Quote:
Originally posted by RBob
Over on diy-efi (I need to move this stuff to moates.net) on the gmecm area, under papers, or ECMs or such. There is a write up on '747 SA. Maybe even titled "running a SA only '747." Or something of that nature.
That paper will answer your questions.
|
Maybe someone can put this in the stickies.
Ignition Control and the '747 ecm
Robert Rauscher, 5/26/99, Updated 4/9/00
Copyright 1999, 2000 Robert Rauscher, All Rights Reserved.
Areas preceded with a '|' (vertical bar), are updated. The
spk747al.bin file is also updated. It is based on the ACSC
bin.
This write-up is on the spark control only. It also does not
include any egr spark conditions. I have been testing an
ignition only '747 setup, and do not have an egr engine.
|First off, I'd like to thank EcmGuy for providing the referance
materials that made this write-up possible. (originally
credit was given to EcmNut, apologies to both, 4/00, rr).
Second, this material is as accurate as I believe, but may
contain errors. If any, they would most likely be in areas
that use bit mapped status words. It can be difficult to
understand the exact meaning of these bits.
/* ------------------------------------- */
For sensors, I've been using:
a) vss
b) coolant temp
c) map
d) knock
There is also a power brake vacuum line restrictor. This
is a stock GM piece that slows the rate that the p/b's
'get vacuum' from the intake manifold. This reduces the
effects that using the p/b's have on the MAP sensor value.
If your not planning on doing the fuel also, try to
mount the tps to the carb. It would be worth it, it
will allow pe/wot to be determined, along with proper
map test being done. Without it, there was intermittent
error 33's: map sensor high codes.
/* ------------------------------------- */
Ecm error codes in use:
12 drp 14 coolant
15 coolant 24 vss
34 map 42 est
43 esc 51 prom
52 cal pack 55 adu
/* ------------------------------------- */
Some notes and associated files:
1) I recommend using promedit for an editor, I have provided
a .ecu file for use with promedit: spk747.ecu
I'm using the version with the build date of: Aug 05, 1998.
2) spk747al.bin, a bin file with high timing for an alum
headed engine. Use at your own risk! But study it.
/* ------------------------------------- */
(SA => Spark Advance)
(wot => wide open throttle)
(PE => power enrich, same as pe)
(map => manifold absolute pressure)
(esc => electronic spark control)
(est => electronic spark timing)
(vss => vehicle speed sensor)
(tdc => top dead center)
(btdc => before top dead center)
(atdc => after top dead center)
(aldl => Assembly LIne Diagnostic Link)
/* ------------------------------------- */
If you have an aldl scanner, you can monitor some of the spark
advance values with some small modifications to the bin.
SA values can be output at locations normally used for other
functions. Then using a little math, the SA in degrees can
be ascertained.
I use locations that the scanner does no math on, the displayed
value will then be as it is in the ecm. Then for SA values, just
multiply the displayed scanner value by 0.35 (90/256 = 0.3515625).
The result will be degrees of advance. Any initial timing will
also need to be added to this advance (see 0x009).
So, if the aldl INT value is 91:
91 * 0.35 = 31.8 Degrees.
With the initial advance set to 6 degrees:
31.8 + 6 = 37.8 degrees at the crank.
I do have a caution on the use of the aldl on the '747 ecm,
it is Slow. Hence, the entire data output during one frame, may
not be relevant to that frame. IE: If your capturing a frame,
while accellerating, things are changing, and the data is not a
'snapshot' of one point in time. The data output is gotten
from memory as it is being output.
Ok, locations I use: BLM, INT, IAC, (0x509, 0x4f7, 0x4eb).
I output RAM locations:
0x0067 (total spark advance)
0x006E (PE/WOT spark advance)
0x0077 (Retard spark amount, only half this
value is sub'd from the total SA)
The value of location 0x0067 will include the SA of the
PE/WOT and Retard spark also. This is nearly the final
value used to program the spark control counters. Only
a spark latency correction is made after this.
/* ------------------------------------- */
Here are the more important spark locations in the bin:
The left hand column are the eprom offset location for each.
0x009: Initial spark advance. This value should be what
the physical distributor timing is set to. This
value gets subtracted from the Total Spark Advance.
(but is physically added back in by the position
of the distributor relative to piston tdc).
This is also called 'static' timing.
0x00C: Max RPM used in the spark slope calculation. This is
maximum value at which additional slope spark will
be calculated to. Once this value is reached (you
have it floored, right)?, the slope increase will
be held at this value.
0x00D: Slope in degrees per 1K rpm. This is the rate that
additional SA will be added.
--- These last two work together to add additional spark
at a linear rate as the rpm increases. This additional
spark starts at 3600 rpm, and increases until the
max rpm value is reached. 3600 rpm is where the main
SA table ends. Above the 'max rpm' the total slope will
be held at the value calc'd for 'max' rpm.
0x014: Warm spark bias. This value is subtracted from the
Total SA. It's purpose is to bias the main SA table,
so that negative advance values can be in the table
as a positive number.
0x015: Cold spark bias. This value is subtracted from the
Total SA. It's purpose is to bias the Coolant
compensation spark table. Again, to allow for a
negative advance in this table.
0x019: Maximum spark (2 bytes). This value is the highest
total advance that can be programmed into the
distributor. Any Calculated Total SA above this
value gets clipped to this value.
--- It's purpose is due to the physical limitations of
the distributor. You can only advance the timing
so far until you are sparking on the previous
| cylinder. With eight spark firings per distributor
| revolution, there are only 45 degrees between each
| set of plug wire terminals on the cap. GM uses the
| value of 42 degrees as the maximum advance. This
| always a safety margin to direct the spark to the
| proper terminal.
|
| However, this is not the limit that can be run at
| the crank. If you set the static timing to 10 degrees
| BTDC, a maximum of 52 degrees can be had at the crank.
|
| The 42 degree limit is only the distributor limit.
| Crank timing is the total of the static setting, and
| the amount programmed by the ECM into the distributor.
| It just becomes a balance of how much total crank advance
| is required, versa, how much static (inital), timing can
| be used (and still hot-start).
(if it becomes a limitation, you need a DIS system).
0x01B: Maximum Retard (2 bytes). Maximum total amount of retard
allowed. For the same reasons as max SA above.
0x023: aldl/closed loop added spark. This spark is added when
in aldl diagnostic closed loop mode.
0x024: Hi-way spark mph threshold. Minimum MPH to start timer.
0x025: Hi-way spark time threshold: Amount of time required before
hi-way spark is added. Must be above MPH threshold for this
amount of time, in order to get hi-way spark.
--- Note: The ecm does an ESC test once each time after an
engine start. This will typically cause two (2) knock
counts to occur (see loc 0x555). Until this test occurs,
and passes, hi-way spark won't be added.
Sooner or later, these counts should show up on the
aldl scanner, under 'knock counts'.
Note: Another parameter (hard coded), is that no err43
(est) must have existed since power up(?). Or, is that
the test for err43 hasn't yet been run(?).
0x026: Startup added SA is reduced (re-calc'd), on this time period.
Every 'n' seconds, any added startup spark is decayed.
0x02B: Knock retard mph cutoff threshold.
0x02C: Knock retard rpm cutoff threshold.
0x02D: Knock retard temperature cutoff threshold.
--- For knock retard to take place, the engine temp needs to be
greater then the temperature threshold and the vehicle voltage
must be greater than 9.0 volts, then, either the VSS or the
RPM threshold must be exceeded.
0x02E: Burst knock MAP threshold, in Kpa.
0x02F: Burst knock TPS threshold, in percent.
0x030: Burst knock differential TPS threshold, in percent.
0x031: Burst knock, duration of retard, in seconds.
--- The last four, plus Knock Temp Threshold (0x02D),
determine whether burst knock retard takes place.
There is also an enable bit at location 0x006, bit
zero, that must be set, if burst knock is desired.
Right off, the engine temp is compared to the knock
temp threshold (0x02D). If below this temp, burst
knock can take place, if above this temp, only regular
knock retard can take place.
The burst knock qualification is as follows:
Option byte @ 0x006, bit 0 set (= 1)
Engine temp below knock retard threshold
Not in closed loop
Map value above threshold
TPS % above threshold
TPS differential above threshold
TPS differential is for more throttle opening
If all of the above qualifies, then a lookup is done
based on current engine temperature. This value is
subtracted from the total timing. It is not available
for output to the aldl link, but is reflected in the
value at location 0x0067.
Burst Knock is an all or nothing deal, no attack rate
or decay rate. Once started, it exists for the entire
duration set at 0x031. The actual retard value is
adjusted each pass through (80 times/second), relative
to the engine temperature (from the lookup table).
Once the Duration is met, the retard instantly ceases.
0x035: Main SA table. This is the beginning for what the Total
Calc'd SA will be. This is a 3-dimensional table with
rpm and map in Kpa as the look up parameters. The value
at that location is the third dimension, which is the
base or starting SA. The warm spark bias is reflected
in these table values. (ie: actual SA will be the lookup
value minus the warm spark bias).
0x10A: Coolant Compensation SA table. This table is used to add
or subtract SA, based on engine temperature, and less
so on manifold vacuum. Typically, more SA at low temp's,
less SA at high temp's, and no change at normal operating
temperature.
This is a 3 dimensional table, with coolant temp and
map VAC (inverse of map Kpa), as the look up parameters.
The cold spark bias is reflected in these table values.
0x157: Hi-way Spark Added. This spark is added while in a steady
state cruise. See 0x024 & 0x025 for qualifiers. If qualified,
a new value is calulated each pass through the code. It is
based on engine vacuum.
0x15E: PE Spark. This spark is added when at WOT. It is based
on engine rpm. I believe that a TPS is required for
the code to recognize a WOT condition.
0x16E: Startup Spark Added. This SA is added after an engine
startup. It then decays at a rate defined in two
other parameters (shown next). The amount of spark
added is based on coolant temp at startup.
If the engine is startup while in clear flood mode, no
statup SA is added.
0x174: Startup spark decay. This is the initial amount of time,
in seconds that the added spark will decay over. The decay
is set once at each start time.
0x17a: Startup spark decay multiplier. This value is a multiplier
of the Spark Decay parameter. Since this value is updated
on a regular basis, as the engine warms up, this increases
the initial decay rate. So, on hot day, the added SA will
decay faster than on a cold day.
0x17f: PE Knock Retard. In degrees, the amount of retard to
use based on engine rpm, during WOT knock detection.
0x188: Knock Retard. In degrees, the amount of retard to use
during normal engine operation. Based on map VAC.
0x18d: Retard attack rate. Amount of degrees per millisecond
that any required retard will occur at, based on rpm.
--- When knock is detected, a count is available to the
ecm code. This is read from the ecm hardware. This
count is the number of milliseconds that knock has
been occurring. This is the 'knock count' displayed
on a scanner.
The total amount of retard is calculated by:
1). Multiplying the attack rate by the 'count'.
2) Lookup on either PE retard table if WOT, if not in
PE, then it uses the normal table.
The amount of added retard will be the lessor of the
above two values; the attack rate, or the table value.
It is then divided by two. Yes, it appears that the code
divides the retard value in half before subtracting it
out from the total spark advance.
Of course, a maximum retard check is then done using the
value in location 0x01b. The ecm will not retard the
total timing any more than this value.
0x196: Retard recovery rate. Percent per millisecond that any
retard recovers back to normal at, based on rpm.
The decay routine runs 10 times a second. The decay starts
immediately upon the addition of retard.
Again, required for a smooth transition?
/* ------------------------------------- */
See Part II
