Hard starting TPI motor - timing off?

New to fuel injection and I could use some advice.

I have a '40 Ford streetrod with a TPI style chevy motor that I don't know much about.  The car runs great once its running.  It starts really hard, with clanging noises that sound to me like the initial timing is too advanced and is kicking back on the starter.  I have no idea how to diagnose this, or time it.

Here are some specifics (or at least what I can tell or have been told):

The short block is a ZZ4 crate.
The distributor is in the back of the motor (not opti-spark style)
The throttle body is an 85-88 F-body style.
The ECM is a 90-91 speed density style tuned port fuel injection.
It has a Howell chip (not the stock chip).
The starter is new.
There are no broken teeth on the flywheel (yet...)

Can anyone tell me what the initial timing should be for this motor, and how to set it?  I know how to time carburated cars, but I've never timed a fuel injected motor before.  From what I understand, there's an electrical connector by the ECM which I disconnect (it has a tag on it), then I can just turn the distributor and put a light on the balancer like normal.  Somebody told me the initial timing should be 0 - is that right?  

I have very little to none experience with TPI's but this was posted on sister magazine's High Performance Pontiac forum by a very knowledgeable member there:

barneyformula

Posted: 01/17/09
04:42 AM


It's been a long time since I set a TPI TA, but the procedure is simple enough. Run the engine until it is warm, shut it off and locate the timing connector. It is a tan colored wire with a black stripe that has an inline plug in it, usually hanging out of the harness near the distributor or along the firewall near the heater box. Unplug the connector and restart the engine, the check enine light will be on, hook up the timing light as you would on any non-efi V8 with the inductive pickup on #1 plug wire (front cylinder, driver's side). Check your emmision/tuneup tag for the base timing setting, I believe it is 0 degrees, but check to be sure. Loosen the distributor hold down clamp and turn the distributor to adjust the timing to spec. On most small block chevy engines, the largest notch in the timing tag is "0", so just line up the line on the balancer with the big notch. Tighten the distributor hold down clamp and recheck your setting to be sure it didn't change, shut off the engine, reconnect the timing wire and disconnect the negative battery cable for a minute to clear the code and turn off the check engine light. Restart the engine to make sure the light is out. Depending on the octane level of the fuel you use, sometimes you can experiment with slightly more base timing, you can try advancing it 2 degrees at a time, but don't go crazy, if you go too far the engine will start to ping and the knock sensor will retard the timing so far that you will lose the benefit of the extra timing and the engine will have even less power than at factory settings.
Steve  

Read more: http://forums.highperformancepontiac.com/70/7329176/misc-upgrade/how-to-adjust-timing/index.html#ixzz1dd9eG6as  

Am I correct in understanding that when I'm cranking the engine to first start it up, only the initial timing is in play?  Is the ECM adding to that timing by sensor feedback, before the motor actually starts?  In other words, will taking some initial timing out solve the problem with the hard starting, or will the ECM just add back whatever I take out?  

an easy way...  to set TPI base timing..  and get it really close..

set the damper to what ever base timing you want..     align the tips of the pick up coil and the reluctor perfectly..  this is where the base timing is set..


on engines without a crank sensor.. // with a pick up coil in the distributer..  below 400 RPMs the timing is controlled only by the module..  directly from the pick up coil..

once above 400 RPMs.. the computer takes control..   it looks at the timing signal from the pick up coil relayed through the module .. then decides how long to wait to send the module to the signal to trigger the spark..

this ability to change the coil discharge signal allows the system to advance the timing from the base timing..

what is hard to comprehend.. is the pulse from the pick up coil is generated..    the module and ECM wait almost an 1/8 of a turn before the module fires..  so when the rotor is passing the number one position in the cap..  that exact signal does not fire right then.. it fires for the #8 cylinder..  

it just works..

the module does have some advance built in ..  just not a lot.. but enough to keep the car running under reduced power. for limp home mode...


one thing.. if your engine is equipped with a small cap HEI.. you will want to pull the rotor and examine the center opening around the distributer shaft..  the first layer on the shaft is the steel reluctor.. its riveted on top of a ceramic magnet..   that magnet is visible around the shaft.. if that magnet has visible cracks.. you will need to replace the distributer..  go for a reman that has a totally different shape of reluctor.. the remans with the different shape reluctor create a much better wave form for the module to work with..

when the magnets are cracked..  the pick up coil creates GHOST wave forms.  this fools the module and the ecm into moving the spark signal around.. makes for some weird misfiring..


hard to start TPI cars...

verify the fuel pressure is up...  running fuel pressure should be 38 or 39 PSI as i recall..  i have just been told on another forum that it is higher on later TPI cars.. i have not been able o verify that yet..

when you turn the key on.. the ECM sees the on signal and sends positive power to the fuel pump relay for 2 seconds..

the ecm will also continue to keep the fuel pump relay engaged while the engine is being cranked by either the wave form from the distributer or the cranking signal from the wire split off from the starter solenoid wiring..   depending upon the ECM you are using..

at the same time..  the 3 wire oil pressure switch should be closing the outer 2 contacts together sending power to the fuel pump..    

the 2 seconds on as the key turns on primes the system...


do you have an ALDL diagnostic port wired up.. so you can plug a scan tool in...

TPI motors usually require just less than 0.40 volts from the TPS (throttle position sensor)  for the system to work properly..   0.35 volts is the usual setting..   if you have more than this..  it might fall into clear flood mode.. shutting down the fuel injector pulse width.. making the car hard to start..

as for will the ECM put in what you have taken out..  probably not .. it will if you have a proper detonation sensor in the side of the block hooked up.. learn how much more timing it can dial in..  

i don't know what the base timing should be on that motor... .  the people who programmed the ECM might be the ones to talk too about it..

please post if you have a scan tool port wired in.. and a scan tool that will work..

if you can.. take some scan tool reading with the engine warmed up..  and idling.. please post the various sensor readings from the scan tool.  i can decipher them...  it will really help..

TPS    0.35 volts.. and the same each time you let off the gas
ECT    190 to 215F  with a 192F thermostat..
MAF   .6 grams per second..  or some variation on that.. 6.0  .6.. .06.. i cannot remember right this second..
injection pulse width..    1.3 to 1.4 MS at idle..
IAC.   20 to 50 if the base idle is set correctly..
O2 sensor   0.1 to 0.9 volts and swinging back and forth constantly.. .
O2 sensor cross counts..  over 10 as i recall

this almost looks like the same problem over on the RRT..

you will probably want to print this.. and go over it several times..  

Thanks Wayne for the 411.  

Wow Wayne - now that's a detailed explanation.  

Yes I have an OBD-1 port wired in.  I also have a buddy with an old OBD-1 scanner.  We got stuck with the scanner though because it wanted a VIN, and the VIN for a '40 Ford is of no use.  A guy on the third-gen Camaro forum gave me a VIN to try.

Not many of the acronyms mean much to me (they will to my buddy).  The one thing I notice is you reference a reading for the MAF sensor, but I'm told this is a speed density setup, not MAF - does that matter?

As for base timing, GM's paperwork for a ZZ4 says 8* BTDC, but I don't think that applies to fuel injection.  

Just to clarify, by "hard starting" I don't mean that it cranks a long time before catching, I mean it has a metal-to-metal banging sound after a few cranks which to me sounds like the motor kicking back on the starter (too advanced).

To re-pose my earlier question, am I correct in understanding that when I'm cranking the engine to first start it up, only the initial timing is in play?  Is the ECM adding to that timing by sensor feedback, before the motor actually starts?  In other words, will taking some initial timing out solve the problem with the hard starting, or will the ECM just add back whatever I take out?  

i got a tpi car from a guy & it spit out no codes.  hard starting & didnt run to great.  the timing was way off.  i do think the initial timing with that wire disconnected is 0, but i dont have a manual handy.  im not sure about the timing on start up, but i think it stays in open loop for a few seconds until it warms up & gets info from the 02 sensor, & temp sensor, etc.  try retarding your timing & see if it makes a difference.  you can always put it back where it is.  
made in U.S.A.

print this chart....

please .. ONLY READ ONE LINE AT A TIME...
these are generic GM numbers and NOT every car will have all the data streams shown..  i normally say to start with the engine cold..  go over your scan tool data.. start with the first line on your scan tool.. find the proper line on the chart. . write it on the left side of that line..  to everything cold...

then warm the motor.. go over it with the engine idling engine hot. this time.. mark the data down on the right side of the page..   then you can sit at the work bench and compare data to what it should be ...

these numbers are kinda generic..   please post your data numbers.. i can decipher them..




as for what the noise is.. banging upon starting..  hmm..

disable the ignition..    crank the engine.. does it still make the noise ?????

if you can  get a good video recording of the noise.. you can post it to photobucket.. and then the 4th link here..  so we can see the video and hear the audio track..  do several attempts..  to get to the 4th link.. hover your curser over the image on the photobucket page.. a tool bar will drop down.. click the 4th one. then paste that link here..  

I will see my buddy with the scan tool on Saturday, and try to collect that data.  I'll see if I can get and post a video with sound.  

are you in oklahoma??? i spent half an hour talking to the fellow with a TPI in a 40 last night..

something... if you can get your scan tool...

throttle position sensor  voltage needs to be 0.54 volts with the throttle closed..

the idle air counts need to be between 20 and 50..  at idle with the engine warmed up..  

if you can get the engine started... with the scan tool plugged in..  see if you can find a time when the IAC.. idle air counts are at 0....   at that point .. unplug the idle air control motor from the bottom side of the throttle body..   you can then..  back off on the idle stop screw until the engine is stable at 400 to 450 RPMs.. just barely running..  look at the TPS voltage..  .. if you can get it to idle with the idle air control motor seated..  shut the engine off..   turn the key back on and check the TPS ..   is it at 0.54 volts..  if not before doing anything else.. adjust it. .. plug the idle air control motor back in.. .  

start the engine.. you might have to hold the idle speed up for 15 seconds..   then SLOWLY. let the throttle close.. letting the engine return to idle over about 30 seconds. so the ECM can learn the new settings...

look at the engine idle speed.. should now be around 500 rpms.. with a coolant temp sensor reading of 192 to 215+   with the TPS at 0.54 volts..  the IDLE air counts need to be between 20 and 50..   closer to 20 is better..  

if the the Idle air counts are less than 20.. open the throttle stop screw slightly..   if higher than 50 .. close the throttle stop screw slightly..    be careful that the throttle blades are not so far closed that they bind when you open the throttle..

with the scan tool..   please verify that  the TPS voltage is the same  each time you close the throttle.   if it varies.. you probably have a worn throttle body shaft bore..  this change in the lowest voltage from the TPS will confuse the computer..

all this has to be set and the engine needs to be started a few times.. perhaps driven  or at least the rear wheels powered and the transmission ran through the gears.. i am taking that the vehicle speed sensor is hooked up to the ECM..

this is not the only thing that will cause problems with starting.. but it has to be perfect..  

Nope, not in Oklahoma, I'm in Massachusetts.  I'm going to print this and try it on Saturday.  

Wayne:

Planned on meeting my buddy with the scan tool today, but can't even get it to start anymore.  Here's a video of the clanking noise it makes when cranking.  



Now after about 5-10 attempts at starting, something that looks like smoke (but doesn't smell like smoke) is coming out the front of the motor.  Pretty sure its from the air cleaner.

Can't tell you much more than that today, because he has the scanner.  He's gonna come over tomorrow, and see what we can figure out.  

wow....

thats BAD  as NOISE goes...

can you get the car up in the air far enough to get under it...

with the key OFF... can you get a 6 point 5/8 socket on the crank bolt and turn the engine with the crank bolt.. see if it makes any odd sounds while you are slowly turning it..

next.. what year is the short block...  pre 1982 or post 1982????  

that noise really sounds like a starter to flywheel problem...

let me explain...

all gm blocks up TO 1982 had starter bolt holes in the block drilled and tapped for 3/8 -16 bolts.. these will have six stripes on the head..   3/8 is .375"

from the beginning of the 1982 model year.  the starter bolt holes in the block were changed to 10MM-1.5 thread..   10MM is 0.3937"

if you use a 3/8 bolt in a metric threaded block it won't really hold at tight as it should..

if you use a metric bolt in a sae threaded block... you will usually either break the bolt on before the starter is pulled up tight.   or  SPLIT the corner off the block..

so.. please test fit the bolt in the block.. do they thread all the way in by hand .. without the starter... or not..

sae bolts have six stripes on the head...  

metric bolt have numbers...     10.9  12.9

additional.. if you look closely.. the starter bolts have a knurl just above the threads.. the block is counter bored...   the starter holes are also different sizes..     the knurl centers the starter and the block to keep the starter gear depth perfect..



if you use a metric starter on an sae block..no matter how tight you get it.. it will wiggle around and make noises quite similar to what you are hearing..

more to come...  

so.. drop your starter...

check the fit of the bolts through the starter in the proper direction. that the fit is tight as shown above..

examine the individual teeth on the flywheel..  you can dress any burs down with a file..  i use a diamond file.. but thats just because i have one...

stick your finger into the opening on the starter nose..    spin the gear..  it should only spin in one direction..  it should not be loose..   it should catch when you try to spin it backwards instantly..  there are springs and rollers to prevent it from spinning the wrong way..  but they do wear out..

check the nose cone for cracks..  they do snap off once in a while..    if you broke it.. and the rest of the starter is good.. replacement NEW nose cones are available from some starter parts suppliers...  

The block is a ZZ4 crate motor, which they haven't made for too too long, so I presume its a post '86 one piece rear main seal type block.  The starter is maybe 2 months old, but I wasn't the one that changed it.

I'll try that tomorrow.  Sure hope it's just that simple...