Tuning the Lithium cars for more IMA power.

10268 Views 101 Replies 10 Participants Last post by PeterPerkins, 2 mo ago

PeterPerkins

Discussion Starter · 6 mo ago (Edited)

Tuning the Lithium cars for more IMA power.

You will have seen my long thread on tuning the Nimh system for more IMA power.

www.crzforum.com

Converting my Nimh CRZ to Lithium (Cells &amp...

I'm a long time member/admin on the Insight Central forum and have worked on lots of the Honda IMA cars developing gadgets and doing Lithium conversions. I've just bought a nice CRZ for myself and intend converting it to Lithium, upping the IMA power output and developing manual IMA control and...

www.crzforum.com

www.crzforum.com

In that I hacked the current and the voltage to double the motor Kw output. It was very lively.

It involved swapping in a Lithium car inverter and dc-dc converter and using the cars original Nimh BCM/MCM module along with an up-rated 200A fuse.

So what can we do with the Lithium cars?

Well we can hack the current in exactly the same was as we did for the Nimh setup.
This will likely get us another 5-7kw or so. Taking us to ~20kw+...

The voltage is more tricky in the Lithium cars especially if we want to keep the Lithium control modules.

The Lithium MCM/BCM gets it's voltage data from three places.

1) The four BMS CAN boards on the pack.
2) A HV input into the BCM. (We can fake this as per Nimh with a very low current HV supply)
3) Serial data from the IPU IGBT module. (We can fake this as per Nimh)

So to tackle 1)

Can we make an interceptor type device that will send out perfect 75% SOC equivalent cell voltages on the CAN bus?
I'm saying 75% as at that voltage equivalent about 3.8V per cell the car will allow full assist and regen.

Yes I think we probably can fool the voltages, but this only gets us to around 190V which is where the OEM Lithium DC-DC and IGBT module max out.

To go higher we might need to use the HCH2 DC-DC which goes to 220V and possibly the HCH2 IGBT module which probably goes to a similar voltage.

That should get us firmly in the 30kw+ camp with a suitable battery pack.

Now to go even higher we need more voltage..

So what IMA cars have higher voltage packs?
Do these dc-dc converters and inverters use the same data stream for control as the CR-Z etc? I bet they are similar.

So would a CRV 2019 or Jazz dc-dc and inverter work in a CR-Z????? Only one way to find out..

But before I start buying expensive parts.........

I'm going to need access to the Honda workshop manuals for these later models to check specs and wiring for compatibility.
Does anyone have current Honda subscription access to this data? Can you message me please.
If people want to support this research in some other way then feel free to message me.

Have a good Christmas and Thanksgiving etc. Peter UK

See less See more

Reactions: 5

41 - 60 of 102 Posts

HondaTuningSuite

· #41 · 5 mo ago

5

Not quite sure what these tables are, but they are outputs of some sort.
X Axis being RPM, and Y axis is unknown, I haven't seen breakpoints like that, so its definitely something specific to the inverter.

Colorfulness Rectangle Font Parallel Pattern

This one is RPM on the Y axis, Unknown X

Colorfulness Rectangle Font Electronic device Electric blue

No idea what this long table is, maybe something with Regeneration.

Colorfulness Rectangle Font Parallel Pattern

Then there are these 2 this could be Assist while the following could be Regen

Colorfulness Rectangle Font Parallel Pattern

Regen maybe?

Colorfulness Rectangle Font Parallel Slope

See less See more

Reactions: 4

da-coder

· #42 · 5 mo ago

HondaTuningSuite said:
Not quite sure what these tables are, but they are outputs of some sort.
X Axis being RPM, and Y axis is unknown, I haven't seen breakpoints like that, so its definitely something specific to the inverter.

Are you using Ghidra for disassembly or something else?
If you want to share the instructions which read this data, i might be able to help there.

Second one might be max regen with X being the battery temps.
This might help:

Font Parallel Rectangle Number Pattern

See less See more

Reactions: 1

HondaTuningSuite

· #43 · 5 mo ago

da-coder said:
Are you using Ghidra for disassembly or something else?
If you want to share the instructions which read this data, i might be able to help there.

Second one might be max regen with X being the battery temps.
This might help:
View attachment 69835

Temps are usually represented as C with a .1 multiplier, I suppose it could be made granular like that but seems odd.

I use ghidra for reversing the roms, I can share a ghidrazip with you, with the table structures setup but its not a complete rom dump until someone wants to give it a whirl, If you have an Openport 2.0 i can send you app to dump a whole bin with, which would be greatly helpful for some of the context of the rom, since its missing the first 64kb of data.

Reactions: 1

da-coder

· #44 · 5 mo ago

HondaTuningSuite said:
If you have an Openport 2.0 i can send you app to dump a whole bin with, which would be greatly helpful for some of the context of the rom, since its missing the first 64kb of data.

Sadly I don't. Peter might though.

HondaTuningSuite

· #45 · 5 mo ago

da-coder said:
Sadly I don't. Peter might though.

v1 through v1.2 ktuner flash is the same cable, and will work

Reactions: 1

PeterPerkins

Discussion Starter · #46 · 5 mo ago

da-coder said:
Is it an open loop or closed loop sensor?

Open loop asfaik. There is no coil around the core..

I will de-solder the chip later.

Reactions: 1

PeterPerkins

Discussion Starter · #47 · 5 mo ago (Edited)

Interesting tables but I know nothing about roms. I can see a steep learning curve ahead.

I suggest the roms dumps etc and disassembly would be better in a separate dedicated thread or this one will get very confusing.

I don't have an openport 2.

PeterPerkins

Discussion Starter · #48 · 5 mo ago

da-coder said:
View attachment 69835

Is this for the HCH2 or the CR-Z (Variant?)

HondaTuningSuite

· #49 · 5 mo ago

PeterPerkins said:
Interesting tables but I know nothing about roms. I can see a steep learning curve ahead.

I suggest the roms dumps etc and disassembly would be better in a separate dedicated thread or this one will get very confusing.

I don't have an openport 2.

Thats a good idea. I will setup a thread for it, since i want to go as deep as i can into all the available systems, even though i don't have a CRZ, Always wanted one

Reactions: 1

PeterPerkins

Discussion Starter · #50 · 5 mo ago

3

HCH2 Phase sensor front/rear with IC (ST?) 0540A1 0622 M56 removed.

Passive circuit component Circuit component Green Hardware programmer Electronic engineering

Passive circuit component Circuit component Green Hardware programmer Electronic engineering

Green Circuit component Hardware programmer Electronic component Electronic engineering

See less See more

Reactions: 1

da-coder

· #51 · 5 mo ago

PeterPerkins said:
Is this for the HCH2 or the CR-Z (Variant?)

Good point, it's for a CR-Z. I'm getting off topic.

da-coder

· #52 · 5 mo ago (Edited)

PeterPerkins said:
HCH2 Phase sensor front/rear with IC (ST?) 0540A1 0622 M56 removed.

So i think i've narrowed it down.

Passive circuit component Circuit component Hardware programmer Microcontroller Electronic engineering

amp 2 operating shunt regulator with red thermistor (R3) and R1, R2, VR1, VR2, C1.
amp 3 being run as differential for the hall effect sensor output.
amp 4 not being used (both inputs are Vcc+).

The resistor and capacitor at the bottom (C2, R8) filter feedback for amp 3.

That leaves amp 1 which I think is changing the hall effect sensor's lower voltage input based on R4.
If not, I don't think it's a quad amp IC. It almost lines up perfectly but not quite.

See less See more

Reactions: 1

PeterPerkins

Discussion Starter · #53 · 5 mo ago (Edited)

@da-coder so how would you modify that circuit to reduce the voltage swing for a given + or - current?

My idea in the example being simulated in Proteus (used in the G1 and HCH1 and CR-Z so far) is along the line of the below.

Rectangle Slope Font Parallel Symmetry

The phase sensor is 0-5V output with 2.5V being 0 amps.
So in the simulation with 1.25V input the output is 1.66V

2.5 - 1.66 = 0.84V change instead of 1.25V

67.2% of the voltage swing is being passed through. 32.8% reduction.
67.2% would have to be increased by 50% to get back to 100% = full output.

R3 is already on the main MCM pcb so that is why I added R2 to avoid any offset.

R1 could be variable as the sensor seem to have quite a good current drive capability.
Making R1 variable with the above values would give you a 0 to +50% current adjustment range.

So each sensor needs four tiny resistors adding.

EDIT To make it simpler to implement I would like to swap out the 10k on the MCM board for a 1k. Then add another 1k pull up to +5V.
This reduces the resistor count and complexity. In effect losing R2 & R3 from the above demo circuit.

Sadly though you can't access the 10k without removing the entire pcb from the IGBT modules. So that's not happening.

See less See more

Reactions: 1

da-coder

· #54 · 5 mo ago

PeterPerkins said:
@da-coder so how would you modify that circuit to reduce the voltage swing for a given + or - current?

Can change the gain on amp 3 by changing R7 and R8.
Transfer function:
Vout = (R7/R5) (V1 - V2) ... R7 = R8, R5 = R6.
So it's a 1:1 change depending on R7.
i.e.:
change output by 2/3...
Vout1 = (2/3) Vout2 ... R7 is the only changing value --> R7.1.
(R7/R5) (V1 - V2) = (2/3) (R7.1/R5) (V1 - V2)
(R7/~~R5) (V1 - V2)~~ = (2/3) (R7.1/~~R5) (V1 - V2)~~
(R7) = (2/3) (R7.1)

I'd suggest just soldering a new resistor in parallel over each.
A 660kΩ (or 680kΩ) on top of R7 and R8 should cause the output to decrease by the same amount as yours.

With desoldering we just need close to 220kΩ in their places.

See less See more

Reactions: 2

PeterPerkins

Discussion Starter · #55 · 5 mo ago

I shall solder in a few resistors and see what happens to the output voltage with my fixed current tests. 10,20,30A etc

PeterPerkins

Discussion Starter · #56 · 5 mo ago

A variable solution would be good. I wonder if some SPI controlled digipots can be attached on a little daughterboard. Or some other sort of switchable stepped solution.

In the earlier models and the crz you can't have more than a certain % current increase at start up as the current into a stopped motor during the start cycle is too high and the mcm sets an igbt short code.

Once the motor is turning the back emf means you can dial up more current increase.

Reactions: 1

PeterPerkins

Discussion Starter · #57 · 5 mo ago (Edited)

Hacking some 270k resistors into the R7 & R8 positions instead of the 330K oem ones gave a current detection reduction of about 20% which equates to a current increase of +25%.

That's probably a good starting point for testing in the car. I have ordered a selection of smd resistors for further testing.

It's possible to do this mod as well with the inverter pcb in position which is great. A steady hand and good eyesight is needed..

Anyway no point getting ahead of myself I need to see if the CR-Z will accept the HCH2 IMA setup in standard form first or this is all academic.

I think I might have just bought a whole car.

Reactions: 2

da-coder

· #58 · 5 mo ago

PeterPerkins said:
A variable solution would be good. I wonder if some SPI controlled digipots can be attached on a little daughterboard. Or some other sort of switchable stepped solution.

Well if we aren't worried about EMF interference we can just make one board and run wires from each digital potentiometer to be in parallel (or replace) those 6 resistors.
Then have one microcontroller run all 6.
Could possibly run shielded cables too.

PeterPerkins said:
Hacking some 270k resistors into the R7 & R8 positions instead of the 330K oem ones gave a current detection reduction of about 20% which equates to a current increase of +25%.
It's possible to do this mod as well with the inverter pcb in position which is great. A steady hand and good eyesight is needed..

Good to hear!

PeterPerkins said:
I think I might have just bought a whole car.

da-coder

· #59 · 5 mo ago (Edited)

If we want to go variable we should go with what you suggested earlier as it's much easier to wire up.
EDIT: removed for simplicity
I'd use an Arduino and an AD5206.

PeterPerkins

Discussion Starter · #60 · 5 mo ago (Edited)

We are def worried about EMI/EMF it's a noise nightmare in the IMA power control zone.
Over 15 years of trying to make gadgets work near the IMA has taught me a lot LOL..
Inside the Inverter sitting on top of the IGBT's is about as near to hell on earth as it gets!

I only talk PIC so would not be using an Arduino in any prototypes.

I'm not sure mine is easier.

Your resistor substitution avoids track cutting and also keeps the layout totally OEM etc.

I think I will try +25% first, if that works then I might consider a variable setup.

I need to investigate the battery current sensor now.

Reactions: 1

41 - 60 of 102 Posts