Honda Insight w/Acura TSX engine - Ecomodding

[serialk11r]

I don't know if 60% is the right conclusion to be drawing, since no load operation is affected by incomplete combustion. It's certainly a horrific number.

How's your IAT? Have you experienced any knock? You could try a more aggressive warm air intake (maybe with a manually controlled air door in case it gets too hot). The engine speed is in a pretty good spot IMO, but you're pulling too much vacuum. If you can successfully get the IAT higher than the engine can tolerate, you should be able to change the cam timing to reduce internal EGR but delay valve closure which will reduce your pumping losses further.

[Ecky]

Originally Posted by serialk11r

I don't know if 60% is the right conclusion to be drawing, since no load operation is affected by incomplete combustion. It's certainly a horrific number.

How's your IAT? Have you experienced any knock? You could try a more aggressive warm air intake (maybe with a manually controlled air door in case it gets too hot). The engine speed is in a pretty good spot IMO, but you're pulling too much vacuum. If you can successfully get the IAT higher than the engine can tolerate, you should be able to change the cam timing to reduce internal EGR but delay valve closure which will reduce your pumping losses further.

I've thought about changing out the head for one that closes one of the intake valves at low RPM, as I figure it would help combustion quality in the regions I operate in, but I'm sure there's some low hanging fruit in ignition timing, air fuel ratio and cam advance.

Intake temps vary; I have some minor issues with heat soak. With my grille block fully removed and outside temps around 80F, I'll see my intake temps hover around 100F while driving. Last summer I recall seeing them get over 140F when in traffic.

The thermostat starts to open around 170F and is fully open around 190F. My radiator fan is set to come on at 195F. I had to pull all of my grille blocks off when it got above ~70F outside as I was seeing the fan cycle under normal highway driving. When outside air is even a bit cooler however its hard to get the engine up to temperature. My radiator is pretty small and is also fairly beat up. What I believe I need is a better radiator and to restrict airflow over the block when outside air is colder.

When coolant temps reach ~200F I start to get knock at WOT between maybe 1500rpm and 3000rpm. Intake temps seem secondary to that. Currently waiting to see if a Facebook contact can come through with a custom radiator that fits in the space of the stock one.

I'm hoping to play with cam advance this weekend, and build separate timing tables for different amounts of intake advance. I expect you're right in that delaying valve closure could help.

[serialk11r]

Seems like this project is getting expensive :P

Sorry I forgot what radiator you're using, is it the stock Insight unit? Bigger radiator is definitely preferable to fans coming on or less grill block for sure.

If you're heat soaking and knocking already, that's not great, I would give DIY EGR a shot. EGR has higher gamma and low oxygen which both help reduce knock compared to hot air. The stock cam timing should be near best compromise for EGR, combustion stability and pumping losses, external would EGR allows modified cam timing while retaining (or even increasing) EGR fraction.

You would need some way of not messing up the fuel trims since the Honda ECU doesn't use MAF. There has to be some kind of internal compensation built in, but without knowing how it works, it's hard to say if you can go mess with it and expect it to work. If it has a full fueling compensation table that can be learned for all closed-loop conditions, you don't need precise electronic EGR control.

A crude system to control EGR would be two valves in series: One that varies aperture in proportion to the throttle (to prevent misfires from too much EGR at high vacuum), and one that closes off near full throttle. This will produce decreasing EGR fraction with increasing throttle.

One passive method would be to attempt to meter a relatively constant fraction of EGR except at idle, which can be done with a "carburetor" like mechanism feeding exhaust ahead of the throttle plate. A signal from the idle control valve can shut off the EGR supply, and full throttle can also shut it off.

As far as fuel economy goes, retarding the intake cam by ~30 degrees will drop VE by something like 13%, and if you can get slightly higher EGR dilution, you might be able to get 20% drop in VE. Bringing MAP from 0.5atm to 0.6atm is probably worth something like 3-4% in fuel savings.

[Ecky]

Originally Posted by serialk11r

Seems like this project is getting expensive :P

Sorry I forgot what radiator you're using, is it the stock Insight unit? Bigger radiator is definitely preferable to fans coming on or less grill block for sure.

If you're heat soaking and knocking already, that's not great,

Yeah, stock Insight radiator. Some more thorough ducting as another member is doing on here could help with this.

I want to say that on my last 10 commutes, only on one did I measure and knock. It was 5 total knocks and they all happened at 1750-2250rpm at full throttle, where volumetric efficiency is highest - I believe I've seen as high as 107% there. I pulled a half degree of timing from the very top of the map and haven't yet seen any more, but I'm keeping an eye out for it as the weather warms.

Originally Posted by serialk11r

I would give DIY EGR a shot. EGR has higher gamma and low oxygen which both help reduce knock compared to hot air. The stock cam timing should be near best compromise for EGR, combustion stability and pumping losses, external would EGR allows modified cam timing while retaining (or even increasing) EGR fraction.

You would need some way of not messing up the fuel trims since the Honda ECU doesn't use MAF. There has to be some kind of internal compensation built in, but without knowing how it works, it's hard to say if you can go mess with it and expect it to work. If it has a full fueling compensation table that can be learned for all closed-loop conditions, you don't need precise electronic EGR control.

A crude system to control EGR would be two valves in series: One that varies aperture in proportion to the throttle (to prevent misfires from too much EGR at high vacuum), and one that closes off near full throttle. This will produce decreasing EGR fraction with increasing throttle.

One passive method would be to attempt to meter a relatively constant fraction of EGR except at idle, which can be done with a "carburetor" like mechanism feeding exhaust ahead of the throttle plate. A signal from the idle control valve can shut off the EGR supply, and full throttle can also shut it off.

The knock only happens at full throttle, where cylinder pressure is highest, and very slight reductions in timing advance and temperature seem to improve it a lot. I'm not getting any at even 95% peak MAP.

I'm not sure I have access to the original Honda maps. Hondata publishes some base maps which are supposedly "equivalent", but the ECU I'm using is for an entirely different engine.

As for tables, KPro has a different set of tables for every 10 degrees of intake cam advance, for both the low and high cams. This includes ignition timing, fuel, knock sensitivity, knock ignition limit, and knock retard.

There are full maps for lambda, and offset tables to modify lambda based on coolant temperature, at low and high load, on it low and high cams, plus a separate table for lambda when it crosses into the "WOT" region (which can be set).

Fueling can be modified by tables for intake temperature (low/high cam and load regions), cam angle, coolant temperature, battery voltage, and what gear you're in. Injector phase can also be modified by map region.

It supports flex fuel (with a sensor or fixed values) and has lots of compensation tables for ethanol percent.

Ignition timing can be modified by whether A/C is running, engine load region, air temperature, water temperature (multiple tables), battery voltage, as well as set dwell angle.

One can also have triggers, both inputs and engine conditions, which cause various events (if then, and/or) such as activating outputs or changing fueling or timing.

There is traction control, the ability to use secondary intake runners, nitrous support, electronic boost control, tip-in parameters, rate of change, but nothing natively for EGR. However, I bet I could set it up with if/then statements.

Originally Posted by serialk11r

As far as fuel economy goes, retarding the intake cam by ~30 degrees will drop VE by something like 13%, and if you can get slightly higher EGR dilution, you might be able to get 20% drop in VE. Bringing MAP from 0.5atm to 0.6atm is probably worth something like 3-4% in fuel savings.

General wisdom in the Honda community is to advance the intake cam to get fuel savings, but I didn't see any. My guess is it's largely due to my not having different ignition tables yet for the various cam advance positions.

I plan to work on that this weekend.

[serialk11r]

If you have to manually compensate for cam angle adjustments, that really sucks. The wisdom of cam advance for fuel economy comes from the fact that you get more internal EGR with cam advance, but it's a tradeoff. The highest efficiency method is maximum intake cam retard and cooled external EGR.

I would try an external EGR tube feeding ahead of the throttle plate with a restrictor orfice (so that most of the restriction comes from the orfice, not the tube), with a shutoff valve. That will get you your extra EGR with little disruption to fueling.

[Ecky]

Filled up with ethanol-free premium this most recent tank, and I had to adjust my fuel trims by 5%. My math tells me AFR for E10 should only be ~3.3% different. Now that it's warming up, I'm seeing some huge fuel economy improvements.

[Ecky]

Broke a 70mpg trip last week. Also got my oil analysis back from Blackstone.



My reply back:

Thanks for the report!

Some information, if it's helpful:
I swapped the 1.0L engine in "Gaptooth" for a Japanese market equivalent of a Honda K24A2 (from the 04-08 Acura TSX). The engine looked good inside and was advertised as low miles, but the actual miles are unknown. I'm unsure where the lead comes from in this engine but it could well be crank bearings.

The oil used was Mobil 1 0w30 (standard) and the engine was run quite hard over a 7500 mile interval, with a lot of bouncing off the rev limiter and WOT pulls. It also has very tall gearing, so there's a lot of very low RPM, high load operation.

The numbers look pretty good to me, but given the information I've provided, do you have any insights or cautions for me? I replaced the oil with Mobile 1 0w20 Extended Protection, and will be sending in a sample of that at the next 7500 interval.

Interesting to me is that a 2.4L engine, driven far harder and with way more moving parts has lower wear indicators. There's also more oil in the pan, but liter of displacement to liter of oil, the K24 has less.

[HyperMileQC]

Wow guys! didn't know you where still at it after all this time. I will need to read this whole thread now

[Ecky]

Had my second 70+ mpg trip yesterday. Admittedly it was overall slightly downhill, but it was still 55 miles and not much elevation change. I suspect that with warm conditions, an already warm engine, and keeping speed at 55 or less, 65mpg is a reasonable target to aim for. The gauge read 70.6 when I pulled in at home, but that doesn't take into account my 1.3% oversized tires.

[MetroMPG]

Salut, HyperMileQC! Long time no see!

Ecky: impressive results, man.