LS2/LS7 coils and DWELL Time

[Nuyskov]

Hello!
Yesterday finished LS7 coils swap into my RB25DET NEO.
And have a question - it's recommended to make DWELL time about 5.5ms, but how can I make it?!
I see two tables: Dwell duty cycle and Dwell time.
But I can change only percentages...
Tell me, please, what should I change to change DWELL time?
Some of aftermarket ECUs have option "change coils" like Nistune has "change MAF"...
Matt, is it available to do something like this? To change DWELL tables automatically with changing coils type?
Thank's!

[skylinegtrhr]

On SR there is table don't know for RB NEO , but before You make adjustment I strongly advise You to read this:

http://efi101.com/forum/viewtopic.php?t=7894&highlight=

[Nuyskov]

Thank's!
I've done it - it's easier that I thought... I've changed all the DWELL times to 3.0ms and the car drives very well!
I have another problem - I always have 21 Error (Primary Ignition) due to aftermarket coils - what should I do to fix this error?

[Nuyskov]

Matt, tell me, please, I've found information about changing ECU for LS7 coils.
I've found in NisTune that we can change 2 tables, but there we can change percents from voltage and milliseconds from RPMs...

I've looked throught constants but haven't found such parameters...

Do we have them?

@@@
To get 5.0 milliseconds of running dwell, the nominal dwell parameters should
be set to:

Maximum Dwell 5.5 milliseconds
Maximum Spark Duration 2.0 milliseconds
Acceleration Compensation 0.5 milliseconds

Battery Voltage Compensation
Setting Net Voltage Dwell Compensation
-4.0 8.0 Volts 2.4 milliseconds
-2.0 10.0 Volts 0.9 milliseconds
0.0 12.0 Volts 0.0 milliseconds
2.0 14.0 Volts -0.5 milliseconds
4.0 16.0 Volts -0.9 milliseconds

This will give 5.5 - 0.5 = 5.0 milliseconds at 14.0 volts while running with the
alternator charging normally.
@@@

[Nuyskov]

Today I've downloaded AEMTuner and seen difference between DWELL tables. Opened calibration with R33 RB25DET, changed coils to LS2 and the DWELL at 14V was 3.3ms, changed coils to R34 OEM and saw 1.6ms. We know, that LS2 coils need to be adjusted to 5ms, so I've found that in AEM our stock DWELL at 14V is set as 1,6*5/3.3=2.42ms. And to receive 5.0ms, we need to scale DWELL vc Battery percents in 5/2,42=2,077 times.
I've scaled in 2.5 times, but tomorrow I'll decrease scaler, but now the car goes perfect at high revs!!!

There are two tables in Nistune - one with Percents from Voltage, and the second is Mss from RPMS.
AEM has such tables but Percents from RPMS and Mss from Voltage.
To get to know exact DWELL time we need to multiply Mss (from RPMS - about 5.0ms for RB25DET) * %s (from Voltage) * something... What is this something?

[Matt]

Address definitions:

DWELL_TIME,&H7B50,16,1,16,1,Coil time
DWELL_DUTY,&H7B60,32,1,32,1,Coil duty

Japanese interpretation:

7B50 TDWELL Voltage correction coefficient 1×16 VB:1.26V 100/51%
7B60 DWLDUT Electrification dwell duty cycle 1×32 HNREPM:400rpm 1/3%

Nistune:
Dwell time [Volts vs Percent]
- Actually displayed as 255/100% to scale this table properly. This appears to match the AEM percentage base also
- VB has a voltage offset of 1.26V. Whilst this is applied to the table internally the cells still show as 1-16V after doing some testing on Z32 with different battery voltages on the bench using HW maptracing making this scaling 2-17V makes the table accurate so I've updated for next release

Dwell duty [Volts vs Percent]
- Done some more investigation on this today. This table is not fixed 'ms' adjustment as such as the amount of ms adjusted is consistent when adjusted on a fixed RPM but varies when RPMs change
- So you are not going to have a consistent ms vs RPM table like the AEM has. It will vary:

For example

RAW / ms @ scope [3600rpm]
10 0.5
20 0.75
30 1
40 1.25
50 1.5
60 1.75
70 2
80 2.25
90 2.75
100 3
110 3.25
120 3.5
130 3.75
140 4
150 4.25

RAW / ms @ scope [5600rpm]
10 0.3
20 0.5
30 0.7
40 0.9
50 1
60 1.2
70 1.3
80 1.5
90 1.8
100 1.9
110 2

- The current ms adjustment shown in the table is based from a calculation

ms = % * 10/(rpm/60

based on some feedback I was given. However I dont find this to be consistent with monitoring the signals on the scope so will put the cells back to '%' to avoid confusion

[Nuyskov]

Hello, Matt!
Thank's for such answer!
I didn't change DWELL Time table.
Changed only DWELL Duty.
From your post, I've understood, in Nistune there is no constant DWELL Time in ms. I've scaled DWELL Duty in 2 times, tried to find information about std dwell time for RB25DET, found 1,6 - 2,0 ms at several resources... So, I've received 3,2 - 4,0 ms... But anyway I still have some pushes with boost, and I'm going to try to scale stock DWELL Duty in 2,5 times...
Is it right?

[Matt]

The 'ms' time appears to change based on amount of RPMs that the engine is currently doing. So you have to adjust per amount of RPMs

[Nuyskov]

- So you are not going to have a consistent ms vs RPM table like the AEM has. It will vary:

For example

RAW / ms @ scope [3600rpm]
10 0.5
20 0.75
30 1
40 1.25
50 1.5
60 1.75
70 2
80 2.25
90 2.75
100 3
110 3.25
120 3.5
130 3.75
140 4
150 4.25

RAW / ms @ scope [5600rpm]
10 0.3
20 0.5
30 0.7
40 0.9
50 1
60 1.2
70 1.3
80 1.5
90 1.8
100 1.9
110 2

- The current ms adjustment shown in the table is based from a calculation

ms = % * 10/(rpm/60

based on some feedback I was given. However I dont find this to be consistent with monitoring the signals on the scope so will put the cells back to '%' to avoid confusion

Hello, Matt!
As I understand, these two raws you gave are mseconds from "something (10, 20, 30...)" during consistent RPM. What is this "something"?

We have two tables - %s vs Voltage and %s vs RPM.
%s vs Voltage is about 17% at 14.0V (meaning offset).
And when we calculate ms with your calculation we see numbers (the same as in the table), but we have about 17% full time scaler at 14.0V, so real DWELL time should be 5.25% * 0,17ms = 0,89ms?! Or I missed something?

So, I mean, can we calculate real DWELL Time in ms with these two percentages (from Voltage and RPMs)?

[Matt]

The 'something' values are the actual 'raw' numbers in the table (the bottom row)

The ECU looks at the current RPM column in the table, takes that number and works out 'ms' based on it.

We may be able to calculate a real DWELL time using the second table but I'm unsure how it could be calculated exactly

[Nuyskov]

The 'something' values are the actual 'raw' numbers in the table (the bottom row)

The ECU looks at the current RPM column in the table, takes that number and works out 'ms' based on it.

We may be able to calculate a real DWELL time using the second table but I'm unsure how it could be calculated exactly

Thank's, Matt!
And the first table (% vs Voltage) - is it a multiplier of an offset?

I've found such discussion... May be this help?
http://www.skylinesaustralia.com/forums ... ty__st__20
There is an evaluation of finding real ms, but there was nobody to test it with oscilloscope... Mat, you have oscilloscope measures - can you test, if that evaluation is good?
Thank's!

[Matt]

And the first table (% vs Voltage) - is it a multiplier of an offset?

Yes this would be an offset. But this first table is just to add extra % for when the battery voltage is < 14 volts. Basically allow a longer charge time when the battery voltage is lower. You should not really modify that table

The second table % vs RPM is another adjustment with amount of % added based on current RPM.

I tested with a scope and posted here the evaluation at two different RPM readings. The ms time adjustment varies based on the RPM. I would have to calcualate on each RPM range to try and work out the ms adjustment for each one

[Nuyskov]

Hello, Matt!
Today I've made several measures of Dwell at different RPMs and %s... (measured primary ignition circuit with oscilloscope with inductive sensor).
Compiled them in Excel file and tried to understand relations between parameters.

The first column is RPM
The second is a point from "Dwell Time" table - % vs Voltage
The third is a point from "Dwell Duty Cycle" table - ms vs Voltage
The fourth is back calculated (with your equation upper) % vs RPM
The fifth is measured Dwell Time in ms.

1. I suppose that relations between real dwell time and %s are not linear, may be there are some more parameters in equations except %s, Voltage and RPM...
2. Dwell vs Voltage is an offset and this offset decrease with high RPMS... Also I think it is not so high as I thought and I think it's near or less that measurements errors...
3. Dwell vs RPM looks like some logarithmic function, because the relation starts to be linear and then is filled up to the right...

Now I'm going to set my DWELL to 5.0ms just with oscilloscope at all the RPMs... And later when I have time, I'll try to find mathematical dependence Dwell Time in ms from %s...

Anyway hope these measurements will help you...

[Nuyskov]

Hi, Matt!
Yesterday I've made a mistake - when I saw influence %vsVoltage on Dwell Time I changed only one point at a time - and I had to change all the plot...

Today I backed up both DWELLvsVoltage and DWELLvsRPM tables and started to correct DWELLvsRPM table... I had to scale all the points x2.5 (from 5ms to 12,5ms) to receive 5ms on the scope, so your 5ms (with your equation in Nistune) was 2ms on the scope.
It worked up to up to 4000RPM (old value 100 vs new value 250), but at 4400RPM I had to set value from 110 to 275 and there is 255 max...

So I've backed up DWELLvsRPM table and decided to work with DWELLvsVoltage table once again. This time I've scaled all the points, so it is a SCALER!!! Not the just offset. With OEM data Dwell Time was 2ms, x2 = 4ms, x3 = huge 6ms.

I've got to know one more thing - if we set all the cells to 5ms for example with scaler x3 (to DWELLvsVoltage table) we receive about 6ms at low RMP, about 5ms at mid RPM (3000-4000) and about 3,2ms at high RPMs.

So with scaler x3 I've decreased DWELLvsRPM table to 4,5ms at low RPMs, at mid RPMs I left 5,0ms and today later I'll increase high RPMs cells to get 5,0ms and tell you...

[Matt]

Okay thanks for the clarification. I should have tested the dwell time vs battery again before stating my thoughts on being an offset... so if this was set to 0% then there would be no dwell effectively

Not sure with setting all the dwell vs RPM cells to the same raw value how that would turn out without scoping it