2nd gen Honda Insight parallel portable plugin hybrid battery pack build thread

[RedDevil]

Back from that weekend out, I installed the car harness in the car and tested it.

Contact on, no batteries attached, flip the switch to see the IMA voltage.
113 Volt. Ah, nice.
Then it starts to fluctuate wildly. Then it says -1. What the?
PING says the dash. I flip the switch off and look at the dash. CHECK IMA SYSTEM.

Oh boy.

I reset the code, try the same with the engine running. But the voltmeter only says 1 or -1.
I know what that means: overload detected.
Now there ain't more than the 200 Volt the meter can show, but maybe its internals got messed up? The box is pretty crowded after all.

Whatever the cause, this won't work.
I remove the harness again, reset the code and go for a test ride.

The IMA system works fine. I did not ruin it, hooray!
But shortly before arriving home another warning pops up: CHECK OIL LEVEL
Err... oil? It did not use a drop previously? This ain't funny anymore!

Home. Check the oil and the CVT fluid level. Both spot on.
So it sends me false warnings? Did I damage the ECU or what?

I get in to check for this on my computer.
My wife says Now you know it's not the oil, why don't you reset that code too and do another test drive to see if it comes back on?
Yeah, let's do that first. Shouldn't hurt.

I key on expecting to see the warning light up to reset it with my UltraGauge. But it is gone by itself.
Test drove another couple of miles, lots of IMA support and regen, good throttle response, no codes whatsoever.
I thank my wife for making me do that next run Finally some good news.

Okay then. I will take the harness apart tomorrow evening to find out what's wrong with it.
I'll sturdy up the volt meter, pair the relay coils with condensers and tidy up the control box interns. Then test again on the bench and hopefully in the car.
Maybe I should not have tested it without the batteries connected.
OTOH, if it shorted then it would have caused a disaster as the batteries, though protected by their BMS, may still yield 55 Ampere.

The harness should be safe to operate without batteries though.
It would defeat its purpose if it wasn't.
Besides, the volt meter could only have caused a minimal draw if working properly, so what caused the IMA fault code?

Tomorrow brings another challenge, so to say

[RedDevil]

The meter failure and IMA warning were probably caused by feeding the internal volt meter with 12V from the car.
I bet the IMA system does not like to be connected to the 12V system in any way, like if the meter has a common ground or something.
That did not reveal in my bench tests (where I used a separate 12V source).

I will check on the volt meter this evening.
If it still functions I'll hook it up with a separate 12V source. I have a bunch of tiny 12V batteries (half the size of an AAA pencil battery) and will use that plus a switch as the source. And I'll put some 2K2 Ohm resistors in the test leads just to be sure.
As often in my life, wisdom comes from messing things up first.

Oh, and the car behaved fine this morning. In fact better than usual, partly due to me booting it a little in my second test drive yesterday evening, so the IMA battery must have held a higher state of charge than usual.
That effect of a higher SOC motivates me even more to get the pack sorted out and operational.

[Baltothewolf]

I can't wait until you get this done and working. I'm extremely curious what MPG gains are going to be in store for you :P.

Also, would this design be tweakable to work in the G1? (Too lazy to read to see if anyone has asked already, just a simple 'yes' or 'no' would be sufficient for me).

[RedDevil]

The G1 pack has a higher voltage and yours is a manual, which gives you more control over charge and discharge of the pack - which may work for the good, but also allows you to damage it!

I know of no G1 parallel pack built like mine, I would have devoured such a thread if I knew of it.

I expect that I'll use about 1 kWh per commute and save half a liter of fuel with it.
At current electricity and fuel costs I need almost 3000 of those to break even.

I theorized that feeding the system makes it a better hybrid; theoretically it could achieve better economy with augmented hybrid use than it would if I use EV mode until it peters out (if that were possible whit the Insight) and then resume normal hybrid mode with the charge/discharge hysteresis losses. In that case the gain will be bigger.

It will definitely be more fun to drive. The way the Insight got off the line this morning with its fuller than usual IMA battery was encouraging.

[RedDevil]

I checked the voltmeter and lo and behold, it had a common ground. So effectively I was connecting the negative terminal of the IMA system to ground.
It did not like that.
Neither did the voltmeter which now only displays a strange reading without much reference to the input, even when fully isolated.

I ripped the meter out and tested all connections for further shorts. None of that, everything should be fine.

With the harness reinstalled, ex voltmeter, it was time to test the waters again.

First, test the harness without batteries again. No problem whatsoever.
Hooked the Fluke up to the current bearing connector terminals and they show 107.5 Volt; about average.

Switch off, hook up the batteries. I can now test the system voltage through the charger connectors on the batteries. 126.0 Volt, the batteries hold 42 Volt each, being charged to capacity.

The moment of truth. Flip the switch!
The voltage races down to 109.5 Volt and stays there for a second or ten. Then I hear the dreaded Ping on the dash: Check IMA system!
The voltage is back up to about 122 Volt. The IMA system disconnected the terminals and the battery shows what its resting voltage is at 99.9% SOC.

Switch off. Contact off and on again. Reset the IMA codes.

What if the engine were running?
I start up, walk to the back, flip that switch again and see the voltage race back down to 109 Volt. And it stays there, no codes.
Success!

Time to put it on the road for a few laps around the block.
It runs fine, but as the engine is still cold there is not much happening IMA wise. The blue light goes off, still nothing much.
Fuel consumption is normal for running around the block with a lukewarm engine. The parallel pack seems to have little effect. When I decelerate it recharges mildly. When I accelerate it uses some boost but not much.

Then all of a sudden it jumps into full EV mode while accelerating mildly.
Ha, at last some effect from the pack!
PING. Check IMA system. Dang.



The engine revs up to 1700 RPM and will go no lower.
I pull up my driveway and do the ritual:
Switch off. Contact off and on again. Reset the IMA codes. Check the voltage. Start again. Switch the pack on. Get moving again.

PING Check oil level.
Oh Honda you funny *******, you know the oil is good. This is just bullying.
Key off-on, warning is gone.
Now for some more speed, as the crawl around the block does not exercise the system enough. But 1 km into the trip, another Check IMA system.

Park. Switch pack off. Contact off and on again. Reset the IMA codes. Check the voltage. Start again. Switch the pack on. Get moving again.

Okay, this was still in suburbia. I'm now close to a 40 mph road.
Then at last it can run free. And it does. Until the Check oil level appears again. I hide it with the Mode controls on the steering wheel.
Fuel consumption is still around normal. Hm.
I play with the gas pedal to provoke EV mode. It happily complies. Gently I start accelerating. The boost needle swings up. It works...

PING Check IMA system.
Ritual. Get going again.

Gently acceleration is the key. One 50 mph road, doing fine. Traffic lights, wait. Engine keeps running (no auto stop) Hm. No error though.
On to the highway. The engine RPM climbs mildly while the boost swings high.
It pulls quite well now for those low revs!
But you guessed it. Again. Luckily, once at speed on the highway the IMA system is not really necessary.

Leaving the highway at my home town, I park and reset the code yet again.
For the remainder I get spared from further codes, until I hit the speed bump area - another IMA fault code to rub it in on the home stretch.

The resting voltage on the batteries is still at around 120 Volt.

-------------------------------------------------------------------

So I have to let this sink in.
The voltage on a fully charged 36 cell LiFePO4 pack seems to be too high for the IMA system to handle with ease.

The batteries themselves hold up well. They stayed at ambient temperature, just like the cables and control box.
The voltage hardly sags; I expected it to drop to around 117 Volt unladen, but even after a 15 mile trip wit at least some extra support, so it has delivered power, yet the voltage hardly goes down.

The relays do not cause problems after all. It was the dodgy voltmeter that caused the spikes I saw.

I could reduce the discharge rate by lining up a couple of lamps. Maybe that keeps the errors away. And I can make a switch to operate the relays from my drivers position instead of in the boot, maybe even a voltmeter (though that obviously cannot be fed by the 12V system).

I will check out on those ODBII fault codes, but first some sleep now. To dream of sheep jumping an electric fence, probably

[S Keith]

120V on an 84 cell pack would be 1.43V/cell. The IMA pack probably only sees that voltage when it's near max SoC and charging at a high current. I expect the BCM is freaking out about high voltage when there is no charging. Can you just pull or bypass 1-2 of the LiFePO4 cells?

[RedDevil]

Yes, I do think an overly high voltage is the issue here. The buddy pack still rests way over 117 Volt. But once connected the IMA battery pulls down the voltage to around 110 Volt and still causes the error.

I could drive for a minute or more without issue. What triggered the error was when I started to use EV mode. That drops the IMA pack voltage and increases the current flowing in from my buddy pack.

I may use a lamp for a shunt to reduce the voltage difference and current.
I will make a driver side controller with a 3 way switch; off, on, lamp. And probably a voltage and current flow meter with an isolated power source. As I feel I need more insight (!) into the conditions that trigger the errors.

[RedDevil]

Hooking up my batteries with the IMA system did result in IMA errors. But what if I use a shunt?

First run - single filament lamp
I hooked up a 55W H7 lamp in series this time.
My pack voltage is at exactly 120 Volt now. Each battery at 40.0 Volt, apparently perfectly balanced. Just like it was after last test.

When I hooked it up without the shunt the voltage dropped to 109 Volt, and it threw a code.
With the light in series, when I flip the switch...
It burns quite bright, not full strength but still impressive. And the voltage on my pack drops to a whopping 118.5 Volt; just 1.5 less!

I guess it burns away at over 3 Ampere so the total internal resistance of my pack must be less than 0.5 Ohm!
Way less than my 0.8 Ohm estimate. So the current was much higher than I expected.
No wonder it threw codes!

Light still shining, no codes whatever, I start the engine. The lamp shines really bright now; it is taking on charge at speed. But it recedes quickly once on the move.

It is quite easy to provoke and sustain EV mode, which makes the IMA voltage drop and the lamp shine white bright. Once it re-enters gas mode the light dies and it even says on the dash that it is recharging the IMA battery.
Meanwhile, for the first time in all my testing, the trip fuel consumption is getting lower than would be possible without the pack. It is doing something. And it is getting better all the time.

Out on the open road, I accelerate gently to 60 km/h. Upon stabilizing the speed it goes into EV mode, which otherwise rarely happens on the level. The light shines bright - too bright. It blows.
The IMA battery, that was taken somewhat above its usual SOC, keeps the FE good for a while. Then it all gets back to normal again.

Second run - double parallel filament lamp
Back home I figure one lamp may reduce the current too much. I also have a twin beam 55/60 Watt lamp. I hook it up in parallel so it could accept double the current.
Hopefully the higher current will reduce the voltage difference so these won't blow.

I flip the switch again with contact on, engine off. PING Check IMA system.
Okay, so this was too much.
But it should work with the engine running.
Reset the code, start the engine, flip the switch.
It runs fine, lamp shines bright. Again I can provoke EV mode with ease.
Fuel economy starts to rise again, leaving common read values behind.

Then PING Check oil system.
Ah, we know that; Honda's way of teasing you one mile after a IMA code reset. Key off and on while on the move, that should clear it. On key on the lamp in the back flashes bright and fails. And the system throws a Check IMA system message to boot. Oh joy!
I limp back home, reset the code and remove the lamp and batteries.

Conclusion
It is a success, at least it showed that one lamp is all it takes to prevent IMA errors. But I need to put 2 in series as it would go bust if the IMA system gets exercised. And I have been burning though my stack of wrongly typed halogens at speed.

And I need a new control box. Because the old lamp box is starting to fall apart, it is not sturdy enough after all.

I want that lamp but I also want to be able to switch it out of the system when I'm on the motorway doing a steady speed. I bet it won't be needed anyway once the pack voltage drops on discharge to below 117 Volt or so.

So I need to route a 3 way switch to my drivers position, and probably a volt and possibly ampere meter too. I'll solder in some high ohm resistors on the high voltage lines to the meter so a short would just cause a tiny current. The meters should not take much current anyway (the ampere meter shunt stays in the back of course).

I'm confident I can make this work now.

[RedDevil]

Yesterday evening I built a new resistor bank by using 6 12 Volt 20 Watt halogen spots from my bathroom which replaced them with LEDs a few years back.
3 parallel pairs of lamps switched in series, they should dispatch 5 Ampere at 24 Volt, which is about as much as the voltage difference between my battery and the IMA system should get.

I mounted the lamps in a door rack from an old fridge, slightly slanted so air can flow around the lamps, and partly open at the back. The lamps are fully isolated by themselves.


Nice, no?

Lets test. Key on, flip the pack switch:


That's about 3 Ampere flowing toward the IMA system. Not enough to trigger IMA codes.
I can see the glow in my rear view mirror.
The IMA battery symbol in the car shows 3 bars below full.
I start the engine. The glow recedes. The Insight starts to regenerate (???) and the battery quickly fills up to one bar under max.

Once the engine is warm, time to trigger EV mode.
It works. The lamps glow yellow and gently get brighter. The battery level indicator starts to drop. And the fun lasted just a minute, the engine kicks in and starts regenerating. The lamps are off again.

The battery replenishes quickly again, at least that part is covered.

Time to leave town and take it out on rural roads and the highway.
Constant speed, CC on.
The economy appears to be no better than usual. But the current flowing in must be small - the lamps hardly glow - and I have my HIDs on.
At least it does no longer throw codes.

A T junction ahead - and on this time of night there are cars speeding. Time for some very brisk acceleration!
No problem. It revs like the best of times and support is maxed out.
Catching up with the cars I just let by, it regenerates again under braking.
A stop at the traffic lights and then an uphill entry to the interstate. Someone is pushing from behind? Aha. Someone is getting smaller in the mirror. Merge into the 5 lane interstate, at more than the average speed there. The someone passes me way over the speed limit.

I gently let the speed drop off, but going against the wind the fuel economy is just normal for the conditions.
Battery level is higher than usual though.
On leaving the interstate the Insight does only marginally regenerate, it keeps coasting by itself. Feathering the brake pedal brings it back to its usual level.

Getting home in the reduced speed residential area I get full EV mode for a long time.
All in all I get more EV mode and more support, and some less regen, than without the buddy pack.
But it hardly affects the feel, and the effect on economy is marginal.

This morning we went to the swimming pool.
This involves a short trip on the highway over a high bridge immediately followed by an exit.
I always get bad FE on those trips.

Today was slightly better, at least on the way in; way better than usual.
Going back it did trip an IMA error when I switched on. I reset it and restarted before I even drove 1 meter, but apparently an IMA code / reset does cause some reluctance to use it well.
In the end the FE was coming back again, and all in all I got a pool trip economy on this cold morning that compares to the warmest of summer mornings.

So, yes it does something. And these lamps (almost) got rid of the IMA codes.

It could do a good deal more once I know the limits better, and actively allow the current to approach those limits at all times instead of just when it is in EV mode.
I do need a controlled current limiter of some kind, and meters to tell me the exact voltage and current flow.

I'll do a few commutes with this setup, I know those conditions best.

[RedDevil]

Monday morning I did the morning commute with the lamps in series.
They did glow up nicely on switching on, then for the rest of the drive I did not see any light in the mirror. Hm. Battery charge level resided one bar under max. It did some EV mode and some light forced regen - just for short bursts.
At work I checked on the lamps. They were lukewarm. So they did something, but not much.

I ended up at work with a 3.6 l/100km average. Not bad, but I've seen better without EV support!

On my commute back I hooked up the battery directly to the IMA system, thinking (without a meter at hand to check it) that the voltage on my buddy pack should have sagged enough to not make it throw codes on the IMA sytem.
And it did not. I could run it on the highway, even got in a bit of a jam, without any codes whatsoever.
But the economy wasn't there either... and the battery charge level actually dropped a bar!

When freed from the jam the average consumption hovered around 3.0 l/100 km, but it showed erratic drops to low 2s. Nonetheless, fuel consumption remained fair but not spectacular.
A long wait at a traffic light did not throw a code.
Only when nearly home, braking for another traffic light did it suddenly trow a code.
I parked it immediately after the lights (where there was a parking spot) and reset the system and switched the pack off.
Again my FE for the trip home was good for the conditions, but not spectacular.

Tuesday morning I decided to put the lamps back in series again hoping to see the light strength in the mirror (the morning commute is in the dark now). But the lamps refused to even glow!
I disconnected everything and just rode it to work and back without support.
Both commutes were 0.2 l/100km worse than yesterday while the weather was warmer.

I measured the voltage of my buddy pack when home. It had sagged just a little, to 118.2 Volt. It would have lost less than 10% of its charge on Mondays commute!

I found that the cause for the lamps not lighting up were the tab connectors on the wires I used to hook up the lamps - the same that hook up the 32 Amps fuse.
I had ground the tabs on the double filament lamp narrow to the ¼" width of the connectors, but the thickness of those tabs had made the connectors bend out a bit and they did not connect properly to the slightly thinner contacts of the fuse and 6 lamp batch (for which I used the connector of the blown H7 lamp).

A friendly nudge from a pair of pliers and the connectors were tight again. The lamps lit up with a golden yellow glow. Tomorrow there's another chance to test them.