CamB

Thanks very much. Will proceed to testing the sensor. Mechanical advance is pretty vanilla*, so will try with that when the time comes. * According to spec. Should probably test that too as now have an adjustable timing light.

Looking to convert a Ford 4 cyl Twin Cam, with a 36-1 crank wheel. The distributor in it has a Lumenition optical ignition module, which has a three wire optical trigger that the Lumenition support people tell me: "There should be 7.0 volts going to the red wire on the optical switch. When nothing is passing through the optical switch, the voltage on the blue wire should be 2.7 volts. And then drops to around 1.0 volt on the blue wire when the chopper passes through it." Third wire is ground. The plan is to remove 3 of the four chopper blades leaving a single one to provide a cam sync. I will try the sensor on 5V supply, but would expect that it will still have a low value that is a fairly significant offset from 0V and a high value that isn't that high. Will this be an issue? Bonus questions: Does the new (to the G4x) MAP Level Trigger 2 sensing work well? It is a moderate compression, higher than average duration cam 8v 1.6 litre engine on ITBs. The plan is to retain the distributor, and I wasn't planning to lock the mechanical advance. I figure as long as the cam sync is a long away away from the missing tooth on the crank there wont be an issue getting sync, and if the distributor still has mechanical advance that the rotor phasing will be better than with locked advance. Am I on track with this or off in the woods?

Thanks @Adamw and @Ducie54. I actually think I'm fairly happy - I know my test wasn't perfect and its not ideal that the flow might be different, but that is the flow I experienced... My key goal was getting a deadtime for the injectors with resistors and confirming they sprayed nicely when driven like that. So this topic has any relevance, I've attached a spreadsheet if anyone is interested. It has a blank worksheet and one I filled in (and modified) for the 803 injectors, plus the Motec data I compared it against. Injector test spreadsheet - Link forum.xlsx

Thanks. I have thought about changing but can't really justify it. The injectors are clean, they aren't particularly high flow or difficult to drive, my spreadsheet shows they deliver fuel linearly in the planned range, and I have the data to make them work well. The spray shape (narrow cone) is suited to the engine too, as its only a 2 valve.

To satisfy my curiosity and give myself decent injector data to tune with, I've hooked up a test bench and run the two sets of injectors I have: Bosch 0280150803 (typically from a Porsche 944 Turbo), which I am going to use with resistors (ECU they will be used with is saturated only) so any deadtime data around is meaningless Bosch 0280150363, which is a very low impedance (0.7 ohm) performance injector which I understand was pretty popular in the past I ran the advanced injector test at various ms, picking a number of events and hz that are (I think) about 50% duty cycle and roughly a minute, then made a spreadsheet. You can work out the deadtime if you know the number of ms requested and the fuel delivered within the injectors linear range using the INTERCEPT function, and the SLOPE function gives the number of ms per fuel delivered so can be used to work out flow. The deviation from this theoretical calculation is the short pulse adder. I only did this at 12v (actually more like 11.7v) and base fuel pressure - I don't have a particularly reliable means of varying either. The results look like you'd expect (particularly if you extract the Motec M1 data and do the same analysis on that - results are fairly similar) - the injectors are linear down towards 2 ms test, then show non-linearity. The deadtime with resistors on the 803 injectors with saturated driver is longer than with peak and hold. However, there are two things I can't figure out: 1) Both sets of injectors show some crazy non-linearity at a certain point - in particular both had a point where the amount of fuel delivered went up when I stepped down to a shorter pulse width. Is this normal? It was slightly unexpected as Motec's data didn't show this. Note - the 363 did it at 1.2ms test and the 803 only did it when run as saturated as I didn't go low enough in P&H (I stopped at 1.3ms as that's really low for an injector of that size bearing in mind it includes deadtime). I guess I am not worried as in each case as both are where requested pw is below 1ms. 2) Both injectors flowed less than expected, and I am not sure what to use as the flow in the ECU for modelled fuel. The 803 injectors flowed slightly less (approx 5%) than Bosch spec / Motec data, which would be mostly explained if pressure is actually 250kpa (what the regulator is supposed to be) rather than the 270kpa a cheap ebay sensor measured (I will use proper sensors when each car is done). The 363 injectors were >10% too low (~8% if the pressure sensor is inaccurate or incorrectly calibrated), and in addition to Bosch data and Motec data I had them measured and cleaned a number of years ago so know they should be a bit higher. I guess I can think of a couple of reasons - looking for thoughts: weighing the fuel delivered and dividing by 0.745 is not accurate I used 4 amp peak, 1 amp hold and this somehow wasn't enough and limited flow the Advanced Injector Test is somehow not a good measure for flow

@Leonv33 - sorry, you are quite right. Frequency in to Clock and then the Data and Q5 should be tied together. Will edit my earlier post.

I have a problem similar to this too. I sort of solved it by reducing the number of parameters being transmitted and the update speed, then doing a lot of reading before writing to the screen. Not very satisfactory. I'm pretty sure I am just going to buy a plex usdm one day instead.

I've been advised that the CAN lambda will manage the heater by using the ECUs running information, so should be connected as per the wiring diagram. Would be worth waiting until someone from Link confirms or you should try the tech support help line.

Thanks Adam. Mine is twin plug and ITB and I was interested as that one you posted is only slightly higher than I would have expected for twin plug (max 28* under load) but slightly more conservative than I would have expected for single plug. I've just been collecting data points. Either way I will get mine tuned properly - I've been told that MBT is pretty clear on a twin plug 911 and occurs at quite low advance.

@Adamw - do you know if that is a table from a twin plug or single plug 911?

There are quite specific instructions in the manual - worth reading all of it as it covers: motor +/- a relay to provide the ECU with power for the above an Aux so the ECU controls the relay plus the accelerator sensor and TPS wiring https://www.vi-pec.com/techdata/ecu-manuals-v-series/vi-pec-manual.pdf

Did you re-check base timing after swapping polarity?

> I understand that pin Q5 is the ÷10 output that gets fed into the ECU's DI input. Correct but also tied to Pin 1 - data > +5VDC from ECU sensor power supply goes to which pin on the IC? VDD - pin 16 > Turbo speed sensor original output goes to what pin on the IC? D = Data - pin 1 Clock - pin 14 > Does the IC need power earth or sensor earth fed into it and on what pin? I'd ground to chassis, all the ones shown (Jam 1-5 and R for reset) and VSS (edited as wrong...)

The most interesting thing for me is that short pulse adder changes depending on your fuel pressure. I'd already worked this out playing with the Motec data, but useful to see it from a vendor. Couldn't tell you why the short pulse is different between the two ID sets though. If you press Alt-F11 in Excel you can unhide the worksheet with the data and have a look but its just a hardcoded table of short pulse by pressure.

That will depend on whether there is an output from the ECU which is useful to you, and it may rapidly use up outputs too. A better idea would be to try and get CANBUS working (adding a CAN shield to Arduino, or I think there are other Arduino-like options with CANBUS built in). I suspect this will give you a better outcome.

I believe it is "Second comms port CAN1 / RS232" (RS232 not for PC Link, for a serial stream).

Would also be nice if you had a way to measure load (MAP or TPS, although I am not sure how feasible either is with Webers as I've never looked into it - must be options) for ignition mapping.

Was just reading this the other day - the manual says it is required (page 12): http://www.linkecu.com/wp-content/uploads/2016/10/CAN-Lambda.pdf

@Davidv - thanks, I should have looked more closely at that! It was easy enough to change the code (I ended up using a custom stream rather than the generic one as I had trouble capturing all the incoming CAN messages if there were too many).

From my messing around with it, all the raw data in the Generic stream is the second parameter + 256 * the first parameter (in the pair). Interestingly when you set a custom stream the order is reversed... Is there a reason not to just use the standard convention? (I guess efficient coding?)

If stalling I would try: - using "idle ignition", so you've got an increase in timing when idle falls below target (there is an HPA webinar on this) - turning off transient ignition retard unless you are convinced you need it, and at least zero-ing it anywhere near idle There's a bunch of other stuff in the map (assuming the first one you posted is close enough to what you are using) that I am unfamiliar with but wouldn't necessarily expect (like quite significant ECT ignition trims, etc). These may not actually be affecting anything - its not possible to tell as your log doesnt have enough parameters. I can, however, see that your first map's ignition table doesn't match the logs (although, without the data, can't say why).

There should be more than that - a relatively recent PCLink upgrade pushed it out to 8x. Try loading the Monsoon example if you don’t have an ECU to update.

Quite hard to read the picture - any chance you can post a higher quality one?

This got me intrigued (can see the benefit), so I had a look at mechanical options. This came up: "Push to make" would be the launch activate button (a momentary), energising the coil and closing the contacts, which because of A-B being closed would remain closed until "Push to break" is open, which could be the clutch switch or I guess an auxiliary output from the ECU for gear change. The Launch input to the ECU would be on the second pole (C-D).

Yep, could do. Send me a PM when you are ready