BWilcox

Hi All,

7730 ECM : 420 CID : Medium Sized Cam : MiniRam : NA.

The BIN is based off of AUJP. Although other parameters (VE, SA) have been modified, the Cranking Parameters have not be touched.

I am presently fighting an issue where a hot (ECT = 180 F) start is very quick (~ 1 s) and a cold (ECT=50 F) start is somewhat slow (~ 8 s).

When the engine finally cold starts, it runs fine. No initial stalling.

The $8D has three tables to adjust the cranking PW.

It appears like the cranking PW is not a function of the MAP. Is this true?

Is there a "best practices" methodology to tuning the cold start?

Best Regards,

Bruce

Orr89RocZ

iTrader: (20)

I play with the coolant temp based crank starts. Crank pw vs coolant. Colder generally needs more fuel than hot

Tough to tune cuz you only get one shot at testin changes per day basically.

JP86SS

iTrader: (1)

seems that a combination of the crank fuel tables vs temp and the IAC park position (hot and cold) help allot.
Park position can have a big effect on how easy it cranks too.

BWilcox

Thanks Guys!

Not sure if it is lean or rich during cranking?

I think I will pull some fuel initially and see what happens. That way I will not flood the motor.

Orr89RocZ

iTrader: (20)

I found that if it cranks and starts with alittle throttle input then its too rich and needed the extra air to fire it up. Lower fuel and maybe increase park steps

BWilcox

Hi Guys,

After much research and RBob's help, I have the Injector Constant and Fuel Pressure correct. Had to increase the Injector Constant and lower the Fuel Pressure. The end result is less fuel will be sent to the motor at a given PW.

In addition, I am making progress at tuning the Part Throttle VE.

Now the Cold Start manners are far worse. The conclusion is the cranking PW is too small.

Is my assumption correct that with the $8D, the Cranking Pulse Width is only determined by tables. The MAP is not utilized until "run" mode?

Best Regards,

Bruce

RBob

iTrader: (1)

The MAP is used to get a barometric read, there is a cranking PW table versus barometric.

If the cranking was OK and just need to scale it for a different fuel injector flow rate change this value:

RBob.

BWilcox

Hi Guys,

Thanks for the suggestions to date.

As I understand the "starting" order of events, the following happens:

1.
Key On a Baro read.

2.
With the Starter running,,,,

The following tables are utilized:

"Crank Fuel PW Mult vs. Baro".

"Crank PW Mult vs. ECT".

"Crank PW Mult vs. TPS".

"Crank Fuel Delivery Delay vs. MAT".

"Crank Fuel PW Mult vs. Reference Pulses".

I am guessing that with the initial start, the Cranking PW is corrected for Baro, ETC, and TPS. Then a delay based on MAT is implemented. Then the fuel starts to flow through the injectors based on the reference pulse count vs. time.

In addition, the VA tables are not utilized during Cranking.

Correct?

What is the purpose for the "Crank Fuel Delivery Delay vs. MAT" table?

The data in the "Crank Fuel PW Mult vs. Reference Pulses" is interesting. Has this been developed to achieve a rich enough mixture with flooding the engine?

Rbob,

Hot start has not been a problem with any of my tunes. The engine fires up well within 1 second of the starter turning.

The problem has always been the cold start. With the old (wrong) Injector Constant and Fuel Pressure setup, cold start (50 F) usually required 6 seconds for the engine to fire up. After I updated the Injector Constant (raised) to reflect the actual Injector Flow rate and set the fuel pressure (lowered) to 44 PSI, cold start is significantly harder / longer. The engine is leaner now.

Seems like I do not have enough fuel to cold start within a reasonable ( ~ 2 seconds) time frame. Thinking the table that I need to modify is the "Crank PW Mult vs. ECT".

My guess is I made the cold start cranking worse because I changed the Fuel Pressure from 48 PSI to 45 PSI. This results in a -3.2% fuel flow rate per given PW. Based on the original cold start times, I was already lean. It would make sense that one would expect even more troubles with cold start.

Never tuned the Cold Start PW before. Thinking that a 3.2% PW increase may not be good enough?

Best Regards,

Bruce

BWilcox

Hi Guys,

I looked at the "Crank Fuel PW Mult vs. Reference Pulses" tables for the following BINS (AXYC, AXYD, AUJP, AXCN).

The AUJP is much different when compared to the others. My modified BIN is based off of the AUJP.

Given AUJP, 4 pulses are seen over 24 DRP.

Given all other BINS, the multiplier is non zero everywhere.

Since I am fighting a cold start (only) long cranking time ( ~ 10 s ), wondering if a non AUJP set-up would be better?

Alternatively, would it be betting to stick with the AUJP set-up and just change the "Crank PW Mult vs. ECT" table.

During the cranking, nothing happens for ~ 5 seconds. At about 7 seconds the engine tries to start. At 10 s the engine starts and runs fine. This problem becomes worse as the FP is lowered.

Seems like the engine is too lean to start quickly.

The voltage @ the ECM is around 10 Vdc during cranking.

A ~ 5 seconds of cranking, oil pressure is indicated.

Have not yet checked the Fuel Pump Voltage during cranking. Need to make sure the Fuel Pump Relay is ok. My guess is it is ok. Since Hot Start is very quick.

Help is appreicated.

Best Regards,

Bruce

Orr89RocZ

iTrader: (20)

I would try leaving the ref pulse table alone. It was popular to take the bins that had no reference pulses below like 10 and make them 0.50 or so down the table to quicken startup. Worked for my maf car back in the day

I would just try adding few % crank fuel vs ect in the cold temp areas to see how it acts

A 420 needs alot more fuel to start than a 350 tpi

84Elky

This discussion about Cranking PW peaked my interest so I decided to simulate the $8d code to see what happened if certain variables were changed.

With all other variables remaining constant:

• Cranking PW is directly proportional to the Cranking Scalar at 0x3AD RBob mentioned in Post #7. If doubled to 400ms from the Factory GM default of 200ms, Cranking PW doubles. Reduce it by half to 100ms and Cranking PW is reduced by half
• The lower the MAP, the lower the Cranking PW. At approximately sea level to 1000 ft, MAP is approx. 100 kPa. At 5000 ft, it's approx. 84 kPa. So Cranking PW decreases with altitude.
• At sea level to 1000 ft, with TPS closed and with Coolant = 71*F, PW = 14.8ms.
• Coolant Temp is used during cranking or at any other time unless there is a coolant sensor voltage error detected. In that instance MAT is used. If a coolant sensor error arises while engine running, a default coolant temp of 145*F is forced for fueling decisions and the CEL is set on.
• IF TPS % is >= 87.5%, Cranking PW = 0 (Clear Flood Mode)

So, what this says is:

• To increase Cranking PW at all temps, increase the scalar at 0x3AD
• To increase Cranking PW at selective temps, increase values in the Table at 0x3AF (Cranking PW .vs. Temp)
• To decrease Cranking PW at selective temps, decrease values in the Table at 0x3AF (Cranking PW .vs. Temp)

A spreadsheet with simulation test results is attached.

BWilcox

Thanks 84Elky,

It is great that you have a bench / ability to dig into the inner workings of the ECM.

For me, I have enough time to mechanically build the engines and then tune them. Understanding the code is just out of reach.

I looked at the stock $8D AUJP tables.

The "Cranking PW vs. ECT" Table indicates an inverse relationship. As the ECT increases, the PW decreases. From a "choke" function perspective, this makes sense to me.

Your simulation data indicates the exact opposite function. Are the published Hacks wrong? Certainly, this is a possibility.

Do not understand the simulation results regarding the Baro. Given a goal of "X" AFR during cranking, as the altitude increases (MAP decreases), the cranking PW would decrease. Does my intuition make sense?

The only table that appeared to use the MAT was the "Cranking Fuel Delivery Delay vs. MAT". What is the purpose of this table?

Best Regards,

Bruce

Orr89RocZ

iTrader: (20)

In my experience with code 59, the colder temps have higher values. Hotter ect will show smaller values.

I'd have to double check 8D again but thought it was similar

84Elky

No bench -- wish I had one -- that's a project for later. Just simulating assembly code segments to get results under varying conditions.

Bruce, thanks for your detailed analysis. In the spreadsheet, a combination of swapped comments on correct data and incorrectly computing the percentage change in Cranking PW based on MAP made a mess of things and I apologize. Just did not apply common sense to the analysis results before posting. This confirms the need for me to stop doing this stuff late at night.

Not sure if you are referring to my post text or the spreadsheet. The spreadsheet shows that as Inverted MAT increases, PW decreases as it should, but the post incorrectly states the opposite.
Also, the initial spreadsheet posted did not display the comments at the top because panes were frozen. Sorry. Take a look at the comment about the MAT temp (L0061) which I believe you are referencing. In summary, it explains that the value shown in the spreadsheet for L0061=MAT is an inverted value. So we have an inverted MAT value (where strangely, a higher inverted value = higher temp) being used to access an inverted table. This results in lower PW being extracted from the Cranking PW .vs. Temp table as Inverted MAT increases, and thus calculated Cranking PW decreases. Make sense? If not, please let me know.
With your comments prompting further review, found that I incorrectly stated that a 25% increase in values in the PW .vs. Temp Table at 0x3af resulted in a decrease in Cranking PW. Not so. Larger table values = larger calculated Cranking PW. Correct spreadsheet data, but wrong comments in spreadsheet and thus wrongly posted. Grrr!

Not sure about the goal of a given AFR, but again you are correct that increased MAP results in increased PW. Common sense! As above, the spreadsheet data are correct but I somehow reversed the comments.
The spreadsheet has been reposted -- same data, but correct analysis. In addition, my original post has been corrected to provide correct information.

A couple of things:
(1) If I'm interpreting the code correctly, the default temp used for fuel calculations is always Coolant. But that is changed to MAT if there is a coolant sensor error.
(2) The table you referenced is part of some very complicated logic which appears to be used to reduce/delay the calculated Cranking PW based on the DRP count during engine cranking. Will take a look at the details and reply if can figure it out.

Let me know if anything still does not make sense. Thanks again or your insight.

84Elky

Apologies, but the text of the original post was changed to the following to correct an error in cranking duration.

Here's how cranking fuel delay and fuel provision works using the 0x39C Table=CRANK FUEL DELIVERY DELAY .vs. MAT and the PW Factoring Tables at 0x3C7 and 0x3D7:

• The Table at 0x39C provides a DRP delay count based on MAT. The count range is 0 to 8 in the $8d Factory Calibration, generally with lower counts at lower MATs. The Table DRP count returned is the number of cranking DRPs required before fuel will be allowed (defines the fuel delay time).
  For example, if cranking at 400 RPM, it takes 37.5ms before 1 DRP is generated. At a MAT in the range 69-89*F (20-32*C), a value of 6 will be extracted from the delay Table. Together, this will result in a delay of 225ms before cranking fuel will be provided (37.5ms per DRP * 6 DRP delay = 225ms). I assume this delay is to allow for some engine movement before slogging in fuel.. What is not understood are the 0x39C Table values for -4 and 8*C being 0, when lower and higher MATs have a non-zero count. This is surely something the GM engineers incorporated through testing.
• After the DRP delay in this example, for each subsequent DRP the computed fuel PW is adjusted by a factor extracted from the Table at 0x3C7=Crank PW Multiplier .vs. Ref Pulses 1-24. The Factory factors are: 0=No fuel; 0.5=50% calculated fuel; 0.8=80% and 1.0=100%. The table extraction and multiplication process continues during cranking for 825ms until the engine starts or until the first 16 Table elements have been accessed (825ms = 225ms fuel delay + 600ms to fully access elements 1-16 of the 0x3C7 Table).
• When 16 table elements have been accessed without starting (16 DRPs detected after the delay DRPs and now cranking for 825ms at 400 RPM in this example), the Extended PW factors are used in elements 17-24 of the Table in exactly the same manner as the first 16 elements. These Extended PW Factors in elements 17-24 are used until the engine starts, or until the 24th element has been accessed without starting, whereupon factor extraction starts again, cycling through elements 17-24 until engine start.

BWilcox

Thanks 84Elky,

I appreciate the time it takes to dig into the GM code.

I am still wondering why GM would scale the Cranking PW on MAT? Seems like it would make more sense to scale the Cranking PW on ECT.

In my case, the ambient temperature during testing is around 55 F. Cold start or warm start. Long cranking times are only seen when the ECT is around 55 F. When the ECT is around 180, the cranking time is normal.

I increased the cold cranking PW and made the situation worse. Had to open the throttle blades to start the car.

Thinking about three variables:

Cranking PW.

IAC Position.

Base Timing.

I think the present Base Timing is around +6 degrees BTDC. Wondering if this is ok for cold start?

Does a non stock cam require more base timing to aide start up time?

Wondering if a non stock cam requires less Cranking PW? Similar to the VE tuning situation at low RPM/MAP areas. Maybe the stock Cranking PW vs. ECT table is too steep over temperature?

Thanks again for the help.

Best Regards,

Bruce

Orr89RocZ

iTrader: (20)

Colder temps will like more timing. If you added fuel and made it worse perhaps you added to much? Its pretty sensitive. But additional iac tuning and spark adder vs coolant temp will help

BWilcox

Hi Guys,

Happy Holidays!

Thanks for all of the help on this issue.

Orr89RocZ, you mentioned that the Cranking PW adjustment is sensitive. What is a recommended Cranking PW change amount? 5% ? 10% ? 20% ?

During cranking, the SA is just the base timing that is set by the Distributor. Correct?

Once the car fires, it runs fine during warm up. It is just the Cold Start that is driving me nuts.

Best Regards,

Bruce

Orr89RocZ

iTrader: (20)

I usually tried 2-5% at a time. Timing adders may be very helpful as well, which is something i did not try as much on my old setup

Also note my injectors were hugggge compared to most so maybe thats why its sensitive

BWilcox

Thanks Orr,

I have played around with the Base-Timing (6 BTDC to 10 BTDC). This is did not significantly modulate the Cold Start cranking time. I could not find any Cranking Mode SA parameters in the $8D hack? I thought that the Base-Timing (from Distributor) is in play until 400 RPM? Then the ECM takes over. Is my understanding correct?

In addition, I have run some more Cranking Pulse Width tests and tracked the Cold Start fire up time.

Before I get into the data analysis, I measured the Fuel Pump Voltage at the Fuel Pump Relay (Pump Side). It is ~ 10 VDC during cranking. The Fuel Pump does prime (good pressure) when the ignition is in the Run mode. Therefore, I should have fuel available to the injectors.

The attached graph represents 4 set-up conditions. Unless indicated otherwise, I have not changed the Throttle Body position or IAC programming.

DataMaster does not indicate the actual Cranking PW. I had to interpolate the Cranking PW vs. ETC Table to calculate the number.

Green : V6 BIN.
After 10 seconds of cranking, the engine sounded like it was on the verge of starting. I shut down the starter at this point and a back fire through the intake was heard. On the second Cold Start attempt, the engine fires up after 6 seconds of cranking. No problems with maintaining a stable idle after the fire up.

ECT/MAT @ Start-Up ~ 56 Degrees F.

Cranking PW ~6.0 mS.

Black : V9 BIN.
The engine starts after 14 seconds of cranking. No problems with maintaining a stable idle after the fire up.

A little bit of Black smoke is seen when the engine fires up.

ECT/MAT @ Start-Up ~ 57 Degrees F.

Cranking PW ~6.8 mS.

Red : V8 BIN.
The engine started after 23 seconds of cranking. Had to open the Throttle Body a little bit to get the engine to start.

ECT/MAT @ Start-Up ~ 57 Degrees F.

Cranking PW ~7.3 mS.

Blue : V7 BIN.
The engine sounds like it will start after 7 seconds and finally starts after 10 seconds.

ECT/MAT @ Start-Up ~ 62 Degrees F.

Cranking PW ~6.5 mS.

Conclusions: Do you guys agree?

1.
V8 BIN is too rich.

2.
A Cranking PW between 6.0 mS and 6.8 mS may be required when the temperature ECT/MAT is in the high 50 range.

As Orr suggested, the Cranking PW may be very sensitive.

Of course, I am assuming that my car should start Cold as quickly as Hot. Relative to the engine displacement (420), the Cam (241/241 @ 0.050) does not seem big to me. I consider this Cam a moderate size. Comments?

The Oil Pressure comes up at around 5 seconds of Cranking. So my guess is the best Cold start Cranking time may be limited by the Oil Pressure build up time. Make sense?

Best Regards,

Bruce

BWilcox

Hi Guys,


The Weather has been bad so I have not attempted additional Cold Start tests.


However, I looked over some old Data Master logs and found something interesting.

Before I started tuning the VE Tables, the car started within 2 seconds @ 70 F.

After I started tuning the VE Tables, the car started in 6 seconds @ 65 F.


During the starting process, the MAP transitions through the following chain of events.

MAP RPM
100 <= 400
90 800
80 1200
60-50 1600


The Lower VE Table update is the only BIN change between Cold Start trials.


The updated Lower VE Table was set up for lower efficiency at lower RPM values.

Does this make sense?

Best Regards,

Bruce

RBob

iTrader: (1)

Is the VE table used during cranking?

RBob.

BWilcox

Hi RBob,

Based on information from other folks, the lower VE table is not used during cranking.

Just found it interesting that I leaned the lower VE table and the cold start time increased.

I was wondering if I need to richen the lower VE table where the engine transitions between cranking mode and run mode?

Best Regards,

Bruce

84Elky

RBob - Quick review of $8d indicates the VE tables are not used for cranking -- only used once the engine has been declared running.

It appears the base cranking PW comes from the table at 0x3af (in uSec, later converted to ms) which is further massaged by several multipliers (TPS%, Clear flood or not, etc.) to arrive at the final cranking PW.

Bruce - Just a guess, but I'd venture that the transition between cranking and running is so quick that the values in the VE table during that transition are somewhat irrelevant provided that VE at idle is correct. But just a guess.

BWilcox

Thanks Guys,

The idle and part throttle sections of the VE Tables seem reasonable with respect to the BLM correction.

However the plugs indicate a rich mixture. My guess is the idle mixture is too rich due to the cam overlap. A false lean condition at idle.

The next round tuning will include massaging the Cranking PW, IAC Park Position, and the NBO2 voltage window (idle).

Best Regards,

Bruce

RBob

iTrader: (1)

This was a rhetorical question...

RBob.

BWilcox

Hi Guys,

I finally solved the cold start problem!

In an indirect way. Sometimes, it is better to be lucky than good.

My 7 year old MSD 6226 coil failed. After reading a many comments on the MSD coil quality, I replaced this coil with an Accel 140011. I am still using the MSD Digital 6A.

Cold start is now just like a new car.

Thanks for all of the help.

Lesson learned,,,,, when tuning does not help, look for hardware root causes.

The low RPM lower MAP VE table now requires some updating. Need to add some fuel. Guess the motor is more efficient now. I was tuning with a less than ideal coil.

Best Regards,

Bruce