and write my own code to respond appropriately, but I lack the ability to tell the MCM what to do.
Right now, I see your project can intercept the battery level signals, resulting in the battery SoC meter rising or falling on command, and the PCM adjusting its requests for assist and regen accordingly. Perhaps this would be start that would allow me to instruct the MCM not to charge the battery by claiming that it is full when cold, until it warms up and can be safely charged at full power. Better yet however, it would be nice to be able to directly tell the MCM the level of regen or assist desired.

 

[retepsnikrep]

Unfortunately as I don't have an HCH2 now my focus has moved to the CRZ.
This thread and HCH2 research is unlikely to be updated.

You will have to watch the CRZ threads on the CRZ forum for updates on CAN advances.
Some may well be applicable to the HCH2 but I will have no way of testing that.

 

[*sean*]

I have an active balancer that would gladly keep the LiFePo4 pack balanced and report all the cell voltages, which I can easily connect to an ESP32 and write my own code to respond appropriately, but I lack the ability to tell the MCM what to do.

Unfortunately as I don't have an HCH2 now my focus has moved to the CRZ.
This thread and HCH2 research is unlikely to be updated.

Since you are abandoning the HCH2 work, would you consider contributing the source to the community? This would let folks like @Techie007 run with it. The best way to do this would be to create a repository on github.com and place it there.

 

[retepsnikrep]

I don't have any working HCH2 IMA code to share.
I didn't get far enough..