K20Z3/KPro refinement improvements

614 Views 49 Replies 5 Participants Last post by mattkosem, 19 h ago

mattkosem

Discussion Starter · 8 d ago (Edited)

I made a breakthrough realization in this thread over here (sorry again for the threadjack @SaucerClone). @Lotus and @LotusElise made some important distinctions that lead me to believe that my tune is, to be kind, unfinished.

My car has a mildly lumpy idle, which some tinkering with the IACV duty cycle has actually already improved quite a bit. I believe I need to spend a bit of time tinkering with the idle ignition control and idle ignition compensation areas a bit to dial it in further.

My larger issue is that it seems like the tune I have has no adjustments in the areas where I'm having issues...

Neither of the 6 ignition tables (low or high) has a single adjustment from the "k20a2 factory calibration" map Hondata provides
The zero degrees low-speed fuel map also has no changes

The car surges badly on/off throttle at low speed (for example, subtle throttle input changes while cruising at 30mph in 4th gear cause the car to lurch pretty badly). Lifting from very low to zero throttle also causes the car to lurch badly. In the discussion above it was called out that too much ignition timing can contribute to this sort of behavior.

If I'm understanding things correctly, it sounds like picking a region like this:

Colorfulness Azure Rectangle Font Screenshot

...and adjusting it downwards by a few degrees should help. This may need to be repeated across the 10°, 20°, etc. tables as well depending upon where the car actually is when the bad behavior occurs.

I'm assuming I should probably data log and adjust fuel down here if it also seems to help. Sorry for the crazy noob questions, but I have to start somewhere. I dug up and watched the old training videos Hondata used to help and read some articles on general theory for doing this, but I think I need to pay for one of these various training courses (Evans, etc.) to get more in-depth background.

In the meantime, though, if someone can confirm my proposal above and tell me if I'm on the right track, it'd help my car be more pleasant at low loads.

Full disclosure, I am not crapping on the tuner or the tune. This was done during the winter of 2020/2021 by a local shop willing to take on my job even though it isn't in their usual wheelhouse. The shop did a fantastic job with the fit and finish of the physical pieces. They seem to typically be tuning much newer cars, almost all of which use forced induction, so I assume that they got it running acceptably, tuned it for power to the extent they were comfortable, and called it a day.

--Matt

See less See more

1 - 20 of 50 Posts

Lotus

· #2 · 8 d ago (Edited)

i think you are getting into the right direction.
Check the low cam map for where you are actually “taking off”. It is usually 0-20° VTC. My Elise does not want 18° ign there as the response becomes to jerky.

Reactions: 1

Lotus

· #3 · 8 d ago

you can leave the idle settings alone for now. Idle duty cycle and fueling is all that needs adjustment of idle for now.
It is easier to map if you run it open loop. The first thing now to get right is also the idle fueling. This is done in the main map.

Reactions: 1

mattkosem

Discussion Starter · #4 · 7 d ago (Edited)

4

I went for a little drive now that I've learned how to datalog. The whackadoo throttle behavior happened right around 4:26 here when I let off the throttle:

Product Rectangle Slope Plot Purple

I put the idle ignition control values back to stock, which seems to have helped a bit more. My idle presently looks like this:

Rectangle Slope Font Line Plot

Fuel and timing seem kinda "jumpy" like the idle. Map fluctuates while this happens, lowest I saw was -14.9 highest was -16.0. I don't see a place in here to log anything about it, but I can hear the IACV make its little hiss while this goes on.

--Matt

See less See more

mattkosem

Discussion Starter · #5 · 7 d ago (Edited)

A bit of reading suggests that the ELD function (enabled presently) can cause this sort of thing on swapped cars. I see my battery voltage is jagged as well, which is an accompanying symptom. I guess I'll shut that off and see how it goes. It seems like it shouldn't be on in this car either way.

Apart from that, it seems like some folks long ago saw similar from the O2 heater circuit drawing too much voltage on a common pin from the ECU (solved by providing a dedicated fuse to the relay, apparently). Hopefully, after all this time, my k-tuned swap harness wouldn't have such an issue. 🤷‍♂️

LotusElise

· #6 · 6 d ago

mattkosem said:
Map fluctuates while this happens, lowest I saw was -14.9 highest was -16.0.

Could you please open the MAP scale to -26 "/psi till 0.5 "/psi? You find it under options - sensors - ...

mattkosem

Discussion Starter · #7 · 6 d ago

Nice tip! Replaced the images above.

Reactions: 1

mattkosem

Discussion Starter · #8 · 6 d ago

As I read/investigate further, it seems like this is probably a contributor to my refinement issues:

Colorfulness Product Rectangle Font Slope

I'm seeing suggestion that, on a cammed engine, dropping down to 0 in all the lower columns at driving RPMs puts a lot of stress on the VTC and timing chain. My understanding is that these things can/should go to zero at low RPMs but shouldn't at 1250 and beyond, for example. In this example, it seems like it abruptly drops by like 15 degrees when drop low throttle to none.

Does this make sense?

--Matt

See less See more

LotusElise

· #9 · 6 d ago

mattkosem said:
My understanding is that these things can/should go to zero at low RPMs but shouldn't at 1250 and beyond, for example. In this example, it seems like it abruptly drops by like 15 degrees when drop low throttle to none.

Honda tunes their engines in low speed cam always to zero VTC below -23" Hg, sometimes, depending on exhaust even lower vacuum level's. The reason is, the lower the plenum pressure, the more prone the exhaust gases will dominate the chamber. To reduce that, the VTC get retarded as much as possible. Mechanical this means movement and bit of more stress during the pin jump in and out to lock at zero VTC. Response-wise this means some delay, not on power, but on lambda, as the moving VTC don't follow the load set point fast enough. The tuner has to decide what is a proper VTC there and can create a different strategy to solve this challenge differently.

Could you describe your engine setup here. Just to understand the WOT and part load VTC better.

LotusElise

· #10 · 6 d ago

BTW, if you check your own approach regarding idle, you still need to adjust your VTC map

.

Reactions: 1

LotusElise

· #11 · 6 d ago

mattkosem said:
Fuel and timing seem kinda "jumpy" like the idle.

If you check the related log, you see why. Hondata don't provide a duty cycle of the IACV, but we can assume yours is quite active. If idle lambda is tuned in, then you should address next the IACV default position.

Basically, the IGT oscillation is showing the idle controller is quite active to adjust torque to keep the set idle speed. Some of it is introduced by the IACV provided torque actuation. Your idle suffers from too much torque from there. You must know, IGT torque actuation need a high sensitivity to adjust. Actually it jumps quite forth and back for not much effort as IGT is already at insensitive area situation. You need more torque from the IGT and less from air/fuel. You adjust this by the duty cycle. If you balance that right, IGT will more advance and the hole system stabilize more. Except, your engine has an untypical friction issue.

mattkosem

Discussion Starter · #12 · 6 d ago (Edited)

2

LotusElise said:
Honda tunes their engines in low speed cam always to zero VTC below -23" Hg, sometimes, depending on exhaust even lower vacuum level's. The reason is, the lower the plenum pressure, the more prone the exhaust gases will dominate the chamber. To reduce that, the VTC get retarded as much as possible. Mechanical this means movement and bit of more stress during the pin jump in and out to lock at zero VTC. Response-wise this means some delay, not on power, but on lambda, as the moving VTC don't follow the load set point fast enough. The tuner has to decide what is a proper VTC there and can create a different strategy to solve this challenge differently.

Could you describe your engine setup here. Just to understand the WOT and part load VTC better.

That makes sense, but Honda's strategy there seems to differ from what I'm finding to be best practices for cammed engines (which suggests that dipping to 0 with regularity puts undue stress on the cam tensioner).

From intake to exhaust, I have:

Hybrid Racing Cold Air Intake
K20Z3 (anything not called out is as provided)
- K20A2 throttle body (Acuity TPS, new MAP from Honda)
- Stock RBC intake manifold
- Drag Cartel 2.2 Cam + Supertech Springs
- Ported K20A2 oil pump
K-Tuned 4-1 header
3" exhaust (all mufflers/resonators perforated/straight through)

LotusElise said:
BTW, if you check your own approach regarding idle, you still need to adjust your VTC map .

Interesting... I was under the impression that idle always drops VTC to 0. Is that not the case?

LotusElise said:
If you check the related log, you see why. Hondata don't provide a duty cycle of the IACV, but we can assume yours is quite active. If idle lambda is tuned in, then you should address next the IACV default position.

Basically, the IGT oscillation is showing the idle controller is quite active to adjust torque to keep the set idle speed. Some of it is introduced by the IACV provided torque actuation. Your idle suffers from too much torque from there. You must know, IGT torque actuation need a high sensitivity to adjust. Actually it jumps quite forth and back for not much effort as IGT is already at insensitive area situation. You need more torque from the IGT and less from air/fuel. You adjust this by the duty cycle. If you balance that right, IGT will more advance and the hole system stabilize more. Except, your engine has an untypical friction issue.

I can definitely tell that the IACV is active, it can be heard plainly. I can see the requested vs actual lambda do vary a bit while idling (see 6:38 on):

Rectangle Slope Font Line Parallel

Is that more sway than would be expected? Would the IACV movement be a cause or response to this? Would the "default IACV position" be zeroed by adjusting the throttle stop such that it doesn't need to actuate? I had the IACV deleted for a bit last year and can confirm that my idle was smooth at that time. I didn't like it otherwise and replaced the IACV with a new one.

At present, my idle air control is 25% in. It was originally in the middle, adjusting it to halfway between there and the current position helped. Adjusting from there to the current position helped more. So, perhaps, I may need to adjust it down further? I'm certainly not seeing any hint of idle dropping too low, and adjusting it downward has had other positive impacts (it drops more predictably to idle without hanging).

Rectangle Azure Font Screenshot Parallel

Thank you for your help with this, BTW. I'm very new to these things, and have found your input insightful!

--Matt

See less See more

LotusElise

· #13 · 5 d ago

mattkosem said:
From intake to exhaust, I have: ...

That is a very nice and simple 86x86 engine setup, which I really like beside the KTuned 4-1 header. Must do a lot of top end power when properly tuned, not so sure about the midrange (header).

mattkosem said:
Interesting... I was under the impression that idle always drops VTC to 0. Is that not the case?

For a really stable idle, see my explanation regarding combustion properties at idle, I would recommend you to zero below 1100 rpm and below -17" Hg. Cold start idle at IAT below freezing point can be as high as 1600 rpm and demand up to -14" Hg. That might be a consideration for a Winter season car, a Summer season car should at least cover the former stated above.

If your IACV demands a duty cycle of around 25 % then I would check for an leakage at the IM/TB interfaces as soon as possible.

mattkosem

Discussion Starter · #14 · 5 d ago (Edited)

LotusElise said:
That is a very nice and simple 86x86 engine setup, which I really like beside the KTuned 4-1 header. Must do a lot of top end power when properly tuned, not so sure about the midrange (header).

I don't love it either honestly. It leaks a bit at the slip joint and I've had to replace 4 springs so far. k-Tuned don't make a header in 4-2-1 that addresses either of those issues without a big power loss. Their standard 4-2-1 header, according to them, would lose me a minimum of 5HP and all the rest demand more space and still have the slip connection.

Perhaps, one day, I'll try PLM's 4-2-1 header. At least their slip fit is held by bolts 🤷‍♂️.

LotusElise said:
For a really stable idle, see my explanation regarding combustion properties at idle, I would recommend you to zero below 1100 rpm and below -17" Hg. Cold start idle at IAT below freezing point can be as high as 1600 rpm and demand up to -14" Hg. That might be a consideration for a Winter season car, a Summer season car should at least cover the former stated above.

If your IACV demands a duty cycle of around 25 % then I would check for an leakage at the IM/TB interfaces as soon as possible.

The car never sees winter. 1100rpm and down at 0° is consistent with what I've otherwise seen. It's the higher points going to zero that seem counter to best practice.

You lost me on that last part. The IACV duty cycle defaults to 50% so I'm at half the standard value. Will check for leaks either way. I had one at the IACV and fixed it, but there may others still.

Reactions: 1

mattkosem

Discussion Starter · #15 · 4 d ago (Edited)

I had a bit of time to smooth out some bizarre sawtooth patterns in my 0, 10, 20, and 30 degree low cam fuel maps and drive the car through them (all lowish load lowish throttle).

Driveability through that portion of the rev range on the street improved massively!

There are some anomalies I'd like to sort out on my cam timing maps next, then I need to spend a bunch of time filling out lambda overlays to find tune the fuel. The low and high cam maps drop to zero at low revs and low vac, which is apparently no good for tensioner reliability so I'm setting them up to taper to zero at low revs only, and only on the low cam. I know I'll never actually be in the low rev range on the high cam, but I would like to make them conform with what I've learned thus far to resemble a "properly tuned" setup without trying to push the performance envelope any.

Haven't pushed IGT anywhere, nor pulled fuel anywhere but idle where I was rich so far, and data logs while driving look safe and clean so far. 🤞One area where I was getting knock in those jaggy fuel areas now seems not to be, so that's a plus. I'm sure I've lost power at the moment...but that can come back with time.

--Matt

See less See more

mattkosem

Discussion Starter · #16 · 4 d ago (Edited)

9

Nothing earth shattering. It was just the 4 tables where I was noticing dips that aligned exactly with annoying refinement problems. No asking for any help or anything, since there's a process for all this I'm only early into, just sharing my progress for posterity (or no one 🤷‍♂️, whatever, it makes me feel better to share for whatever reason).

0 from:

to:

10 from:

to:

20 from:

to:

30 from:

to:

I know, I know, really nothing major here... The doozy on the 20 degree map caused a noticeable annoyance, which I can no longer reproduce.

--Matt

See less See more

mattkosem

Discussion Starter · #17 · 4 d ago (Edited)

4

Tentatively leaning toward moving low cam timing like this:

->

And high cam timing like this:

->

VTEC cutover is 5000-5400, so the differences in the edges are likely inconsequential. I can tell pretty clearly that the shop definitely tuned this area, so my intent was to preserve what they did (+/- a bit here and there) in that spot but make everything else more gradual.

Should hopefully have some time over the weekend to suffer through some miles of open loop logging. 🤞

See less See more

Bjorn

· #18 · 4 d ago

40 degrees seems easy too much. I found no real power gains going over 30 degrees on my k20. It's good that you're systematically tidying up all the vtc maps. I feel like this gets neglected a lot on people's tunes.

I'm curious about why you think 0 deg vtc is hard on the timing chain tensioner.

mattkosem

Discussion Starter · #19 · 4 d ago

Yeah, it probably is. 42 is the max Drag Cartel recommends with the pistons I have, and it seems uncommon to need that much VTC.

It's not my own idea. The instructor in the online tuning classes I'm taking mentioned it a few times over the dozens of courses I've watched so far. Evidently on engines with already exacerbated oil pressure fluctuations from bigger cams, as VTC drops to 0 the timing chain has a tendency to lose tension a bit. I may be misstating some details there but even on one of the live tuning video segments on a car with DC2.2 cams it was recommended.

Might be more of a concern with Skunk2 cams. 🤷‍♂️

--Matt

yaggs

· #20 · 4 d ago (Edited)

Those with more experience: please correct any errors i may make in this post
I had to rewatch alot of those jeff evans videos several times. I spent alot alot alot of time making revisions to my street tune but i am extremely happy with how it came out. It's kinda addicting to get in there and mess around while striving for more performance and drivability.

My take:
make sure you understand the concepts (fundamentals) before moving to the specific application videos. I learned the most from the "live tuning session"

Understand that kpro interpolates between vtc, ign, and fuel so unless you are directly in the middle of a specific cell at exactly 0, 10, 20, 30, 40, or 50 degree cam the surrounding data points will effect the input (think 3d). For example: at 33 degrees cam, both the 30 degree map and 40 degree map affect what fuel and ign get delivered.

You're going to have higher and lower values in your fuel and ign graphs. You should see a general trend though. Smoothing the ign and fueling goes a long way. It makes the transitions throughout throttle and rpm more linear and smooth. You will become a master with the "interpolate horizontally/vertically" options

My k20 also likes 30ish degrees on the high cam. I think i remember making more power going into the 20s near redline. 40 degrees straight from crossover to redline seems unlikely the ideal tune and tbh seems a little lazy on the tuner

Take a 20-30 minute drive while logging. Look at your lambda values in your fuel maps. This will allow you to fine tune the fuel values at partial throttle. Make changes in blocks, not individual cells. Keep those graphs smooth.

I had the same question about vtc btw:

www.k20a.org

Cam angle on deceleration

Hey all, In reguards to Kmanager for Kpro, Conrad talks about setting the cam angle to zero on the first two columns on the low cam. Quote from his advanced tuning write-up: “Looking at the low-speed cam angle map, the cam angle in deceleration columns 1 and 2 should be zero or close to zero...

www.k20a.org

See less See more

1 - 20 of 50 Posts