Honda Insight Forum banner

161 - 180 of 219 Posts

·
Premium Member
Joined
·
1,850 Posts
The kit will include a fully assembled PCB... these parts are too small for the average solder-gun wielder to handle... I'm using a pick & place machine and a five zone reflow oven (that i just happen to own as part of a legitimate business venture).

These kits will include all parts needed to complete the conversion, except for the actual lithium packs. You'll start with an OEM battery pack (NiMH cells), and end up with a functional lithium pack.
I apologise for the ignorant question, but I really don't understand the inner workings of these batteries. What is the AH rating with the 48s configuration?
 

·
Premium Member
Joined
·
1,850 Posts
Ah ty. That's perfect then.
 

·
Moderator
Joined
·
7,223 Posts
Also to answer Balto's cost question - I'm not 100% sure yet. But it will be more than Battery Hookup's price... :oops: That was a crazy deal.
 

·
Linsight Designer
Joined
·
2,172 Posts
Discussion Starter #165
Hey gang, I haven't worked on this project in a few days now... but will get back to it later this week. Instead, I've been working on getting a retaining wall permit, dealing with a cat that got hit & run, doing consulting work, etc. Tomorrow I've got a new CNC machine coming in, but I should get back to this project Thursday.
 

·
Registered
Joined
·
504 Posts
It is not. Silver lining is she was killed instantly. Unfortunately I have the whole act on video.
Sorry for your loss. Losing a pet is the toughest thing I've dealt with in my life so far (I guess I'm fairly inexperienced/sheltered).

Just thought I'd check in to see if any of the work done here would helpful/reusable. This project uses the LTC6811/LTC6812/LTC6813:
EnnoidMe/ENNOID-BMS-Firmware: Firmware for ENNOID-BMS (github.com)
ENNOID - BMS GEN 1

I haven't worked with it personally, but this is the BMS my FSAE team is using this year.
 

·
Linsight Designer
Joined
·
2,172 Posts
Discussion Starter #170 (Edited)
@GuySmily Thanks for the condolences.

I'll take a look at that design tomorrow. One issue I have with LTC6811/12/13 is those devices operate in series only... whereas the LTC6804-2 operates in parallel (i.e. each IC is individually addressable). Series operation would require different firmware for each series cell quantity (e.g. onboard 42/48 or external 60/72). Series operation also might not work with the 8b MCU I'm planning to use... because I have to store the entire shifted result simultaneously in RAM (versus just processing each ICs' QTY12 values at a time). In industry series operation has certainly won, primarily because it lowers the isolation voltage required between each cell set in the stack. However, in industry the number of series cells is fixed, hence only one firmware version is required.

However, I'll definitely look at the firmware implementation in the linked repository tomorrow... I know I've chosen a weak MCU, but I'm quite familiar with it and I'm certain it's up to the task. However, if someone else has already done the heavy lifting with a more powerful MCU, I'll definitely consider it.

...

Daily update:
With some slight modifications I'm able to fit everything onto the PCB without the odd Bay2 BMS connector shenanigans I mentioned in my previous video... everything fits comfortably now (the orange highlighter no longer means anything; green is now HV isolation):
89760


I've spent a few hours roughly throwing things into a schematic. The linked file IS ABSOLUTELY NOT FINISHED, but some people have expressed interest in seeing it. There's not much there right now, except for the actual BMS circuitry on page 2 (using the LTC6804-2).
 

Attachments

·
Administrator
Joined
·
12,725 Posts
On the module BMS 12/18S connectors what did you decide to do to interface with the plugs in the light of the unobtanium male sockets?

IIRC you mentioned pins on the pcb arranged as per BMS plugs..
Do you have any drawings or specs for that?
 

·
Linsight Designer
Joined
·
2,172 Posts
Discussion Starter #173 (Edited)
Yeah it's just un-shrouded 3mm spaced pins. It's not an ideal solution, but it works.

The 12S connector has two rows spaced 3 mm apart C2C. Each of ten columns is 2.2 mm spaced C2C.

The twelve interior columns and rows on the 18S connector are identical to the above. However, both outside columns are spaced 3.5 mm C2C from their neighbor, and these outside pins are spaced 4mm C2C from each other. I haven't made a final drawing yet, but when I do I'll post it here.
 

·
Administrator
Joined
·
12,725 Posts
It def might be useful to have some little breakout pcb boards for those 12 & 18S connectors based on your measurements.

Small pcbs with pins straight to a line of SIL 0.1" pads/holes so we can solder wires or use screw terminal connectors if we want.

I'll be using an Orion2/CR-Z BMS as well as your BMS pcb with a couple of separate builds.

89906


You get the idea..

Someone else on here might be able to knock them up from your drawings. :)
 

·
Linsight Designer
Joined
·
2,172 Posts
Discussion Starter #175
That's a great idea. I'd use 180 degree screw terminals (green in your picture), so that all the wires point the same direction.

Daily Update:
I should finish up the schematic tomorrow. It's been going slower than I wanted, but it's getting done. Here's the (still incomplete) schematic as it stands now:
 

Attachments

·
Registered
Joined
·
971 Posts
Fun to see a real schematic!!! This is not a design review or intended to be design input; it's just some thoughts.

thoughts

  • nice to leave a means to extend the isoSPI bus to outboard modules.
  • thanks for the tip on the HV DC - LV DC converter!
  • interesting circuit providing HV to the BCM - is that PWM from the Arduino, filtered by C7? I may borrow that.
  • ditto for the VPIN drive from the Arduino - PWM filtered by C7 (you have C7 in two places)
  • i did not know HLINE was 5volts. I would have assumed 12V like the OBD2 line. learned more than one thing new today.
  • SPI/isoSPI appearing on BCM connector B. Interesting!
  • neat: variable-current grid charging

questions

  • reason for two 5M resistors instead of one 10M on the transformer common? in case one shorts/arcs short?
  • 22 ohm load for balancing. The Fit EV LTOs used 47 ohm with a lower cell voltage. I think that's more than a half watt dissipated at 3.7V. Am I wrong?
  • does this literally use an Arduino Mega board plugged into the main board? (do you consume all the serial ports, or is one left over? I use a serial Bluetooth module in my BCM to spit tap voltage data to my phone.)
  • shorting TX0 & RX0 - no USB reprogramming of the Arduino Mega without unplugging it from the board?
  • would one-wire temperature sensors be a non-starter in this noisy environment? I was thinking of using quite a few of them in my eventual LTO build.
  • i think you noted that the packs had an outer cage that could short if crushed. Is that wired to the pack's connector and if so, are you using it to detect a crush event and shut things down?
  • does inclusion of the grid charger invoke a requirement to perform Part 15 FCC certification testing (device with oscillators powered off the mains)?
 

·
Linsight Designer
Joined
·
2,172 Posts
Discussion Starter #177 (Edited)
Fun to see a real schematic!!! This is not a design review or intended to be design input; it's just some thoughts.
Thanks for the feedback. This is exactly what I'm looking for when I post things!
Upon looking at the schematic this morning, I see that the net names aren't visible... this is because I work with the schematic in 'dark mode', but forgot to invert things for the pdf. I'll make sure the net names are visible in the final schematic.

  • nice to leave a means to extend the isoSPI bus to outboard modules.
I hesitate to include it as a breakout, because I'm worried that people won't realize that they're working with high voltage on the breakout boards. Even though isoSPI is an isolated bus, the standard only relies on isolation transformers. I've added series-coupled capacitors to further increase safety, but I do worry that someone might not know what they're doing and connect an unisolated pack to it. Rest assured there will (eventually) be an exhaustive discussion on how to safely interface with isoSPI.

FYI: Serial-configured isoSPI devices certainly would be easier to use (e.g. 681x parts, 6804-1, etc)... but they're unobtanium right now due to the 2020/2021 chip shortage.

  • thanks for the tip on the HV DC - LV DC converter!
Yeah I was going to design my own, but this part is only $6 and is a finished solution. It requires no external feedback. Better yet, its idle current is almost nonexistent... on the bench I measured 120 uA at 150 Vin (18 mW), although the datasheet only guarantees it's less than 70 mW. Put another way, the 42S battery configuration will run this converter (at idle) for about 15 months.

FYI: I purchased all of DigiKey's EPM1210SJ inventory. The great parts shortage of 2020/2021 is certainly changing how I procure parts: Before choosing any specific part, I make sure to buy it first ;). At work we've resorted to buying existing parts on the gray market (e.g. aliexpress, etc)... they're just unobtanium in traditional channels.

Related: My goal is that the entire system will run for 6 months without starting the car. I've spent time ensuring that those systems that aren't required when the car is off can be turned off (e.g. with open drain FETs, etc).

  • interesting circuit providing HV to the BCM - is that PWM from the Arduino, filtered by C7? I may borrow that.
I'm not sure which circuit you're talking about. This will be easier to discuss once I add reference designators ;).

  • ditto for the VPIN drive from the Arduino - PWM filtered by C7 (you have C7 in two places)
Sorry, I haven't updated reference designators in a while... FYI: these are actually "C?" because they're presently unassigned. I'm not 100% certainly how quickly the MCM needs to see an update, so I might need to adjust the first order lowpass filter values.

  • i did not know HLINE was 5volts. I would have assumed 12V like the OBD2 line. learned more than one thing new today.
Yeah it's 5v... but I had to double check that just now ;). I'll add some protection circuitry to it on the final schematic.
Honda didn't make this signal robust because it's a diagnostic-only tool.

  • SPI/isoSPI appearing on BCM connector B. Interesting!
I thought I figured this neat hack out on my own... but it turns out Peter has already stuffed pins into the (relatively empty) OEM connector in his previous designs. It's a clean solution.

  • neat: variable-current grid charging
Yeah I want to be able to control just about everything on the grid charger via the microcontroller. This same circuit will work with the Lambda GEN high current chargers (e.g. a 200 volt 16.5 amp unit). However, I don't intend to officially support anything beyond the specified part number.

  • reason for two 5M resistors instead of one 10M on the transformer common? in case one shorts/arcs short?
Safety, creepage and clearance, etc. I don't skimp when it comes to safety isolation. I've even got a gas discharge tube on the HVDC inputs. I worked on a modular (PXI) 1000 volt multimeter many years back now... it was amazing to see how much damage our customers managed to do to them, without hurting themselves or bricking the rest of their hardware.

  • 22 ohm load for balancing. The Fit EV LTOs used 47 ohm with a lower cell voltage. I think that's more than a half watt dissipated at 3.7V. Am I wrong?
The 2nd page of the schematic is lifted from the original Linsight design. I haven't updated some of the component values yet (see note at top of page). My original design goal was to be able to restore a 5% imbalance in 18 hours. With a 66 Ah cell (e.g. Nissan Leaf in the Linsight project), that required ~20 Ohm resistors.

For LiBCM the 5 Ah cell can be discharged at much lower current. I haven't quite settled on the final resistor values. Per the note at the top of page 2, right now I'm considering 33 Ohm resistors (500 mW @ 4.1 V), but I haven't settled on a final value. These resistors are rated to 1.25 watts, and are cooled via fan whenever they're active. Assuming I want to restore a 5% imbalance in 6 hours, that would only require 100 Ohm discharge resistors, which would only consume 170 mW. I'll probably strike a balance somewhere in between.

  • does this literally use an Arduino Mega board plugged into the main board? (do you consume all the serial ports, or is one left over? I use a serial Bluetooth module in my BCM to spit tap voltage data to my phone.)
Yes, using an Arduino Mega, although the firmware will be written in C.
For the low cost of Uno/Mega boards, I almost never design them out, unless space is an issue. It's also really easy to troubleshoot when things aren't working as desired.

Send me your serial bluetooth schematic and I'll put it on the board (if it fits). I don't intend to initially support it, but it won't hurt to have it there... I assume you have an external antenna (since the IMA enclosure is a good Faraday cage)?

  • shorting TX0 & RX0 - no USB reprogramming of the Arduino Mega without unplugging it from the board?
Good catch... I was getting tired last night, but might not have caught that today. I'll move to UART1 when I connect BATTSCI/METSCI today.

  • would one-wire temperature sensors be a non-starter in this noisy environment? I was thinking of using quite a few of them in my eventual LTO build.
Right now there are:
-QTY4 RTDs on the PCB proper (to measure balancing resistor temperatures).
-QTY4 RTDs reused from the OEM battery pack. I suspect they'll be used to measure cabin, inlet, outlet, and grid charger temperatures.

Whatever leftover pins I have, I plan to add some unassigned open drains, GPIO, filtered, and maybe even a high side driver (TBD), via screw terminal header... probably QTY8 total unassigned pins.

  • i think you noted that the packs had an outer cage that could short if crushed. Is that wired to the pack's connector and if so, are you using it to detect a crush event and shut things down?
This didn't end up being the case... the insulation between the cells and the outer enclosure definitely exists, but it's not detectable (at least there's no wiring for it). Unfortunately, I can't "shut things down" because only the MCM can do that (e.g. open the contactor). I guess technically I could throw a crowbar onto the fused 12 volt rail (thus blowing the fuse). But no, I don't have any plans to do any of that.

  • does inclusion of the grid charger invoke a requirement to perform Part 15 FCC certification testing (device with oscillators powered off the mains)?
The off-line converter is already FCC approved. I am very well aware that this approval does not cover my derivative design, so I'll self-certify with a declaration of conformity (otherwise known as an "honor system approval"). I simply don't have the equipment, time, or energy to test this device for RF emissions. Note that I've ship many products containing off-the-shelf off-line converters, and haven't tested any of them since I left National Instruments in 2013 (NI has its own stringent internal compliance lab).

Given that this design will be used within a nearly sealed aluminum enclosure, any Hams out there will be glad to hear that I do not expect any RF emissions issues.

But yes, Part 15 FCC technically applies. If I add a Bluetooth module it gets even more murky. Note that if anyone comes knocking, I'll cite 15.103(a), which exempts "digital devices utilized exclusively in any transportation vehicle".
 

·
Registered
Joined
·
971 Posts
1. The Bluetooth modules I use are the HC-05 and HM-18. There are others in between. Basically, +3.3 or +5, TTL serial TX & RX. A couple of other pins are not used if you plug in a module you've programmed elsewhere. https://components101.com/wireless/hc-05-bluetooth-module; Amazon.com: DSD TECH HM-18 CC2640R2F Bluetooth 5.0 BLE Module Compatible with HM-10 for Arduino: Home Audio & Theater. The four pins that count are the center four; some modules only include those four pins. Looks like uart0 tx/rx (USB) could double for this. But yeah, I would not mess with integrating Bluetooth or even offering it as a DIY add-on unless you're prepared to drop a whole lot of money for certification. I build the things that use these for myself and only have built a few devices that use them.
2. There's a gap at the back of the IMA compartment between the driver and passenger seats where the rf leaks out if you put the bluetooth module close to it. I have not tried the module elsewhere inside the compartment.
3. I liked the color schematic
4. I see you added the 120 resistor to the hline. I meant to mention that but wasn't sure it was needed (have done this for homebrew obd2 and kwp2000 interfaces using Arduino)
5. I like how you are using an optocoupler/driver to scale down the pack voltage to what the MCM needs to see. I will be curious to see the firmware for voltage management. seeing the batsci and metsci protocols documented in a repo once and for all will be awesome. I believe that if the MCM is provided a lower voltage than the pack is actually at, the effect will be a power "boost".
6. The downside with using rewritable firmware is the ability for someone tinkering with it to burn down their car / garage / house / passengers / bystanders / put-your-imagined-world-ending-scenario-here (and blame someone else).
7. (here i digress) I am interested if you have examined the HCH2 and HCH3 reluctor-wheel (?) method of motor phase/position sensing, and whether that can be be adapted back to the hall effect sensor (?) position output from the three Insight motor sensors that the MCM expects. This would allow the higher power HCH2 motor to potentially be retrofitted. The HCH3 motor has 33% more windings and so this signal is potentially 33% faster and it is unknown if the tach signal going into the MCM is compared to this signal from the motor. The HCH3 motor has 2x the power and torque capability and looks like it has a similar bolt pattern, though the case may be wider than the Insight motor and may cause issues with motor mounts, driveshaft lengths, etc.

The concern I have with overdriving the Insight motor with a current mod + a higher pack voltage than what the MCM is told is that this really has not been tested in a high stress environment like found at altitude in the US Southwest. Thus one potential future input will be temperature sensors in the motor to back down assist/regen if a boosted motor becomes too hot. On the other hand the more powerful motors may be more fun but the power to recharge them has to come from somewhere. The regeneration mechanism (brake light = turn on regen) may be too simple if the highest regen is to be extracted out of the motor - it may simply brake the car too much for just touching the brake pedal.
 

·
Administrator
Joined
·
12,725 Posts
The concern I have with overdriving the Insight motor with a current mod + a higher pack voltage than what the MCM is told is that this really has not been tested in a high stress environment like found at altitude in the US Southwest. Thus one potential future input will be temperature sensors in the motor to back down assist/regen if a boosted motor becomes too hot.
I have overdriven the G1 motor upto 30kw at 240V for short periods without apparent ill effect.
I have also discharged 40ah 160V packs through it over relatively short periods. ~1hr

It probably gets hot, but I suspect a pack of these new 4.5ah cells just doesn't contain enough energy to seriously overheat it.. Even forced cycling up and down would likely overheat the batteries or MDM as quickly as the motor.

I did add temperature sensors to an IMA motor over ten years ago but never felt the need to take readings.
As the gearbox etc has to come out to do that it's not going to happen for 99% of people...

It might be possible to extrapolate from that research paper testing of an IMA motor for ballpark figures.
Spare IMA Motors are ten a penny as they don't break in normal (or even overdriven use) so if we overheat one it's no big deal.

You could extrapolate the MDM and/or battery temperature as an indicator of likely motor temperature/duty to add limits, or count current versus time for average power etc etc without having to actually physically measure the motor temperature directly.

If the MDM/battery is working hard, so is the motor... ;)

Bottom line IMHO it's likely not a problem..

On the other hand the more powerful motors may be more fun but the power to recharge them has to come from somewhere. The regeneration mechanism (brake light = turn on regen) may be too simple if the highest regen is to be extracted out of the motor - it may simply brake the car too much for just touching the brake pedal.
I never noticed/considered braking regen (dabbing the brake pedal) as being too powerful but we are all different.. :unsure:
 
161 - 180 of 219 Posts
Top