Technical RPM counter

[nmrmak]

The engine uses wasted spark, so in one rotatation it should fire each plug twice. 1(4) 3(2) 4(1) 2(3).

And since there are two coil packs, i'd probably be looking at 2 pulses per rpm per coil pack.

Pictures (all ~1.5MB):
ECU with connectors: http://mrmak.deveta.co.yu/images/img_0093.jpg
ECU sticker: http://mrmak.deveta.co.yu/images/img_0087.jpg

EDIT: I suppose it's 4AF then.
EDIT2: Are there pinouts for it somewhere?

 

[randomizered]

No pin outs as far as i know for the 4AF, i cut off my connectors to fit megasquirt, I may have a colour guide somewhere i made up, I'll see if i can dig it out.

If you disconnect and check the LT connections though, 2 are the same colour, they are +ve, the other two are switched earths and are the same colour at the ecu end as the connector end.

You can remove the connector covering without damaging the connectors and see the wires from there.

There will be 2 pulses per rpm from one LT connection. The timing will not be spot on as they control the dwell of the coils, and will signal before a spark occurs, but you could use it to give a very close estimate of the rpm

Kristian

OT: are you Yugoslavian? You have fantastic english, no worries there

 

[nmrmak]

OT: Yeah, I'm Yugoslavian (although that country doesn't exist anymore, we're supposed to move on to the .rs TLD soon). It could be that my written english is okay, but you wouldn't want to have a spoken conversation with me

You keep mentioning LT, what do you mean by that abbreviation?
If it grounds the line for each spark fire, what's the signal level the rest of the time? 12V? 5V? Or will i just have to measure that?

EDIT: I'll have a go at an oscilloscope and the rest of the goodies on thursday, so you don't have to dig for that info, it should be easy to get, i'm just thinking about what stuff i need to do to get the signal to my microcontroller, so i can start ordering the right parts now.

 

[randomizered]

The coil has a HT (high tension) and LT (low tension side)

The coil is fed by 12V high current. This votage is there all the time the ignition is on protected by a 20amp fuse.

Until the LT side goes to ground, the coil will not charge.

When the LT side goes short for a period, current flows through the coil building up voltage in the HT side.

When the LT side cuts off the ground, the stored energy in the coil will be dispersed through the coil and spark plug.

The reason this can't be accurately mapped to spark/rpm is that the coil needs a to charge at a different rate depending on rpm, this is called the dwell period.

The coil also has a discharge period that needs to be calculated for as well.

I'm not sure how the stock ECU deals with this, but I'm sure that for each spark the LT side will be open then ground to charge coil, then open.

If you scope this, you maay be able to use it as an rpm output, the rising or falling edge will be very close to rpm/x

Kristian

Kristian

 

[nmrmak]

What do you think about using the crank sensor for the rpm indication? would that be a better option?

EDIT: How do sei sporting rpm counters work? using the LT line? I think i read it connects to pin 23, which is "Output Engine Speed Limiting (Fuel Pump) Relay Fuel Pump" or something?

 

[randomizered]

If you can safely intercept the the signal before the ecu, then yes, it will work.

It has 60 teeth and two missing teeth for sync, so can provide a resolution of 6 degrees per revolution

Depends if you can decode that signal and interface it tough.

Kristian

 

[nmrmak]

The max rpm i'm looking at is in the order of 7k?

 

[randomizered]

yes. The rev limiter on the mpi is 7K area. Maybe use 7.5K for a bit of headroom.

Kristian

 

[reddy4bed]

Would not reccomend picking up from crank sensor, the voltages are relatively low and need screening to avoid noise etc on line, also the 2 signals are opposite phase and would need to be put through voltage comparators to detect the crossing point (as the ecu does). Much simpler to get the signal from ign coil LT as said (pins 1 or 19), put a resistor in series, say 10k, strap this output to ground with a 5v zener to prevent any voltage over 5v from line getting into TTL.

 

[nmrmak]

I was thinking along the lines of optocoupling the LT line, same effect but i don't risk altering the signal much. Putting a 10k resistor in series and a 5v zener would modify the signal at the coil as well i think. (i actually need to draw a schematic, could be that you're right). The first step will be to measure the voltage levels and frequencies tomorrow.

 

[nmrmak]

I was looking today at the coil packs, and i can't see the high tension and low tension inputs. Maybe i got it wrong. There are two connectors, one on each coil pack, that seem to be low voltage (thin isolation) and there are 4 connectors, 2 on each, that go to the spark plugs. Are there any additional HT connectors, and are the LT connectors the ones that don't go to the spark plugs?

This system is a bit different from my old car, and i don't want to burn the scope testing it. My old car had a distributor with 1 HT input and 4 HT outputs that went to the spark plugs, but i can't see a HT input on the coil packs, just don't want to get it all wrong.

 

[allanhelen]

The LT is the input the small wires one plug per coil pack the HT is the output the two large leads to the plugs

 

[nmrmak]

Is there a HT input? Where does the high voltage come from?

 

[allanhelen]

no HT is output only the coil pack generates the HT

Unlike coil and distributer setups the ECU pushes the LT to coil pack to generate the spark rather than distributing generated HT

 

[nmrmak]

Thanks, with this info, i'll be able to hook it up to a scope tomorrow hopefully. The problem is that the scope isn't mine, so i can only use it when the guy doesn't need it.

 

[nmrmak]

Allright!

Tested my prototype today and it works. Shows ~650 rpm at idle, which is about correct. The scope shows ~75ms rev duration, which is okay too. Interesting thing is that i think that one coil pack actually makes one pulse per revolution, or well it at least gets one pulse per rev from the ECU. I will have to see through my software to see again if this is correct.

I will post the code as soon as i comment and document it and schematics as soon as i make them (i was working on a development system and added a prototype board to it). The system picks up the signal directly at the LT side of the coil pack, and uses a zener diode (thanks reddy4bed, turns out i didn't need optocoupling or anything) to limit it to ~5v (down from ~12V). Pictures coming as soon as i build the pcb, and start installing it. As i said before, i'll give you the code and the schematics so you can play with it, and quite possibly improve it. An analogue display could be an option, but i can't deal with that right now (uni and all)


Any ideas how to get the cables (not more than a few square mm) from the engine bay to the dashboard? I'm sure they pulled the original cables somewhere, but i can't see that spot. It's going to be 3-4 wires, 12V from the ignition, ground and the LT signal from the ecu, possibly from a ribbon cable, so it's not too thick.

Cheers and thanks for the help, i really really appreciate it, i know it sounds cheesy, but i couldn't have done it without you guys.

 

[cc1]

The system uses wasted spark so fires every cycle but 180 deg out of phase between the two coils.
The rev counter connector should be behind the middle air vents on a seicento.

 

[nmrmak]

Is it easy to remove the air vents?
Is that the lead to the LT side of one of the coils?

 

[randomizered]

No, what CC1 is saying (i think) is that most sei's have the wiring for the rev counter installed, but not linked to the pod, behind the dash. It may be a differnt signal, just one wire for signal, a power and gnd.


Don't know why i didn't think of using that before


Kristian

 

[randomizered]

https://www.fiatforum.com/cinquecento-seicento-guides/86286-temperature-gauge-fitment-gars4.html


those are the wires and connectors.