Hondata

- serial number of K-Pro: 11084
- ECU part number: 37820-PRA-E12
- serial number: 188226-4113
- KManager version: 3.1.2 up to and including 3.2.0
- what sort of vehicle you have: Lotus Elise

Perhaps this is simply the way it works and a by-product of the software being geared towards the PRB ECU with a wideband, but I just thought I'd ask..

When I make some datalogs of my car I noticed that the reported primary O2 sensor voltage is not between the expected 0-1V, but is only reported (when running) as cycling between 0-0.25V or so.

The car is running closed-loop just fine, fuel trimming is done (although long-term seems nailed at 0) and an OBD reader connected to the DLC shows it cyling between 0.2-0.8V as expected, so the datalog showing an odd voltage doesn't 'break' anything in itself.

Just curious if it's supposed to do this.. I worked around the problem by setting the variable range smaller so the 'wave' as it cycles lean-rich becomes better visible on datalog graphs.

Attached is a small excerpt from a datalog from my car..

BTW.. The datalog does show another oddity which has been noted on the forum here as well.. The acceleration retardation seems to kick in even with very tiny or no visible throttle change (looks like it's reacting at a hair-trigger..). I have a K-tuned TPS on order and will swap that for the original OEM one first before looking further into this as well as reduce the tip-in retard gear muliplier..

Bye, Arno.

The wideband is current based and not voltage based.

Spunkster wrote:The wideband is current based and not voltage based.

OK, but the PRA uses a narrowband which is a simple voltage based cell and current is irrelevant for these? (up to a point ;) )

I suspect you mean to say that the narrowband O2 sensor on a PRA ECU with a K-pro is not really scaled or set up in the software to return a clean 0-1V signal in the datalogs so this is basically 'how it works' at the moment?

No biggie.. Some re-scaling of the O2 voltage variable gives a useful output in the graphs anyway..

Bye, Arno.

The ignition retard you are seeing is 'acceleration smoothing retard'. See http://www.hondata.com/help/kmanager/ge ... meters.htm

Yes, the narrowband o2 is voltage based and would normally give a 0-1V output. The scaling may be out, but as you pointed out, it still works ok.

Hondata wrote:The ignition retard you are seeing is 'acceleration smoothing retard'. See http://www.hondata.com/help/kmanager/ge ... meters.htm

That's correct, but the ECU seems to decide I'm accelerating and need acceleration smoothing retard when the TPS is not changing at all or at most only 1% or so when I look at the datalogs.

I'm curious as to why the ECU would decide that it needs to engage acceleration smoothing retard action at all for such minute TPS changes..

I'd personally only expect it to kick in when it detects something like a 10 point positive delta in the TPS within <0.5sec or so which would indicate a fast throttle change and the need to apply the acceleration retard (and probably kick in the acceleration enrichment too).

Geeking out a little here, but I hope you don't mind :)

I get the impression that the code is somehow reaching an incorrect conclusion at small(ish) steady throttle openings and any resulting small variation in throttle position readout.

Almost like it's not doing any hysteresis damping and not disregarding small 'noise' like throttle changes in favour of latching onto any change to trigger acceleration smoothing retard.

For instance, I can imagine that if the current algorithm to determine if it needs to engage acceleration smoothing retard uses a throttle position change trigger that's a percentage of the current throttle position that it would trigger very/too rapidly on low throttle positions, but work OK on bigger throttle positions.

Eg. if the trigger is defined as a 10% change in throttle position then the absolute delta at 50% throttle would be a delta of +5% or more, but at a 'cruise' 4% TPS the change required to trigger would only be a delta of +0.4% in throttle position which would be the tiniest change in throttle pedal position..

Hmmm.. Something I can test out.. I'll have to try this to see if the effect indeed varies or increases with a smaller or larger base throttle position. Even if it doesn't return anything it'll keep me busy :)

In my case it could still be (extremely) short, but large, spikes in the TPS signal that don't show up on the logs from a TPS that's going bad. Which is why I'll replace the TPS soon with a new one (once it finally arrives that is!).

Wiring itself all checks out OK, even when testing it and moving the loom and connectors arround to simulate vibration and movement.

If that doesn't help I'll try reducing the tip in retard gear multiplier to see if I can take the hair-trigger away from the tip in. Should help as it's obviously a lot less or even gone in higher gears.

All in all it's not really a big issue, but the seemingly random 'stutter' during steady state driving caused by the acceleration smoothing retard is somewhat annoying.

Anyway.. Stuff to test and figure out! Good!

Yes, the narrowband o2 is voltage based and would normally give a 0-1V output. The scaling may be out, but as you pointed out, it still works ok.

Yup.. Seems that internally it's working fine and correctly adjusting the mixture, it's just in the sensor data that the value displayed is roughly 1/5th of the actual voltage on my PRA ECU.

Bye, Arno.

I think the Honda smoothing is mostly based off crank acceleration. I've seen problems, like yours, with a good TPS sensor.

Remember that Honda designed the smoothing to work on a 100% stock vehicle, so once modifications have been made it may not work as designed.

My personal opinion is that it was a bad idea and I always disable it.

Hmm.. You may be onto something there..

The engine itself isn't that much modified (K20A2 with DC5 pistons, TODA A3 cams and valvesprings), but the engine is mounted pretty solid in the car (and in the back of the car vs. the front ;) ) compared to the stock setup on an EP3 Civic so the crank speed may well vary much more simply beause of changes in road conditions (bumps, ruts) feeding back from the gearbox to the engine as it can't flex as much to compensate.

I'll play with the settings first. Probably works as I already noticed that 5th and 6th gear show hardly any of this and it's set to -100 in the default config.

Thanks for the info!

Bye, Arno.

Just disable it - gear multiplier to -100 and duration to 0.

We have a Lotus with a Honda engine in the back here too.

So its best to disable smoothing as well on my CTR EP3 modded with injen CAI?

Try it both ways.

Is the basemap of EP3 tuned with CAI already set up with smoothing disabled then?

So whats the effect if i set the tip in retard to 0 everywhere?
Tip in retard gear to -100 everywhere
And duration to 0 everywhere

Does this have impact on gas response then? Do i need to correct my normal timings after doing this?

Try it and see.

Yesterday evening I did some testing for 3 hours with the acceleration smoothing settings.

What I can conclude is that the car drives much smoother without the smoothing actually.
Gas response seems better as well, weird thing is that under 4000 rpm it drives a littlebit different, like it builds up torque a little better.

Can these settings have any impact of the performance in higher rpm +7000?
Normally this should have no impact on the horsepower peak i guess?

Or is this just an imagination that the car runs smoother without the ignition retard?