RhinoPower Ltd

TechSupport · Guru

RE: ECU ID 1902, P/N #33920-65DA1

fordem · Senior Member

Maybe I'm missing something or maybe that screencap is too small for my aging eyes, but that looks pretty good to me - so it's off a degree or two here and there - surely that's better than a completely uncalibrated dashboard gauge that tells me hot cold and somewhere in between.

To come from a different angle - do you know of any way to calibrate an AFM - or to measure airflow? How do we know what output from the SDL's AFM addresses correlates to what airflow?

Also have you considered displaying the ECU's target RPM?

On the datalogger page - I think that's the one that let's me select a graph (and which I have not used or rather figured out how to use) - I can only select from a predefined list of sensors which would depend on the ECU mapping to define an address - would it be possible to enter an address directly or select from a list - I'm still trying to figure out where the ignition timing advance would be.

TechSupport · Guru

I've been through all the calculations for the WTS and verified these by logging the programme execution on a spare ECU. Its not coming out very linear despite the look up table. It looks like the sensor characteristics are a little different to what is stated in the FSM or my Steinhart-Hart coefficients are out. The nearest I can get to a straight line approximation is:

y = 0.7053x - 41.575

where X is the SDL output and Y is temperature in degs C.

-- Edited by Rhinoman on Wednesday 21st of October 2009 11:17:39 PM

TechSupport · Guru

I've spent some time looking at the code and the schematics for my ECU. All the code does is linearise the temperature sensor readings. Surprisingly although the ATS and WTS have the same characteristics and input circuitry they have slightly different mapping. I've calculated the Steinhart-Hart coefficients for the sensors and created a resistance v temperature graph and calculated the ADC voltage. I will use the ECU look up tables to determine a slope for the response which will give the correct temperature readings. The good news is that all the Zuk ECUs that I've looked at so far use the same sensors.
I'll add the LUT locations to my Tuner Pro definition file.

TechSupport · Guru

I have the characteristics for the WTS and ATS but the output from the data link isn't a raw voltage, the ECU manipulates the data in some way. I have modified the code in my spare ECU to increase the number of diagnostic addresses so I can compare the raw value with the modified value. That should give us a better idea of whats going on.

fordem · Senior Member

Status flags - there appear to be two address locations handling status flags, possibly three.

#37

37 bit 8 senses electrical load - it is a 1 if the lights or rear screen defogger is turned on and a 0 with them both off (display varies between 1 & 129 - if the engine is off - and between 5 & 133 if the engine is running - see below)
37 bit 4 - I believe this may be the fuel pump status - engine off it is 0, engine running it is 1 (see paragraph above for display).

#38

38 bit 5 senses closed throttle - it is a 1 with the throttle closed and a 0 with the throttle open (display changes between 16 & 0)
38 bit 3 senses a/c load - it is a 1 with the a/c on and a 0 with the a/c off (display changes from a 16 with the engine idling and a/c off, to a 20 with the engine idling and the a/c on, or a 0 with the throttle open and the a/c off to a 4 with the throttle open and the a/c on.
38 bit 2 senses p/s load - it is a 0 with the steering released and a 1 with the steering turned (display varies between 16 & 18 with the engine idling if the steering is turned and released)

#39

This is the one I am not certain of

39 is a 0 with the engine running and a/c & blower off, changing to a 1 with the blower on, and 49 with the a/c compressor running.

Notes - this vehicle has no O2 sensor or EGR, so I am unable to find those flag positions, I observed no changes when pressing & releasing the brake pedal and looking at the ECU connector pin out I see no input from a stop light switch.

fordem · Senior Member

I don't know how much help these will be, but I found them in the FSM and thought I'd post them.

Left is ECT, right is IAT

-- Edited by fordem on Thursday 15th of October 2009 02:48:44 AM

TechSupport · Guru

Referring back to my notes from around a year ago, ambient was -1c !!
I have two tests, I did these with an infra-red thermometer. The first test isn't very clear:

C         WTS   ATS
-1        44      60
7?        47      84
10       82      110
30       130?   120

The second set is much clearer:

WTS    C
110     24
128     34
132     40
139     45
145     50
150     56
153     60
159     68

Which tallies with what you get. The ECU applies some conversion to the data, I'm not exactly sure what, the code is quite complex and there is a lot of error checking. I need to get my spare ECU fixed and log the code execution from the external bus.

-- Edited by Rhinoman on Tuesday 13th of October 2009 10:00:47 PM

fordem · Senior Member

I did a "temp calibration" but this time in *C

Ambient was 30.7*C, the engine was not quite cold when I started..

Fluke (*F) - ECT (#10) - IAT (#29)
34.6 ...........121.............114
35.0............125.............115
36.0............128.............115
37.0............130.............115
38.0............131.............115
39.0............132.............115
40.0............133.............116
41.0............134.............116
42.0............136.............116
43.0............137.............116
44.0............139.............116
45.0............140.............116
46.0............141.............116
47.0............142.............116
48.0............143.............116
49.0............144.............116
50.0............146.............116
51.0............147.............116
52.0............149.............116
53.0............150.............116
54.0............151.............116
55.0............152.............116
56.0............153.............116
57.0............153.............116
58.0............154.............116
59.0............156.............116
60.0............157.............116
61.0............158.............116
62.0............159.............116
63.0............160.............116
64.0............161.............116
65.0............162.............116
66.0............163.............116
67.0............165.............116

The output seems to be a 1:1 relationship with *C with an offset of 97 - but there appears to be an anomaly in the lower readings, which may be either a nonlinearity in the sensor output - or - may relate to the way the reference measurement was taken - the thermister bulb for the Fluke was tucked under the edge of a rubber hose holding against the outside of a metal coolant pipe, there is probably some thermal mass involved and the coolant may heat faster than the pipe.

With a cold engine the ECT & IAT sensorsshow roughly the same but with an offset of 97, an indicated airtemp of 116 would translate to 19*C rather than the ambient which was 30*

TechSupport · Guru

TPS - The Baleno ECU exhibits exactly the same behaviour with a difference of 31 across the whole range, on the Vitara I have applied a scaling factor of 0.9 (TPS Angle = Raw data*0.9).

-- Edited by Rhinoman on Monday 12th of October 2009 07:21:06 PM

fordem · Senior Member

Last two - key on, engine on - at 2000 rpm - the vehicle is stationary and I have my foot on the gas pedal holding it open.

Originally I thought that only one address changed - #38 - going from a 16 with throttle closed, to a 0 with the throttle open, but I notice from the screen captures that #37 has also changed from a 1 in the key on, engine off shot to a 5 when the engine is running.

Looking back at my notes I can see that #38 also changed from 16 at idle with the a/c off, to a 20 at idle with the a/c on.

OK - now, so to speak, you're seeing what I'm seeing - maybe you can see something that I'm over looking

fordem · Senior Member

The next two are key on, engine on - at idle - the engine is cold, so the idle is a little on the high side.

fordem · Senior Member

I figure the easiest way to do this might be to post a couple of screen shots rather than create a bunch of tables, and I figured I might as well throw up the page 1 addresses, just to be complete - so here we go - first two are key on, engine off.

TechSupport · Guru

I'm wondering if the software is correctly reading addresses above 31, maybe it isn't and the timing location is up there. What data are you seeing in the upper addresses?

fordem · Senior Member

Now that you mention it - there appears to be a horizontal "bulge", possibly another resistor in the heat shrink tube between the pot and the connector - take another look at the "in situ" picture. I don't think I have that size heat shrink tube handy, so I'll forego cutting it open for now.

There is no locking mechanism that I can see - apart from the latch on the connector to stop that from vibrating loose.

TechSupport · Guru

That looks like a bit of a lash up! Thanks for the pics, the value of the pot is nice and clear, 50K is quite high so I would assume that its in parallel to another resistor inside the ECU. Is there any kind of locking mechanism for that? it looks like it could vibrate out of adjustment.

fordem · Senior Member

CO adjusting resistor - left is in situ, right is a close up

-- Edited by fordem on Monday 14th of September 2009 07:31:19 PM

TechSupport · Guru

The speedo gauge itself is an analogue gauge and is likely to be non-linear across its working range. This makes it a little less accurate than the ECU which uses a timer derived from the main crystal to calculate speed.
Naturally both the VSS and ECU readings will change a little with tyre size/wear and pressure.

I'm slightly surprised that your CO resistor is adjustable, they had those in some early Swifts. I understood that over here the resistors are fixed and are swapped by the dealer to fine tune the CO. This may be different because of our emissions regulations and the presence of the catcon. Can you take a pic of the resistor please?

I can't take any readings on mine for a few days as I'm upgrading the brakes and I have a bit of a vacuum leak where the servo should be!

fordem · Senior Member

On this vehicle VSS is a sensor on the side of the transfer case that is wired to both the ECU and the speedo - wouldn't that give them similar accuracy?

TechSupport · Guru

I think that you're right on the km/h, the VSS should be a little more accurate than the speedo. I'll compare the temperature readings with what I get. I've been away for a couple of days so I've got some catching up to do.

fordem · Senior Member

Well - I finally found the timing mark - a single notch in the rear edge of the rear pulley - and more by scraping with a screwdriver than by eyeballing it - cleaned it up, put a dab of white paint, and got out the timing light.

The engine seems to be set for around 20* BTDC - with the jumper in - and there is around a 5* variation (at idle) if the jumper is removed & reinserted. This is way more than the FSM specifies but is not that unusual here in Guyana, we get only one grade of gasolene - super unleaded - and it is common practice to advance the timing to take advantage of the power available.

More to the point - none of the addresses on either page seems to relate to ignition timing - I've tried inserting & removing the jumper, both at idle and at higher rpms. and none of the addresses seem to change as the jumper was inserted and removed.

VSS

Could this be calibrated in km/h ? See the tables below

km/h - #9
..30......17
..40......22
..50......26
..60......33
..70......35
..80......43

These were taken against a GPS and allowing for a +/-1 accuracy of a digital counter, subtracting 2, +/- 1 from the value displayed and then multiplying by 2 seems to report the indicated speed in km/h - #9 is at 0 with the vehicle stationary.

mph - #9
..20.....18
..25.....22
..30.....25
..35.....30
..40.....35
..45.....40
..50.....43

fordem · Senior Member

I can now confirm that #17 & 18 work with the MAF sensor - with the vehicle stationary, MAF connected and engine running these valuse will change as the throttle is opened and closed - if the MAF is disconnected, they will remain at zero under the same conditions.

Unfortunately I am still completely stuck on the ignition timing/advance - I am still unable to locate the timing mark on the crank pulley - I've looked at the FSM, I can find the pointer on the timing chain cover, but not the groove on the pulley.

Looking at the dataviewer with and without the timing jumper inserted - nothing jumped out at me - there is of course the possibility that the cut paperclip was not making electrical contact, however, with it in place the behaviour of the check engine light seems to change - it flashes briefly when the ignition is switched on and then stays off, even if the engine is not running.

fordem · Senior Member

Rhinoman wrote:

Do you know what the flow rate of thr J20 injector is? I want to add calculated mpg.

-- Edited by Rhinoman on Saturday 12th of September 2009 12:10:43 AM

The injectors are orange and marked INP771 - this appears to be a NipponDenso number and from the charts I can find should be 200 cc/min or 19.6 lb/hr.

fordem · Senior Member

You had mentioned that you expected MAF to have two addresses like RPM - these are possibly #17 & 18, this is almost certainly a hi/lo byte pair and I can play with it with the throttle with the vehicle stationary - I'll try to confirm this by disconnecting the MAF sensor over the weekend.

I did some temperature scaling - using a Fluke 179 with the temp sensor bulb wedged between the rubber of the radiator top hose and the metal pipe - it's not in direct contact with the coolant but as the coolant gradually warms, so should the pipe, so the readings should be reasonably correct

Fluke (*F) - ECT (#10) - IAT (#29)
90...............114.............116
101.............134.............115
110.............139.............115
120.............144.............115
130.............153.............115
140.............158.............115
150.............163.............116
160.............167.............116
165.............172.............116
170.............174.............116
175.............178.............116

Coolant temp stabilizes at around 175*F - it never reached 180.

Ignition advance/timing - I have been able to get the timing light to trigger on the COP setup, but unable to locate the timing mark on the crank pulley - apparently the mark is a groove about 1/16" long on the "front edge of the rear pulley", so now that I've been shown what to look for I'll have another go at it.

If I'm correct there is a jumper that can be inserted to lock the timing to 5* BTDC, so I may use that to try to pinpoint the address - with the jumper out it should increase with rpm, but hold steady with it in.

-- Edited by fordem on Saturday 12th of September 2009 04:13:27 AM

fordem · Senior Member

This is what my injector pulse looks like ...

Engine is warm and at idle - RV0.3 shows #15 = 2, #16 = 147 (note #16 has a small fluctuation in value)

TechSupport · Guru

The MPI ECUs use a conventional set up. The SPI needs a relatively large injector to get enough flow at maxium load, it then needs a more complex set up to achieve a short enough pusewidth at idle.
Do you know what the flow rate of thr J20 injector is? I want to add calculated mpg.

-- Edited by Rhinoman on Saturday 12th of September 2009 12:10:43 AM

fordem · Senior Member

Man - for a forum with so few members, this one is pretty dynamic - posts being replied to and moved whilst I'm thinking about my response - now I don't know which thread to reply in. biggrin

I'll stick with this one for now, since the response is pertinent to this ECU.

Looking at the FSM diagrams, on this vehicle the positive side of the injector coil is fed from a +12V line controlled by the main relay which is how I, so to speak, expected it to be, and the negative side is wired back directly to the ECU, so I would guess that we don't have that feed back circuit that's being discussed in what is now the "other" thread, I also experience no apparent change in engine operation due to loading of the drive circuits by the scope probes.

fordem · Senior Member

Juha wrote:

Could your "interesting feature" be just ordinary coil mutual induction? When current flow is interrupted inductance try to maintain current flow. This cause voltage to rise toward infinity.
If this mutual induction spike is clamped it cause solenoid (injector, relay etc) to hold longer. Hold time depends how much energy we have stored in our inductor and how low clamping voltage we use. Solenoid releases easier/quicker proportional how high we can let the voltage rise. Unfortunately in real life we have quite low limits for clamping voltage. In practical circuits I have seen values between few tens of volts up to 150V. Ignition systems are different they are usually clamped around 400V. This is protection against open secondary condition. Normaly primary voltage rise only 100-200V before the spark ocures.

Looking your scope picture I am more puzzled why voltage after injector pulse is left higher level. It seems to be around 13V before pulse and around 14V after the induction spike have calmed down.

That was actually my initial thought and the reason I paid no attention to the "spike" when I first looked at the injector pulse width - I was expecting a "turn off" spike from the collapsing magnetic field, and I saw what I assumed was one - but when James suggested that this was a "spike" being sent by the ECU to turn off the injector faster, I didn't question it - this guy knows more about these ECUs that I do...

-- Edited by fordem on Friday 11th of September 2009 07:00:17 PM

TechSupport · Guru

http://www.activeboard.com/forum.spark?aBID=133095&p=3&topicID=30468381

I've moved this to try and keep this thread on topic.

-- Edited by Rhinoman on Friday 11th of September 2009 03:24:05 PM

Juha · Member

Could your "interesting feature" be just ordinary coil mutual induction? When current flow is interrupted inductance try to maintain current flow. This cause voltage to rise toward infinity.
If this mutual induction spike is clamped it cause solenoid (injector, relay etc) to hold longer. Hold time depends how much energy we have stored in our inductor and how low clamping voltage we use. Solenoid releases easier/quicker proportional how high we can let the voltage rise. Unfortunately in real life we have quite low limits for clamping voltage. In practical circuits I have seen values between few tens of volts up to 150V. Ignition systems are different they are usually clamped around 400V. This is protection against open secondary condition. Normaly primary voltage rise only 100-200V before the spark ocures.

Looking your scope picture I am more puzzled why voltage after injector pulse is left higher level. It seems to be around 13V before pulse and around 14V after the induction spike have calmed down.

fordem · Senior Member

Mmmm - tomorrow I'll try to take a second look at the injector pulse width - and see if I can find the scopemeter software so that I can get a screen capture.

I just hung the scopemeter off of the ECU connector - #4 if it makes a difference - and told it to give me the negative going pulsewidth, without paying that much attention to the actual pulse shape - I recall there was a positive going spike, but I cannot remember if it was on the leading or trailing edge.

You are correct on the cat assumption, there isn't one - but, if you mean the CO adjusting resistor when you say fixed resistor, what I found appears to be a 10 turn precision pot, with a little knob where the adjusting screw would normally be.

With regard a connector - take a look at the picture which shows the top side of the ECU board, find the processor and then go up to the connectors and look to the right of the capacitors - you'll find three solder pads with a rectangle silk screened around them - but I have no idea what it's for.

TechSupport · Guru

As you don't have an O2 sensor, I assume you have no cat either. It would be an interesting experiment to hook up a variable resistor in place of the fixed resistor to see how much adjustment you have.
SH7202 is reflashable although I don't see a position for a connector or any pads for a bed of nails.

TechSupport · Guru

The P is a typo, I must have pressed both keys at once. I measured the injector pulse width on my 8V tonight.
I measured the injector pulse width on my 8V tonight, the scope probe messes with the injector circuit and I can't get a proper idle but the waveform is clear. This is the signal on ther ground side of the injector, its the easiest signal to analyse because you see very little of the PWM, just a couple of small bumps at the end of the main pulse. The big surge at the end is an interesting feature, this is caused by the circuitry reversing the current across the injector to slam it shut, I am seeing pulsewidths of less than 1mS at idle.

inj_pw

Rhinoview v0.3 was reporting around 5.2mS. I then managed to capture fuel cut:

Fuel Cut PW

The ECU reports this as 100 (raw data) RV shows 0.4mS so my original assumption that this was injector PW in us was correct. Although the first waveform looks to be 1.83mS, I'm obviously not seeing the actual closing point.

-- Edited by Rhinoman on Thursday 10th of September 2009 07:37:33 PM

fordem · Senior Member

OPBD2 ?? Is that P meant to be there or is it a typo?

I'm also thinking that perhaps I should send you those pictures full size - the processor is a MH7202F, which I understand to be flashable, but I haven't really explored in that direction.

By the way - I was able to locate (after some searching) the CO adjusting resistor - up behind the dash on the passenger (left) side, you have to remove the lower glove box to get to it.

-- Edited by fordem on Thursday 10th of September 2009 04:41:19 PM

TechSupport · Guru

Injector PW is out by a factor of 4. I will xcheck mine later if I remember to take the scope home. There was a bug in v0.2 so I'm not sure what scaling factor it was applying, I think it was 1:1.
Your ECU looks very similar to a UK (OPBD2) model, the SH7052 is an Hitachi processor that apparently uses the Motorola/Freescale architecture. The bike guys have been successful reading and writing these. You have to use a Background Debug Mode (BDM) interface (Hitachi call it AUD), Darcy and I have been working on this on the later Swift and Baleno ECUs, we have comms with the processor but haven't yet read the memory.

fordem · Senior Member

I edited the original post to include some Injector pulse width details.

Yes I have a GPS, the speedometer is reasonably accurate (vehicle is on standard OEM size and brand/model tires)

I'll take a look to see if there is a second address that works with the MAF - I did notice what looked like another "hi byte/low byte pair" whilst trying to find the VSS - if I recall correctly the hi byte went to 0 with the vehicle stationary but there was a reading in the low byte address and it did not have the nice linear correlation that #9 did.

I'll take a look and see if I find the CO2 resistor - I've just assumed there is one.

As regards the ECU - I took some reasonably clear pictures of the insides - I'll email them to you.

TechSupport · Guru

Do you have a GPS that you can check your speedo against?

Also I forgot timing. This will probably be around 45degs at a fast idle/engine cold, dropping to whatever your base setting is (8/10 degs?)

What does your ECU look like inside? it is readable?

TechSupport · Guru

TPS is an odd one, I get the output in degrees but I convert it to % to display. Address 11 is obviously angle because its corrected to 0 at idle. I will have to insert a scale factor for this.
WTS and ATS, I will compare this with what i get but you aren't getting an output in Feahrenheit either.
MAF, I think that should be two addresses like RPM.
O2S, no O2S means no fuel trim values either. Do you have the CO adjusting resistor?
ISC should be around 128 with no additional load, 20 looks like it could be right, theres always the possibility that your idle screw needs adjusting. I'm going to scope my ISC when I remember to borrow a Picoscope, my laptop died and my replacement doesn't have a parallel port to run my old PC scope.
Status flags on a 1701 are at 38 so it looks like yours is the same, good find. Bit4 is idle, bit2 is AC on. I have these flags as being at 38 on a 1701:
                  - egr valve (open)
                 - fuel pump (on)
                  - stop lamp (off)
                  - power steering sw (off)
                  - a/c sw (off)
                  - idle sw (on)
?                  - O2 sensor activity B1 (off)
?                  - O2 sensor activity B2 (off)
                  - electric load (on)

clearly theres a problem here as there are 9 flags listed but only 8-bits.

Battery Voltage. On the ECUs I have looked at the battery voltage is divided down by a potential divider to suit the input range of the A/D, the scaling factor is 13.653. Now 166/13.653 = 12.16V, there will be tolerances in the 5V supply and the resistors so thats pretty much spot on.
39 I'll have to think about

I'll start creating the definitions for you.
Thanks
James

fordem · Senior Member

TPS - I seem to have three addresses #11, #12 & #38 - this is a 3 wire TPS by the way

Throttle closed #11 is at 0, #12 is at 31, and #38 is at 16 - as the throttle is opened 11 & 12 increase gradually until at max throttle 11 is at 172 and 12 is at 204 - note that at max throttle the difference is also 31, this difference remains virtually constant throughout the throttle pedal travel.

#38 exhibits a different behavior - as soon as the throttle is opened it drops from 16 to 0 and remains there, as if it were perhaps an idle position switch

WTS - #10 - Engine off, key on - #10 is at 115 and jumps to 255 when the CTS sensor is unplugged (ambient is around 90*F)

ATS - #29 - Engine off, key on - #29 is at 129 and jumps to 255 when the IAT sensor is unplugged (ambient is around 90*F)

As the engine warms up #10 increases to perhaps 170, #29 actually drops slightly to 122

MAP - this vehicle does not use a MAP sensor but has a hotwire MAF sensor.

MAF - #27 - as far as I can tell the MAF sensor uses offset #27 - this varies with engine rpm, but if the sensor is disconnected, it will jump to a 127 setting as the throttle is opened - with the sensor connected it will not go above 30 with the vehicle stationary and the throttle opened (same conditions under which it jumps to 127 with the sensor disconnected. (The engine runs very rich and will not idle with this disconnected.)

O2S - this vehicle does not have an Oxygen sensor.

Engine Speed (RPM) - #6 & #7 - these behave exactly as defined the notes, #6 is the high byte, #7 is the low byte and no scaling is required.

Vehicle Speed (VSS) - possibly #9 - there appears to be a direct correlation between the vehicle speed and this value, perhaps with a scaling factor ..

20 mph - 16
25 mph - 20
30 mph - 25
35 mph - 29
40 mph - 34
45 mph - 38
50 mph - 42

Injector Pulse Width - possibly #15 & #16 - these vary in a fashion similar to rpm, and drop to zero if the engine is on the "over run" or decellerating in gear. They are at 0 with the engine off, key on - with the engine on and at idle, pulse width is 2.9mSec, #15 is 2, #16 is 200, increasing RPM has very little effect, with the engine on, at idle and the A/C on to increase load the pulse width goes to 3.7 mSec and #15 is at 3, #16 at 160 - fiddling with lights etc., to vary the load it seems that they are a low order byte/high order byte pair.

Battery Voltage - possibly #25 - Engine off, key on, battery voltage (measured at battery posts) 12.31 gives a value of 166, Engine on, battery voltage 14.14 gives a value of 192.

Idle Speed Controller - possibly #20, although this is not clear - I see a number of address change as the load at idle varies

if the A/C is switched on and off

#18 changes from 214 to 78
#20 changes from 186 to 244
#38 changes from 16 to 20 (could this be a single bit change on an I/O port)
#39 changes from 0 to 49

if the rear screen defrost is switched on and off

#18 changes from 214 to 229 - I would have expected a change in the same direction as the A/C if this were the ISC, but perhaps with a smaller numeric difference - see #20.
#20 changes from 186 to 196
#38 does not change
#39 does not change.

Ignition Timing - to be determined - this is a "coil-on-plug" system so I have to figure out a way to get the timing light hooked up.

-- Edited by fordem on Wednesday 9th of September 2009 08:29:36 PM

-- Edited by fordem on Wednesday 9th of September 2009 09:47:48 PM