changing injector dead time, what about VE?
 

Hello,


I'm having some minor trouble with the engine idling @ 900rpm. It's a 1,6 '90 running batch fire and it should be a bit smoother (less shaking and less vibrations) More like OEM, which was really sweet.

I'm running MS2PNP with Yellow RX8 450cc injectors @3bar and saw that my injector dead time was standard programmed at 1,2ms. The right number for this should be 0,8ms as I've read on the forum.
The req fuel number is 6 at the moment. I'm idling with a VE number 30 at 900rpm, 25kpa, 13,5 AFR(batch fire)
I would like to have the VE numbers a little higher, so when going up or down the afr would change in smaller steps


When I change my injector dead time from 1,2ms to 0,8ms. with what number do I have to multiply my entire VE table? (if it's that simple?)


Thanks

change req_fuel or ve% until the AFRs goes back to where they were.

youre going to ultimately end up adding more fuel, since you're telling the MS there's less deadtime and more time fueling.

So when i select my whole VE table and add a certain percentage until my idle is back at 13,5 AFR anything should be fine?

correct.

When I switched to ev14's and had to adjust deadtime/reqfuel/ve everything was off enough that I had to re-do the ve table. Maybe you won't have to since you're going ev6->ev6 but I think you probably will have to as well

I hope Braineack' s right...


thanks for the info guys

I think he might have misunderstood the question, or I missed something

Dead time should be measured for your injectors, I was using .9 on rx8s. The dead time correction value will help the car hold VE better when you've got electrical loads going. You are likely going to need to retune the whole map by changing dead time, I do not think the ratio is even across the entire VE map in other words.

To get more resolution you need to increase the VE numbers, to do that you need to reduce the required fuel. If you idle at 50 with a req fuel of 5.0 your new VE at idle would be 100 with a req fuel of 2.5. You now have double the resolution, basically.

Changing dead time will alter actual fueling more at low loads than high loads. Expect to see yourself adding more to the VE table at idle than at 100+ kPa.

It's like this: You are removing 0.4mS from everything in the fueling equation (1.2 to 0.8mS change). If the original pulse width at idle was 1.8mS (total called for), you will now need the same amount, but, after the dead time adjustment, and before the VE table adjustment, you will suddenly be running 1.4mS total. A huge percentage (-22% change)

At high loads, you may have been more like 13mS total. That will fall to 12.6mS, so not nearly as much percentage change, and therefore a small change in VE to accomodate (-3% change).

And yes, I think the 0.8mS is a better number.

Would the 0,8ms improve smooth idle or better operating engine?

I mean, is it worth to redo the VE mapping or should I leave it like it is at 1,2ms

if you're asking if correct dead time matters, the answer is yes, yes it does. you will have weird inconsistencies with it off

ok, Autotune will do the trick I hope.

Thanks

so if you're trying to make it idle more smoothly, changing the dead time isn't directly going to affect that. It's just part of the pulsewidth calculation.

You've got a few knobs to turn, so to speak. You can richen the AFR. Try that first and see what it does. You can change spark timing and airflow (example: increase IAC duty but retard spark to make the same idle speed). Since you are batch fired, you can't change injection timing so that's not an option.

Correct

And Correct

Well, I'm going to correct the dead time as it should be and redo the entire VE map. Thats for sure now.

With the Mazda ECU you could place a glass of water on top of the engine, the water and glass would barely move...

I don't know it if it's possible to get the same smooth idle as the stock ECU does.
At the moment it's idling at 900rpm (+/- 30rpm) AFR 13,5 (tried 13AFR with no changes) and 25kpa vacuum. Tried 10 to 16° advance (no changes)
The idle is certainly not bad and I've seen lot worse but still think it can do better...

Ik keeps idling with some vibrations that you even feel when sitting in the car.These were not present before the MS2 conversion with the RX8 injectors.

Is there any room for improvement or should I take peace with it like this?
Any other settings that I can try to change?



Thanks

huh? MS can do at least as good, and often times better idle than the mazda ecu. you're just not there yet.

perfect post. PM your address to receive a cookie.

idle should be just as smooth, deadtimes wouldn't case that.


are you sure youre running on all four?

:rofl: yes, I'm sure, it's not THAT bad. Engine is in perfect working order.

I agree, but I'm wondering what part of my tune needs improvement to get there?

This will sound corny, but: it's really about creating harmony between ALL the settings to get a good idle. Then it's about smoothing everything over and dialing in enough correction and buffer zone such that extremes (high power load, cold temps, ac kicking in, fans kicking in, etc) don't disrupt the idle and are compensated for timely and efficiently.

You need proper afr
You need proper timing
You need to pay attention to the cells AROUND both those areas too, they affect it
You need to do idle valve test to determine iacv range
You need to tune CL idle control settings, PID, and dashpots
You need proper AIT compensation
You need proper EGO correction settings
There's probably more I'm forgetting

A lot of this stuff gets overlooked

You're right about that, but I think most of that's covered. I've attached my tune a few posts earlier.
rpm is stable enough in my opinion but it's the vibration I want to get rid of...

I will post a log when idling, maybe you guys notice something wrong.

Thanks

idle searching is very often caused by large variations between affected and surrounding cells. needs moar smoothing (but that's just a guess, I can't open your MSQ from work computer)

here you are

My experience with the 1.6 is that it's sensitive to injection timing at idle. You can smooth it out by changing end of injection timing in the idle area if you are sequential. Since you are batch fired you don't have that option.

<p></p><p>why arent you using idle VE table?</p>

I would like to, but can't activate it right now I think...


Is this possible with MS2 firmw 3.3.3?

I thought that was ms3 only

You are correct, no idle VE cells outside the normal VE map in MS2.

I don't think 3.3.3 adds that feature either, but as a 3.3.2 user I'd sure like confirmation.

<p>no i was thinking ms3.</p>

too bad, it's a nice feature in my opinion

problem solved, bought a Link G4+ :rofl:

The fueling calculation is as follows:

PW = DT + (ReqFuel * MAP * VE[RPM,MAP] * AirDen * BaroCor * corrections)

Forget the corrections and we get:
PW = DT + (ReqFuel * MAP * VE)

Rearange the formula to find VE:
VE = (PW - DT)/(ReqFuel*MAP)

Assume we have 265cc injectors and a deadtime of 0.8ms, a required fuel value of 5.51 and want to find the PW at 80kPa with a VE of 55%:

PW = DT + (ReqFuel * MAP * VE)
PW = 0.8 + (5.51 * 0.80 * 0.55)
PW = 3.2244 m/s

We can rearrange to find VE as below:

VE = (PW - DT)/(ReqFuel*MAP)
VE = (3.2244 - 0.8) / 5.51 * 0.80)
VE = 0.55

Re scaling for new injectors example:
We upgrade to blue 540cc injectors, an assume a Required Fuel value of 2.25, and a dead time of 0.95

If we want to re-scale the VE table for new injectors we first solve to find the current PW (3.2244m/s in our example).
Then find the correction factor for the new injector. In our example using 265cc and 540cc injectors:

Correction factor = Old Injector cc min / New Injector cc min
Correction factor = 265 / 540
Correction factor = 0.4907

Applying this correction factor we find the new pulse width:

New PW = Current PW * Correction Factor
New PW = 1.58 m/s

Then we calculate the NEW VE value to populate our VE table as below:

New VE = (New PW - New DT)/(New ReqFuel * MAP)
NEW VE = (1.58 - 0.95) / (2.25 * 0.80)
New VE = 0.35 or 35%

Resurrecting an old thread.

Some more stuff. I have too much time on my hands at present!

I have had the wrong dead time in my MSQ for nearly 3 years. RX8 injectors. I had dead time set at 1mS. When I changed it to a "better" value of .8mS idle would go stupid lean. Not wanting to retune the whole VE, I had until now decided to just leave it at 1mS and live with it.

BUT...

You can use this spreadsheet that I've designed to recalibrate the VE table. I tried it, seems to work fine at idle, (the worst scenario, low pulse width). I still have to VEAL higher Pulse widths, but I think they should be pretty close hopefully. Use at your own risk, and if you don't understand why you might need it then don't use it.

https://drive.google.com/open?id=1lTiwZ ... uT4rBK2EhD

Some more stuff. I have too much time on my hands at present!

I have had the wrong dead time in my MSQ for nearly 3 years. RX8 injectors. I had dead time set at 1mS. When I changed it to a "better" value of .8mS idle would go stupid lean. Not wanting to retune the whole VE, I had until now decided to just leave it at 1mS and live with it.

BUT...

You can use this spreadsheet that I've designed to recalibrate the VE table. I tried it, seems to work fine at idle, (the worst scenario, low pulse width). I still have to VEAL higher Pulse widths, but I think they should be pretty close hopefully. Use at your own risk, and if you don't understand why you might need it then don't use it.

https://drive.google.com/open?id=1lTiwZ ... uT4rBK2EhD