Honda Insight Forum banner

381 - 400 of 456 Posts

·
Linsight Designer
Joined
·
2,371 Posts
Discussion Starter · #381 ·
Quick update (I'm heading out the door with the wife in a few minutes):
This afternoon I finished up my first attempt at firmware MVP... I've run out of time to test it, but will do so when I get back from social function in a few hours.
 

·
Premium Member
Joined
·
67 Posts
It's just a splash page for now, but:
Nice, and a web page means it's real!!!! Oh wait, the Elio I put a deposit on 7 years ago has a web page......
 

·
Registered
Joined
·
194 Posts
if Momma ain't happy, none of us get to be...

Have fun, enjoy the time off!
 

·
Linsight Designer
Joined
·
2,371 Posts
Discussion Starter · #384 ·
Alright, got LiBCM installed in the car... it fired right up!

...I backed out of the driveway, then gunned it... immediately got P1586(23).
That means LiBCM's current sensor value doesn't match the MCM's current sensor value. This isn't an issue with LiBCM's analog measurement... it's just something wrong in firmware.

I'm a bit worried that the Arduino serial connection halted immediately afterwards... that's going to be a fun one to figure out. Hopefully it's not an RF noise issue.

It's all part of the experience. I'm done testing on the road tonight... it's been raining off and on all day and I'm not wanting to get stranded right now; better to do that during the day.
 

·
Linsight Designer
Joined
·
2,371 Posts
Discussion Starter · #385 · (Edited)
Shifting gears a bit, the new connectors came in while I was out!

They both fit correctly! Thanks @Bull Dog for finding the manufacturer and series for these connectors!

Note that I installed the batteries backwards in the video... points I made are still valid, but I'm not going to re-shoot it.

Word to the wise: As I've mentioned a couple times now, LiBCM is still in the prototyping stage, so if you're playing along at home (e.g. making your own cables), keep in mind that the drawings I've provided are not yet a released product. It's possible that I'll end up rotating the lithium modules 180 degrees (bottom becomes top) to make it easier to plug them into the RevB PCB. If I end up doing that, then the cable lengths will change. So please wait for the final release prior to cutting up that expensive 4 AWG cable!
 

·
Linsight Designer
Joined
·
2,371 Posts
Discussion Starter · #386 · (Edited)
Oh wait, the Elio I put a deposit on 7 years ago has a web page......
Same, but for the Aptera. Somehow I put another deposit down for one... probably because it was only $100 this time.

This is also why I never took deposits on Linsight (now 'LiMCM') several years back... I don't want to be beholden to anyone on these hobby projects. This is also why I haven't taken deposits on LiBCM. I could always get hit by a truck tomorrow.
 

·
Linsight Designer
Joined
·
2,371 Posts
Discussion Starter · #391 ·
Is that code checked into the repo?
Yes, and it's bare-bones and terrible. I'm just going for the MVP, which right now is a superloop. Once I get the barebones working, I'll refactor the code into a cooperative multitasking arrangement... but for now I'm just verifying basic functionality.

main() is in MVP.ino.
 

·
Premium Member
Southern California
Joined
·
865 Posts
Alright, got LiBCM installed in the car... it fired right up!
Love seeing Pegasus on the dashboard. :) Can't wait to finish that up! And congrats on the partial success! Very good for a first run.
How are you getting data through OBDIIC&C and Pegasus at the same time?

Also, for the ribbon cable connections at the PCB, I'd really love if you could find room to add a couple holes on either side of each connector. I'd want to 3D-print some brackets that would then be screwed to the PCB and keep the cable from vibrating out. Or do those connectors have some kind of retention?
I definitely want to make sure everything is positively retained. I could probably also drill a couple holes in the acrylic cover and attach a part there to press on them.

Are you going to make a cable test jig to verify all the balance cables were made correctly?

Also, change the termination resistor on the METSCI line from 120Ω to 100Ω. That's the value on the instrument cluster's board, and I assume on the stock BCM as well. Not sure why Honda did that, typically the standard is 120Ω. Maybe 100Ω impedance twisted pair was cheaper/easier to source.
 

·
Administrator
Joined
·
12,892 Posts
I found this chart of OCV versus SOC for LiPo, not sure if any use.

Code:
Table 4.1 Look-up Table for Cell SoC Estimation Based On Cell OCV
Cell OCV(V) SoC
[4.2-4.1) 100%
[4.1-4.05) 90%
[4.05-3.99) 85%
[3.99-3.945) 80%
[3.945-3.983) 75%
[3.983-3.847) 70%
[3.847-3.812) 65%
[3.812-3.7755) 60%
[3.7755-3.7405) 55%
[3.7405-3.670) 50%
[3.670-3.658) 45%
[3.658-3.618) 40%
[3.618-3.565) 35%
[3.565-3.505) 30%
[3.505-3.430) 25%
[3.430-3.350) 20%
[3.350-3.240) 15%
[3.240-3.140) 10%
[3.140-3) 5%
[3-0] 0
 

·
Linsight Designer
Joined
·
2,371 Posts
Discussion Starter · #394 ·
Sanity check: Bam (IMA Battery Amps) should be negative when providing assist, right? I should know this, but I'm too far in the weeds.
 

·
Linsight Designer
Joined
·
2,371 Posts
Discussion Starter · #395 ·
Love seeing Pegasus on the dashboard. :) Can't wait to finish that up! And congrats on the partial success! Very good for a first run.
How are you getting data through OBDIIC&C and Pegasus at the same time?
Arduino Mega configured as a Man-in-the-middle bus arbiter. MEGA queues requests from each device, interrogates ECM, then ping pongs the results to whomever asked for it... if I can't get new data in time (due to bus conflict), I just send the last received value for that data type back to the requestor... this seemed to be necessary for OBDIIC&C to not barf; Pegasus seems to work more gracefully. It's not pretty, but I got tired of not being able to use both... particularly clearing the numerous CEL codes I get from all the tinkering. Obviously not the intended use for either device, but I got tired of not being able to use both.

Also, for the ribbon cable connections at the PCB, I'd really love if you could find room to add a couple holes on either side of each connector. I'd want to 3D-print some brackets that would then be screwed to the PCB and keep the cable from vibrating out. Or do those connectors have some kind of retention?
I definitely want to make sure everything is positively retained. I could probably also drill a couple holes in the acrylic cover and attach a part there to press on them.
The area around those connectors is too dense, and installing screws after plugging in is too dangerous. Further, these screws would need to be touch-safe, which is not possible given the 3 mm creepage I'm using on the board. If I end up needing to spin to RevC, maybe we can revisit... but those connector traces are pretty compact, given that they're wedged between the fans and narrow PCB sections.

I do not believe vibration is an issue for any connector on LiBCM. For certain there's zero change the Mega would ever vibrate off... I practically have to pry it off when I want it off... pin-mates add up. I've shipped maybe 20,000 products used in high vibration applications and have never had a daughterboard fall off.

Note that if you want positive retention, it won't be from the acrylic cover... the battery connectors attach on the other side. On the other side, you could place a foam pad between the lithium modules and the PCB headers. A better option would be to place glue on the PCB connector after plugging in the harness. I do not believe this is worth the effort.

Are you going to make a cable test jig to verify all the balance cables were made correctly?
Yes, using NI hardware. I'm also going to test each cable I make with the LiBCM I ship it with... during FVT.

Also, change the termination resistor on the METSCI line from 120Ω to 100Ω. That's the value on the instrument cluster's board, and I assume on the stock BCM as well. Not sure why Honda did that, typically the standard is 120Ω. Maybe 100Ω impedance twisted pair was cheaper/easier to source.
You are absolutely correct...
...but I didn't see any reflections on METSCI during my hardware testing. I'd love to put a 100 Ohm resistor down to meet the design intent, but I'm a reel minimalist and I needed a 120 Ohm already... and didn't want to add both. Again, the signals are very clean on both sender (MCM) and both receivers (instrument cluster & BCM). METSCI is such a low speed signal - and the LTC1487 is such a beast with RS485 - that the 20 Ohm difference doesn't matter in this application.

Please elaborate if you foresee any problems... I'm not seeing anything at the moment.
 

·
Administrator
Joined
·
12,892 Posts
Bit 15 of your amps word in the 87 & AA packets should be 1 when assisting and 0 when regening IIRC.

My old integer only maths BCM gauge AMPS code from 15+ years ago..

Code:
'Amps = ((BCM87[4] * 128 ) + BCM87[5]) - 2048    'Calculate Battery Current

Result of above might go negative setting bit 15.. 

'if Amps.15 = 1 then    'Calculate AmpSign
'AmpSign = 45    'Set AmpSign to 45 (-)
'else
'AmpSign = 43    'Set AmpSign to 43 (+)
'endif

'Amps = ABS Amps
'Amps = Amps * 2500
'Amps = Amps /  512
 

·
Linsight Designer
Joined
·
2,371 Posts
Discussion Starter · #397 ·
OK, then I had my sign backwards... makes sense that the MCM wouldn't like the BCM telling it "+100A" when it was reading "-100A". I'll have another test drive in the morning... see if we can drive to the top of Lookout Mountain in 3rd gear without breaking anything... that's constant assist for several minutes.

Unrelated, but annoying: I can't update LiBCM's firmware when the key is on, or else the MCM gets all pissy and throws the contactor open, then throws several P codes.
 

·
Administrator
Joined
·
12,892 Posts
Unrelated, but annoying: I can't update LiBCM's firmware when the key is on, or else the MCM gets all pissy and throws the contactor open, then throws several P codes.
YEP its a PITA. :(

I have always had that firmware upgrade problem with stuff that is communicating with the MCM when the car is on.
When you update and your device stops working (sending data) for a few seconds the MCM throws a whole bunch of codes.

I try to arrange my gadgets so the PIC/CPU can be independently powered by the cheapo programmer and updated with the car off. I usually use a simple SIL 0.1" jumper to isolate the cpu +5V feed from the rest of the PCB/MCM/BCM stuff etc etc The pic programmer then just provides +5V to the PIC and not the rest of the board.

Your's is a bit more tricky/complicated. :unsure:
 

·
Linsight Designer
Joined
·
2,371 Posts
Discussion Starter · #399 · (Edited)
Since LiBCM doesn't turn off when the key is off (unless the IMA battery is low*), users can just update the firmware with the key off (via USB).

*Note that even if LiBCM has turned itself off (to save power), if you apply power to the USB port (e.g. plug it into a computer), then LiBCM will ALWAYS turn back on (even if the IMA battery switch is turned off). Therefore, you can always update the firmware by plugging the USB port into a computer and uploading the new firmware.

...

Here's a video I took prior to reading your comment, that happens to address this indirectly:


I'm excited to go for another test drive tomorrow.
Hopefully whatever was causing the USB connection to fail was a fluke... cause otherwise there's gonna be some headachy RF exploration... maybe there's a reason the BCM is inside a metal box ;).
 

·
Registered
Joined
·
1,036 Posts
For certain there's zero change the Mega would ever vibrate off...
Probably not, but there are these (they came with headers for a DIY shield but I have not tried fitting them. So as length goes, ymmv)
90851

I'll admit they are a PITA to work with when they don't line up. That might make them worse than not using them.
 
381 - 400 of 456 Posts
Top