Login Register

Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Driving 97 kick tach after J18A engine swap
#1
Hello there!

I am a new member, though I have been on this site for a few months now. The info on here has been a huge help with the swap mentioned above (Baleno J18A into a G16B 97 kick). As you can tell by the title I need help wiring up the tachometer to read the RPM. I tried hooking it up to the RPM output on the ECU, but without luck. After a quick search on here, I found that the tach reads a 50v impulse and the (most likely) 12V signal from the ECU cannot drive it. I even built an arduino circuit utilizing the "tone" code to generate different frequencies, but 12V does not seem to cut it.

I wanted to ask whether the tach takes a square wave or a sine wave input, and whether I can bypass whatever filter it has inside the cluster to use a lower level impulse. Even if the pulse count at set RPM doesn't match up I could build an arduino circuit which I can then calibrate with code to show the correct rpm. It's worth noting that I tried hooking the tach up to the IGN Ground on the ECU but still no luck with that. 

PS: I did an immo delete on that ECU, so if you are interested I can write up a tutorial!
Reply
#2
ive no idea at all what tachometer you bought or are making, guessing 1997 stock cluster.
the j18  on up all have 2 to 4 coil packs. COPS,  (1 cop split to 2  or 4 cops) varies world wide how down by year.
no single coil to use a RPM (primary minus lug of coil) now.
I have no drawing on j18,  but one.


a pro grade store bought tach can be connected to any coil and set to work, right, 1 o 4 coil setup.
now a word on cops.


are you asking how to make stock cluster tach G16 97 factory tach.
the stock tach is very crude, analog tach, it is just  signal intergrater,  crude as can be.
and only works with the brown wire G16, coil supressor module and only on single coil  engine, and  has back EMF of 50v. spikes seen on a scope.
The 50v is noise, back EMF the signal is IGNITOR output;.

to make stock tach work like this takes  a 1 to  4 up-scaler, that is 1 pulse in sets 4 pulses out. a freq. multiplier.  x4
and ignore the 50v.
here are my signals  .
coil minus, G16 (1 coil spark engie)  0v 15v(gen voltage) 0v repeat is all it take, the back emf is not needed to make the tach work IMO.
scope.
https://fixkick.com/look/coil-prim2w.jpg


the stock tach simple measure the hit rate to 0v I think, coil charge time  is integrated to make needle amp go higher,  dwell time , times hit rate. 4 hits per 2 crank turns..

[Image: tacho.jpg]


the ignitor was moved off the dash board engine side to inside each cop in 1998 in think. j18
so this signal is different  from 1996/7 raw COP engines. not sure yours.

I do not know how your donor car rans its tach, nor methods.
http://www.fixkick.com
Reply
#3
arduino
1 cop steal , input signal as input 1 to ARD. (is 4 times too frequent)
run x4 freq scale up algorithm. if possible IDK if is. ill look that up later.
set output to 0 to 12v square wave
and SYNC'd right.
makes sure output of arduino is polarity sync'd correct, 0v means coil is charging and 12vdc coil is dead time.
this make take a 5volt out to 12vdc output driver, to do that, arduino is not 12vdc device.
but is easy to do with one transistor and 12vdc supply inputted, then corrrect sync polarity to get this right./

the stock tacho is just the time it takes to charge 1 capacitor inside the tach the more hits of 0v the faster it charges and needle goes higher,
0v = coil charging, or dwell , old skool.
http://www.fixkick.com
Reply
#4
no tacho schematic but when I look inside the input is transistor inverted
so 0v becomes 12vdc and then via 1 resistor charges the integrated cap. this in turn, is connected to one more transistor that runs the tacho analog meter.
IIRC there is only 2 transistors there simple as heck.
last looked 10 years ago. but this is my memory from then.
http://www.fixkick.com
Reply
#5
PRT correct
and duty cycle correct, for sure 0v. out.
and 4 times input. 1 coil to 4 coil conversion.
and sync polarity correct (0v = dwell charge )
50v spike not needed or wanted.
0 12,0v square wave.
http://www.fixkick.com
Reply
#6
the new j18 cops if using those IDK, here is one example, but not sure yours is same.
has 0v 5v signals that are inverted, seen here best i can tell.
5v turn on coil for dwell charge.
i have no idea what year any Baleno did this change, 1998 ? or otherwise
1998 sport here, if you tap one cop here, it needs 1 to 4 freq upscaler. to work.   The ecu has series resistor driving this line. not sure its value . needs scope seen.
on some J engines there are 2cops in 1 pack and only 1 transistor drives both, for wasted spark application.

[Image: 98COP-unique.jpg]
http://www.fixkick.com
Reply
#7
Wow! Thanks for the quick and massive response!

I guess I should have mentioned:
-The engine has 4 COPs with a separate ignitor (only 2 wires go into each COP).
-The arduino circuit I built utilizes the transistor method that you mentioned. When the out pin is HIGH the transistor is on and sinks current to ground through a resistor, when LOW, the transistor is off and you get current going through the output of the circuit. I'll try to draw it up later today.

A little more on the arduino sketch. I coded it to sweep a range of frequencies based on the output of a pot that I had wired up. It was about 2 months ago, so I don't really remember the frequency range, but the fact is that the needle did not move. I don't have a scope to test the output of my circuit, but I have another arduino nano and I've heard of people building scopes with them, so I might try it.

Once again, thanks for the huge help!

I really wish that I had access to the donor car, but my dad bought the engine from a scrap yard and the old fart there just cut off all the wiring from the ECU and didn't even give him the immo box and key (hence the immo delete).
Reply
#8
(08-31-2020, 07:40 PM)maroka Wrote: Wow! Thanks for the quick and massive response!

I guess I should have mentioned:
-The engine has 4 COPs with a separate ignitor (only 2 wires go into each COP).
-The arduino circuit I built utilizes the transistor method that you mentioned. When the out pin is HIGH the transistor is on and sinks current to ground through a resistor, when LOW, the transistor is off and you get current going through the output of the circuit. I'll try to draw it up later today.

A little more on the arduino sketch. I coded it to sweep a range of frequencies based on the output of a pot that I had wired up. It was about 2 months ago, so I don't really remember the frequency range, but the fact is that the needle did not move. I don't have a scope to test the output of my circuit, but I have another arduino nano and I've heard of people building scopes with them, so I might try it.

Once again, thanks for the huge help!

I really wish that I had access to the donor car, but my dad bought the engine from a scrap yard and the old fart there just cut off all the wiring from the ECU and didn't even give him the immo box and key (hence the immo delete).


great, input.
yes the Arduino comes on many forms, and the 3.3v i/o version is no good. mostly, due to cars are 5v or 12v to15vdc. and not wanting to blow up Arduino.
I guess you are  talking the Arduino i/0  outputs using totem pole drivers pins? (programmable I/O is what is inside)

the way to test the speedo cluster tachometer pin is use any pulse generator to test it SIG GEN, square wave 1000Hz or so range.
set it to same rate as say 3000 RPM ,  (4 pulses per 2 crank turns) for G16 cluster (one coil spark is what G16 uses here, with a real distributor. then)
to do so,  generator output would need to be 12vdc square wave.
and see the tachometer work on a bench. test.

this test proves 2 things, tacho still good after 23 years and theory is correct.  12vdc square wave at (4 pulses per 2 crank turns (4cycle engine rules and one coil rule)

ok a COP that runs on 12vdc good,  that is more easy, large signals but the bad side is EMF (100v?) will blow any Arduino to H3LL.  so we use the BASE pin like E113(5) is safe.

IDK the base drive signals,  say on E113  is it 5v or only 0.7vdc. IDK  as scope answers that.
Arduino UNO (orig).  with 40mA output limit on any output digital pin.  is very limited , hardware wise. (a buffer cures that ) for sure to drive the tachometer.

Suzuki sport j18 has fire wall ignitors. as you can see blk/white is ganged 12vdc power, key on, and the  2 pins E112 and E113 are prime to usage.(1 example)

the E112 will nasty EMF , inductive kick backs on that pin, even 100v & no Arduino   will survive this but E113 is safe, I also bet the ignitor drawing below is not accurate
in that the snubber clamps are missing and there may be base resistors here for example E113 ,  Suzuki loves to use simplified drawing making work like this very hard.



Solving this takes a scope or luck with any DMM.
one thing to do is use the DMM diode test on (red test lead to) pin 5 E113 to , pin 3 ground E113 set to diode test and the meter shows MV drop. This may show you if there are hidden resistors on the BASE pins, but real scope makes all this so much more easy. see the signal and knowing that how to scale it in to the Arduino correctly.


[Image: 96-97Cops.jpg]
other unknowns here,  do coil 1 and 4 fire at the same time,? as in wasted spark mode, I have never tested for that. but is  factor on all cars to consider
and this is not too important as all we care is just  to use  1 coil input as the tach signal. (and does fire 1 time per 2 crank turns and not 2 time as in wasted spark mode)
lacking a scope LT/GREEN wires are same BASE drive,
 
here is a list of goals.
1: not blow the Arduino UNO to hell or worse the 3.3versions.(DUE or XERO) the brown pin 1 wire above sure will. (Q1 lets call coil 1 , transistor Q1 for easy refr.)
1b: not allow the mod we design and make, to cause Ignitor to fail ever. and engine die.
2: use MOSfet transistor buffer from Arduino  output pin (40mA rule not violated) (vast exist, even 2n7000) to the TACHOMeter input pin, problem solved easy.
3: The Arduino needs a good but safe input from say E113(5) pin the BASE drive signal may be safe for the Arduino input pin;.  only scope knows for sure.
4: Arduino INPUT needs  below 0.3vdc to be a logical ZERO, and above 0.6vdc is a logical one. the above igniter will not work here. but is easy fix if too high-v. use a voltage divider resistors to lower the signal levels or again a buffer transistor to do the same.  (ViH and ViL corrections)

the problem here is 2 fold, no scope sure, and no schematics exist on any ECU driver for pin 5 above and inside PCM(ECU)

We do not know the ECU drive single levels or currents to pin5. knowing this makes the mod easy, but we do not have that information.
I suspect that the drive signlal is a bipolar transistor(s)x4 and 1 series resistor to pin 5 (location of Resistor  unknown but a scope tells you this !)
say 500ohms and 0/5v drive to that resistor giving only 0.7v drive to the pin 5. not 5vdc.
if say the scope shows a 0/5v signal swings, on pin 5 then then the igniter has its own 500 ohm (or so) resistor on that pin or the drawing is a lie (common in car worlds) and is a ignitor MOSFET that loves 5vdc inputs and 2v threshold too boot. SEE?

the key here is to know that many drawing on cars are super simplified,not accurate,  making cars not easy to fix or work on. ever.

first the level in and out of Arduino must be good, not too large. (no damage allowed by us)
and the freq scale'r must be 1:4,   example, 1000 rpm in is 4000 out. (4 coils are 4 times less pulses than one  coil.)

we also do not know how the J18 cars ECU provided a tachometer on that car, at all.
  my guess a dedicated pin and not documented at all. other than pin # /color wire.
most cars now are digital controlled, not analog at all. this path. (or sends serial data from ECU to cluster)  all this evolved fast.


here is the very old suzuki 1 coil ignitor the books never tell the truth , this is what I found.
even my drawing is not very good, I was un willing to smash up this ignitor to find all parts inside, the DRAIN pin , I am sure has a snubber there to prevent Q1 from EMF coil kickback and destruction a clamp diode is there IBET.  There is resistor inside, this is so if you remove the A12 pin side the coil does not turn on fully and burn up the ignition coil the 630ohms there prevents burn up, of coil.  No coil allows 100% dutycycle, or burns up. the dotted resistor in read is test resistor.
The transistor is not BiPOLAR it is MOS (n-ch). there is never any BASE current in either polarity just 650 ohms.
91kick,TBI ,1 coil car: A4 pin on ECU on 16verion version.  the below is how to test coil and ignitor as  pair, for testing spark good or bad.(A12 unplugged)

 
[Image: igniter1a.jpg]I show this so you can imagine what might be in yours. IDK yours. mine yes.
top sig is COIL Minus. at transistor turns off, coil makes HUGE  spark.  my $50 scope. (rebuilt by me)with LED black lamps.
the spike there goes beyond 80volts,  not sure how high, but dont care, it is way too high for many devices, for sure Any Arduino. cas= cmp cam sensor.
[Image: cas-1000rpm-2w.jpg]

more HG = more vacuum (Hg= mercury column vac)  spark advances on HG. changes.
http://www.fixkick.com
Reply
#9
Arduino inputs. i/0 pins

'[Image: avr_io_current_limiting_resistor.png]

5v version of Arduino not 3v,  see D1 ? do not blow up D1.
to not blow it up use R1. outside The Arduino.
over 5.5vdc is not allowed, or it burns up the inside of the CPU.
same goes for inductives kick backs, EMF and blowing up D2.
always clamp EMF at the source and not allow it to enter the Arduino pin. (be it relays or any coils of any kind, COPS, Injectors, solenoids, all make huge EMF kicks)
this is the first steps in Arduino use. do not blow it up, rule 1. but yes resistors work great here. (I like not having the 100v spike at all, best)

now the topic of using a cheap cpu as  a frequency mulitplier, 
you need to use the input as interrupt.
The CPU is great at division,  or divide by 2, easy or a counter that counts down  to say 4 and sends out 1 pulse  per  4 inputted. (and wrong and backwards, no good this)
but counters do not multiply at all. they divide. 1,2,3,4 bang, we want 1,BANG BANG BANG BANG, .(x4 not div by 4)

with an Interrupt (program sketch) we can fire off one interrupt service routing per one coil firing then the routine (sub) sends 4 pulses outputted.
hitch 1:
(wasted spark engines,(cops) are like 2 cylinder engines. yah. 1:2 ratio. (I sure hope yours does not fire cylinder 1 and 4 at the same time wasted spark mode) as some do. but most those have only 2 coil packs.
I do not know any other ways with Arduino to do this,  below link takes this....
https://forum.arduino.cc/index.php?topic=118649.0


NEXT UP CHIPS:
but gee, slow speed, freq multiplayer chips are sold. (or made up as set)


other ways. (look ma' no CPU no programming)

https://bestengineeringprojects.com/freq...r-circuit/

https://swharden.com/wp/2016-08-31-ics50...ultiplier/

one good thing is the output is low frequency and easy 1000s of HZ only.
and simple square waves not complex analog waveforms here needed, it is measuring the rate of  COIL dwell hits. (tacho)
http://www.fixkick.com
Reply
#10
my scope are here, besides real modern 350mhz scope.
old and restored.

https://fixkick.com/Tektronix/index.html


a good old scope can be had for $50 and ways 50 lbs, and is called boat anchor by some now but is 100%solid state , and works like new. just heavy.

any scope works, real, they have 300 to 600vdc max voltage limits inputs, unlike cheap Pico scopes blow up over 5v or 12v or 20v. sad that.
a real scope will not blow up, on any car, (not directly to HV spark leads but I do have HV probe too and does work there to my scope)
the analog scopes 100mhz are vast models, they made so many at that spec, and is overkill for this car, but is cheap to buy used.
one that still works sure. TEK is the best IMO, I can get it to trigger on anything, so unlike other brands (HP is good too)

we called it a wiggle watcher. in the day. avoid all tube scopes. too old. (a guy that love tubes saying that, have HAM gear here with that. )
http://www.fixkick.com
Reply


Forum Jump:


Users browsing this thread: 3 Guest(s)