ECU Tunes and Installs

[Jason Broadhurst]

005nap

[Jason Broadhurst]

So 60TJZ needs a tacho.

On a modern car, this is usually a simple job. The tacho itself reads a 5v square wave output from the ECU, or from an ignitor output. Most ecu's then give you the option to adjust the frequency/count of pulses per revolution so you can match the tacho to the engine even if the tacho is from a 6 cylinder and the engine is now 4 cyl.

Most old cars however run off a signal from the negative coil. This signal is the back EMF voltage from the coil, which is quite high and something an ECU can't generate easily. This voltage spike happens when the coil charge finishes charging and is discharged. The MA70 tacho needs 6 pulses from the engine per cycles, so you can easily connect the tacho to the negative of a coil on a 7M running a distributor, as that one coil fires 6 times per cycle. Once you fit a 2J, you can connect the tacho to the negative of a coil on plug coil, but it will only read 1 pulse per cycle, so 1/6th RPM. An option is to connect the tacho to all 6 coils, and use 6 diodes to stop crossing signals. This works fine, but is hard to wire and also not good for diodes to see 100+VDC so i have seen them fail and tacho drop by 1/6th every few months.

The other option is to use the tacho output from the ECU, and amplify it so that it can trigger the tacho and use the ECU adjustability to get the tacho reading correctly. To do this, I emulate an ignition coil even and isolate the ECU from the damaging voltage spike with a transistor. The following circuit gets this done, and then you can use ECU settings to adjust tacho scaling.

2k ohms resistor - limits the current on the ECU tacho output, in case the output isn't capable of outputting much current
Diode - stops any potential high voltage or current going back to the ECU. This can be any low current 50V diode
NPN transistor - like a digital switch. Takes the ECU output and switches the inductor on and off without being directly connected to the coil and potential high voltage. This can be any NPN transistor, it just has to be capable of the coils current.
coil/inductor - creates the EMF which in essence when switched off creates a high voltage. This could easily be a proper inductor, but I chose to use a relay coil because I had one handy and they are a good mix of resistance and inductance. 1H odd is enough.

The 2.43kV is the simulated max voltage. While the theoretical max voltage is infinite, this simulator takes it a bit far still. A real world voltage is upwards of 100's of volts still. The voltage spike can be seen here and measured at over 50V, easily enough to trigger a tacho. The scope has a 5v input which is actually just the calibration output for the scope. Knowing i have current limiting on the input signal due to the 2k resistor, I knew it wouldn't upset the scope cal signal. It generates a nice spikey output which I will feed to the tacho.





If your ECU has an aux out which can be set as a tacho, you can even simplify the circuit by driving the coil straight off the ECU.

I

[BigMuz]

Nice.

I'm putting a shift light in my sons sirion. Can't find a wiring diagram. What's the simplest way to find the wire that sends rpm to the existing tacho? Probe for voltage or frequency? I I trued frequency to ground yesterday but couldn't get anything that seemed right. The meter might be a bit slow responding.

[Jason Broadhurst]

Freq will find it usually, but it will be hard to confirm because it could be inj, ign or tach.

[BigMuz]

Thanks. To ground or to 12v?

[Jason Broadhurst]

Either will work it will just invert the signal but a multimeter on frequency won't mind.

[Jason Broadhurst]

Hilux on the dyno, making 300 odd HP at 7PSI.

[elfturbomax]

What fuel?

[garvice]

Even has the sun damaged front lights for maximum sleeper effect 😆

[I'm DJ!]

Hilux on the dyno, making 300 odd HP at 7PSI.

You could fit my house under that curve!

[Jason Broadhurst]

98ron. Flex sensor is installed, just waiting for new rods to turn it up

[burn is weird]

thought I'd make a post here about road tuning (or even dyno tuning if you prefer logs)

So I havent been able to get to a dyno so have been tuning my VE maps on the street with logs. My ECU is using a TPSxMAP VE model for fuelling that take into account air and fuel density, stoich ratio of current flex fuel mixture and uses a Target lambda table to work out how much fuel to inject. the important bits for this model are accurate injector flow setting, deadtime and linearisation details, and a TPS vs RPM volumetric efficiency table, along with a MAP vs RPM "4d" table as I have an ITB turbo engine. in this type, the VE of the engine can change with TPS, but also with Mass flow and exhaust backpressure at different boost levels. so you tune the TPS map on minimum boost, then slowly increase the boost to tune the MAP VE correction table.

If you are starting from scratch you'd build a TPSxRPM (or MAPxRPM if you're going that way) VE table that goes from 60% at closed throttle to and idle, up to 100% or so at WOT and peak torque rpms and above. thats a starting point.

then you go for a drive with closed loop fuel corrections disabled and log the following at a minimum:

RPM
TPS
MAP
coolant temp
total fuel corrections
Target lambda or AFR
actual lambda or AFR

I am using some software called megalogviewerHD (http://www.tunerstudio.com/index.php/megalogviewer) which is GREAT for analyzing logs. the paid version has calculated fields, log filters, a log display, a scatter plot and a histogram table generator.

you want to filter out any cold engine stuff, and anything that has significant fuel corrections so nothing is interfering with the fuel mixtures and any sudden TPS changes you can use the filters on the right hand side of the screen to do this (excuse the scaling, I have a 4K screen and windows object scaling doesn't like it). then create a calculated field "AFR Error" = 100*Actual AFR / Target AFR. set the X and Y axes to match the values in your ECU, and you have a % correction map! the color of the cells is based on the number of 'hits' for that cell that have been averaged, so the greener it is the higher the quality of the data. white cells might only have one sample, so not necessarily a reliable number.



put this in excel and interpolate missing values, or put 100% in the blank cells and apply this % correction to your VE table, do a visual sanity check of the 3d plot for any skewy outlier values and you're tuned!

whenever you go for a drive, log to a USB stick and analyise again, and over time your corrections should approach all 100% meaning no correction.

there are a few other ways of looking at stuff too. this becomes quite useful for finding correlations, visualizing knock, setting up boost control duties and EGTs VS ignition timing for example.

[Jason Broadhurst]

That software is amazing! It sure beats using matlab.

What sort of calculations can you do? I suppose it you need mega math calcs you can just do them in the ecu and log that.

[burn is weird]

That software is amazing! It sure beats using matlab.

What sort of calculations can you do? I suppose it you need mega math calcs you can just do them in the ecu and log that.

https://www.efianalytics.com/MegaLog.../formulas.html

have a look at that page.. pretty much anything you want. you can also reference samples using offset. for example, you can get a derivative(rate of change) plot by something like ([TPS] - [TPS-4]) / ([TIME] - [TIME-4]) which takes change between current sample and 4 samples in the past.

also if >200 then valid, else 0 kind of stuff. super powerful.

if you find your wide band has a fixed delay, you can shift the wide band value into the future to line up with the engine conditions that caused it. quick TPS stabs will show you this delay.

better to do the calcs here from the base data unless you need them in the ECU specifically (or have to super-sample for averages etc) for faster logging (the CSV logging in SEKUKU anyway) the USB logging with the 16 channels is a set data rate.

calculating rate of change of RPM (or road speed) is useful for ignition tuning because as long as the road is flat, it is proportional to torque. so you log a few runs, bumping up timing as you go then do a filter for 100% throttle values and do a scatter plot of RPM along X, rate of change of RPM on Y axis, and Z axis (dot color) as ignition angle, and you can see what angle provides peak torque at each rpm point. I have logic in the ECU that instantly cuts power as soon as noise within a crank angle window (-10 to 40 degrees TDC) exceeds the background noise by a certain value, and only resets when you go to 0% throttle as my failsafe. along with noise cancelling headphones, which plug straight into the adaptronic to listen for knock.

the log viewer software can also trail a live ecu CSV log so you can do analysis on the run, on a single continuous log file.

[taijohnsen]

I'm totally sending you my logs...