that makes sense.
06/05 = msb+lsb = RPM, very good ! 16bit word, and of course needed.
i will confirm that
can wait to find short and long turm fuel trim CMD, param.
will look for a 44 pin PLCC test clip, with cable if possible.
still looking for spark advance. param. or MPH.(VSS)
i have a box full of 100 min. toggle switches, and will wire them all in to my harness.
for the binary inputs. or just a simple dip switch would do.
the hard binarys are , Closed Loop flag, o2 activation in manual.
and , fuel pump activate (i will find it), Egr solenoid,Evap solenoid,
but CL is my most important flag, gotta find that puppy.
cheers.
__________________ http://carfix.stufftoread.com/
click me.
91'-8v,2dr,4wd,Conv. Market#E03
My server has Engine Repair proc/data. Check it out.
Top prize parameters:
Closed loop bit flag, (not found, yet)
SFTM AND LONG TERM FUEL TRIM, LTFT.
ALL SENSORS ( we got them ) MAF,map, ect,iat, oxy(pending).
with just these , troubleshooting any 89-95 kick would prove vastly easier.
no?
keep in mind. 128. =80 hex and 80 hex could be analog sign bit for A/D converter but it is also be the center range bit for both the
STFT and LTFT counters. ( fsm claims that it is)
I think they would start at 80 ,at key on , and memory wire power first power up. those trim values.
wonder if i can trick my bench jig into fooling ecu to go in to closed loop.
(.5v frozen oxygen sensor input would help, pray not needing to swing it)
Close loop addr. 01,03,04,08, 15 , might be one of these.
cheers,
__________________ http://carfix.stufftoread.com/
click me.
91'-8v,2dr,4wd,Conv. Market#E03
My server has Engine Repair proc/data. Check it out.
The 80h register values aren't A/D related, all A/D outputs have been logged. I also think that this is the centre value for fuel trim values. Shouldn't be too difficult to find in the code if we can find the counter for the O2 sensor.
The O2 sensor must be toggled, it is a digital input to the processor. The O2 input is connected to a comparator with the threshold at approx. 0.4V so output is just high/low. On my ECU comparator output goes to a via and is passed to an inner layer. I have hunted for the other end of that track several times but cannot find it.
I don't think that it will be an input capture because it toggles very slowly (approx. 1Hz). I may be able to get my Vit into closed loop mode if I lift the rear wheels off the ground and run it in gear. From the various FSMs the vehicle must be over 2000rpm and at normal operating temperature.
A number of functions are related to MAP, RPM and ECT/WTS, these are fairly well defined in the FSM so should be apparent in the code. Do you have a transfer function for ECT and MAP?
YES I DO, both factory and measurement on new sensors. (did that ,so I can make my megasquirt to run factory sensors, it does! and no cut wires in the CAR !)
thanks for that tip , didnt know they did the o2 , outside with a comparator.
makes sense , cpu not able to do 10 conversion a second and still have time for other things. so the compar.so it just says or , rich lean . ( after all NB sensors only do that any way)
so i will build a 8 cycle per/second 1v sign wave oscillator ( i have a function generator in my shop that cand do that easy)
i can cruise with my bench setup , easy. no wheels to catch my neck tie. ahahaha
Im told these early ecus ,store the prestart baro reading in a seperate variable for
altitude compensation, ( some one Hondas have 2 sensors , one just for altitude)
but not these old dogs, just 1.
in your code somewhere.....
off to make O2 work on bench jig.
i can run any RPM at any load now , no problem, Have all inputs controlled.
just the O2.
as your sure about 1hz swing , all my oher cars swing 6 to 8 or more times per second.
I have not scoped my real cars O2 yet. called crosscounts .
which slow with ageing sensor.
I have the sensors characterised I did then when I designed my own ECU. The sensors are actually quite accurate, the big variations quoted in the FSM are to allow for measurement errors. I used the Steinhart Hart algorithm to plot the complete characteristics
What I need is the actual/binary ratio for the diagnostic link. Then I can look throught he comparisons in the code to match values with functions.
Quote:
Originally Posted by jtgh
Im told these early ecus ,store the prestart baro reading in a seperate variable for
altitude compensation, ( some one Hondas have 2 sensors , one just for altitude)
but not these old dogs, just 1.
in your code somewhere.....
The baro reading is in the code somewhere. There are a number of MAP registers, the CPU reads the A/D and copies the values into a whole bunch of registers, this is more efficient than fetching them each time that they are needed and then moving them a working register. Unfortunately I have to search the code for lots of registers rather just the one but I will find it.
Quote:
Originally Posted by jtgh
that rom mark is probably really a gatearray ASIC. but not sure.
Only Joe Suzuki know, eh ?
thanks for all .
Definately not a ROM it is a CPLD/ASIC implemented as a port expander.
On mine L0002 is a port, it has a bunch of timers on it. Bits 5 and 3 are flags so you need to isolate them.
L0019 - interesting, I have one address that I have noted as "pointer?". I'll recheck my notes when I get home from work. My notes are a bit scattered as they have been written over several months while sat in the car surrounded by test equipment. Presumably you can't actually read throught the entire PROM just the first 256 bytes?
I will do some more logging with the Logic Analyser at the weekend. This is the way that I hacked the ECU - read through the code, do some logging, check, relog, read some more code............
Injector pulse width should be variable using the ATS/IAT sensor input.
I saw your pages and they are great and very good for the sensors and I see your home grown ECU, very nice....
do you have a chunk of Borland 6 code I can use for just logging raw data. to screen or file. I just got Borland 6
EG: (psudo code)
send cmd
get data
log data
repeat 32 addr?
I dont have the RS232 serial interrupt drivers for BL6.(yet)
I will try to figure out the binary ratio on the ECU for the two thermistors.
just make a table with 2 deg. reso. ought to be good enough.
well at the accuracy of my Potentimeter that is. and shaking hands....
thanks for the nfo on that ASIC. chip.
Trying to make a OXy sim input so I can find the proper addr. in ECU.
5 HZ might work.
__________________ http://carfix.stufftoread.com/
click me.
91'-8v,2dr,4wd,Conv. Market#E03
My server has Engine Repair proc/data. Check it out.
According to my calculations at 23C, the voltage input to the ECU should be 2.5V (so mid-point is biased to ambient!). So the reading should be 80h. I have logged the reading across the diagnostic link at 7A which is pretty close (my thermometer battery was flat!). WTS and IAT are the same sensor and the interface circuitry is identical.
With the sensor disconnected the input is biased at @ 4.75V, so a constant 4.75V will trip a WTS high voltage fault code - this should be easy to find in the code!
it flips to 70 hex, for shorted sensor and infinity .
nfo;
-------------------------------------
ECT transfer. ( with pot and switch so i can measure ohms accurately)
ohms ---- data---------
--------------------
short= 70 hex
open = 70 hex. infinity
85hms -lowest allowed restance [ above boiling] , below which 70hex flag
97 ohms = E0 (hex)
384 = 9F [ about 80 deg. F}
581=90
815=84
2.4k = 65
4.7k = 55 ( max on 5k pot)
10 (9.9kreal) resistor = 45 (below zero temp. )
I will get a bigger (20k pot and find the max low temperature value.)
__________________ http://carfix.stufftoread.com/
click me.
91'-8v,2dr,4wd,Conv. Market#E03
My server has Engine Repair proc/data. Check it out.
I've had a look at the code. ATS and WTS I had aready put in. I'm not sure if I've uploaded this version of the source code. My comments (not logged to confirm) give 12h and F6h as the WTS fault thresholds, 94h - L006C is the default value. I can only see one fault flag (not high and low?) which is L0082, bit6.
ATS thresholds are 15h and F5h, default value is 7A (doh, this is what I got I later found a problem with the ATS connector!). Fault flag is L0082, bit7.
TPS flags fault at 0? (L0082 bit4 and bit1) and 3D? (L0082 bit5) - really not sure about this, TPS value is loaded from L0070 - maybe TPS angle/idle switch.
Elsewhere TPS_Raw fault is given if lower than 08h and higher than F8h (L0082 bit3)
L0046 bit5 is related to delta TPS so maybe a flag for fuel cut or fuel enrichment.
MAP thresholds seem to be 0Bh and E8h. I think that MAP high is L0082 bit0and MAP low is L0082 bit1.
I haven't found anywhere that L0082 is copied to a diagnostic location, it may be an indexed copy. L0081 and L0083 also look to be fault flags.
Linear Mode
