TimoL

Oh, I see. But can't really go wrong with RCM, they're the quality stuff. And yes, I'm still running those injectors. I had some minor difficulties with passing the emissions testing but that might have been due to a bad O2 sensor. I'm planning for a full engine rebuild & hopefully maxing out these injectors with E85. And most likely exploding another transmission

Hello @DenisAlmos did you manage to test the injectors? Would love to see your results.

Hello, I've been rather busy with my work so I didn't check this thread regularly. Thanks for all the replies. I checked the battery voltage one morning and it was about 14.5V after cold start, engine running. The ECU was showing about 12V. I have one 20 sec PC log that to took after about 20km drive. I tried various electrical loads by turning the lights and blower on or off. No change in reported voltage.I believe the reported voltage was wrong during the whole drive, it was running very rich whenever I had to stop or drive at very light loads. I also noticed that I can't use the search function for the battery voltage. -Timo batt.llg The ECU config file I posted earlier contains the settings that caused this issue. Maybe they can test using it? Anything else I can do to help? No problems since I turned off the logging, but I think it's a bit too early to tell for sure. -Timo

So here's the full story, Few days ago I started having problems with the car running really rich. First I thought it has something to do with MAP sensor or FPR but I couldn't find any fault. Then I went to check ECU readings and I noticed battery voltage was stuck to around 11.5V. This would explain why the car was running too rich because the ECU adds too much injector dead time to PW. I'm 100% sure the alternator or battery is not at fault here. There was zero change in the voltage reading when I applied full electrical load on the system vs. load at all. This has been happening for a few days now. It's not the same reading it gets stuck to. And it's always on cold starts, but I haven't done that many warm starts. If I shut the engine down after a while and restart, the voltage reading is working again. I've attached a log from this morning when I started the engine and shut down a couple of times. After I saved the log and started the engine again the voltage reading started working. Changes I've lately to tune is to enable internal logging. The voltage reading got stuck again tonight. When I disabled the internal logging and saved the settings it started working right away, no restart was needed. I'm not sure if this issue has something to do with the logging or if the settings save triggered this fixing effect. Any thoughts? I can provide more logs or do testing if needed. But for now I've turned of the internal logging to see if it helps. -Timo battvolt_starts.zip currentConfig.pclr

Well, that's good to hear. I might go DBW after all.. I assume most of the OEM solutions have a deliberate lag added to them (something to do with emissions & economy)?

Good point. E-throttle could be a solution but I don't want more lag in throttle response

I'm looking for a cruise control unit with an ECU control input. It would have to be a drive-by-cable model. I believe some WRX models (or just Imprezas) had CC since the throttle body has support for a second throttle cable. I want to use it as a cruise control and to increase the throttle opening for cyclic idle, thus requiring the ECU control input. I guess the actual cruise control could be done with the ECU as well? Do you folks know any cruise control systems that would work for this purpose? Edit: My car is 1999 WRX GC8 -Timo

Is it tuned for gasoline too? I know this might not be the solution you're looking for but have you tried to add some gasoline to the mix? I found that running E65 - E70 helped tremendously with the start up. It was around -5 deg Celsius when I tried this. Haven't had the change to test more since I blew the gearbox.

I can't seem to be able to figure out how to setup a secondary open loop AFR target table which I can switch to whenever I want and disable the closed loop lambda at the same time. I'm going to be running a switch connected to a DI. I've found the lambda overlay table but I suspect it cannot be used for this purpose. Is there any way to achieve this? I could run a secondary fuel (VE) table but hen I would need to maintain two VE tables. And how about disabling the closed loop with the switch?

I tried to adjust the injection timing as suggested but I made a table that has ECT as the Y-axis. When the engine is warm a timing of about 400 works best in the idle range. I have verified this by observing the AFR and transient response. But things seem to be very different when cold. When I dropped the injection timing by 90 deg I noticed a drop of lambda value from 0.86 to 0.77 and the transient response was improved. The PW was constant. I should try the ign cold advance as well. I pre viously had good results with my motorcycle (e85 too) wiith MS2. Here's what I think: When the engine is warm the fuel is injected on the hot intake valve (& port). This vaporizes the fuel just before the intake valve opens. But when the engine is cold, injecting to the closed valve does not help the vaporization, instead the fuel sticks to the valve & port walls. So the best way to get a biggest amount of fuel in to cylinder is to inject on the open intake valve. Smaller value means that fuel is injected later since its "BTDC", right ?

Thanks, I'll try if this works me Yes, I will definitely need different Warm up tables for Gasoline & E85. I have no plans on driving on E85 in winter tho .

I'm not sure if this topic should actually be in the tuning section? But here's the deal: I've switched to E85 for now and having difficulties tuning cold corrections. The starting is okay and it runs great if it just let it idle. The real problem is when I attempt to take off. Even if I apply just a tiny amount of throttle or the engine speed drops for any reason the engine will most like stall. I believe this to be because of change in manifold pressure which affects the evaporation of the fuel. Fast throttle applications actually work very nicely, the accel enrich seems to tackle that very well. I'ts not even very cold, about 15 Celsius (Finnish summer, yay ). I bet that at colder temps the car would not be driveable unless it's been let to idle several minutes. When engine is warmed up none of these problems are present. Fuel and ignition map have been tuned up to 180 kPA absolute (had to stop there because of fueling problems but that's another story). I've attached a log and the current tune. This was a cold start (15C temp). I actually managed to stall the engine once. The logs don't show this but I'm 100% sure it was due to AFR leaning out. I took off lightly and then tried to replicate the problematic conditions by shifting a lot and decelerating & accelerating (see "Accel_and_decel.PNG"). I'm not sure if this can be tuned just by acceleration enrichment, unless switching to MAP based? To combat this, I think there would need to be a temperature based Accel Hold & Decay values. I was thinking of utilizing the Warm Up Enrichment table with the MAP Delta as load axis. But when looking at the logs, the MAP delta variable seems to hardly change at all? -Timo CurrentTune.pclr Log 2018-06-30 2;40;04 pm.llg

That's exactly what I thought. Maybe I could cobble up an Arduino based voltage logger and connect it to the CAN bus. I do not want to compromise other sensors since I'm running a fuel pressure sensor and differential fuel pressure compensation.

I would like to measure the voltage at the fuel pump terminals and be able to log it with the ECU (Link G4+ WRX ver5/6 plugin). Is this doable with any of the free inputs?

I also ran some tests on very low pulse widths to find how the injectors behaved in the non-linear range. All of the injectors have a similar curve for low PW linearity but the peaks reside at different ms. The peak will be flattened (if not removed all together) when you calculate average for all of the injectors. Since the low PW adder is same for all injectors you would need a low PW (and dead time) matched set of injectors to have full benefit from the low PW calibration table. But since these are 1000 cc injectors, I expect to be operating in linear range and the low PW calibration will not be needed. Interesting results nonetheless. I could not wrap my head around how to calculate the low adder table but I think what's all that about. To visualize it: The blue line represents the expected flow per millisecond and red actual flow. For example to find out the needed PW addition at 0,625 ms you would measure the distance from "the expected flow" line at that x-coordinate to the x-coordinate where the actual flow has the same flow value. I've marked these distances with green lines to the chart. Is this how the short PW adder works?

Finally had the time to finish the tests. Results are in the attached spreadsheet, feel free to use for your own measurements. It's not really clean and has a lot of unnecessary data, I just like to calculate things... Here's what I found: The dead time has some variation between injectors, lowest being 0,715 ms and highest 0,812 ms (at 14,0V, 350 kPa). The average was 0,763 ms for all injectors. I also measured the dead time using a MS2 based ECU and got an average result of 0,80 ms (also at 14V, 350 kPa). Overall, the dead time is pretty much what I expected for these injectors. I only used one injector (Inj4) for measuring dead time at different voltages. I only went down to 10 Volts on the test bench and extrapolated the low end. -Timo deadTimes.ods

Thanks for sharing the link David, that's a great write up BTW. Since I will be using the Link ECU for driving the injectors and the voltage reported by the ECU for cal table x-axis, I believe it should be accurate enough. Also, the same fuel pressure sensor (connected to the ECU on both cases) will be used for the testing. -Timo Yes, they should be the same. I'm actually more concerned about the flow rate differences between the injectors. The seller claims that they are flow matched but I paid 140€ for a set of 4. You know, you get what you pay for... I confirmed the wiring by measuring those pins from stock ECU while it was powered on. The wiring seems to work on the test bench. I could not do any tests yet, because the test bench fuel pump was shot. I believe it to be because it was used with ethanol and then stored dry (=corrosion?). I will be posting results soon as I have some. -Timo

Thank you for your input Adamw. I am going to use the Link ECU for the dead time testing. I took a look at PCLink manual and it seems that I can test only one injector at a time and only using injector channel 1 because I need the advanced mode. Is it right? It's not really a problem, I was just hoping to be able to test all 4 injectors at the same time. Also I'm not certain as how to wire the ECU. The installation manual has at least some of the pins explained. Based on the manual I would wire the ECU like this. Is this correct? And how do I wire the injector +, just straight to +12V? -Timo

I just bought JECS 1000cc side feed injectors for my 99 WRX but the seller could not provide any dead time values. So I build a test bench for testing injector dead times. The setup: -Adjustable fuel pressure reg, 2 - 9 Bar -Adjustable voltage supply, 9 - 15V, 36 A -MS2 based ECU But I started thinking that maybe I should not do this with other ECU than the one I actually use on my car (Link G4+ WRX Ver5/6) with these injectors? I plan to get fairly accurate data for dead time and possibly short pulse width, but I have no experience of the latter. On the other hand, wiring the Link G4+ to my test bench should not be very hard. AFAIK only 8 wires would need to be connected: -12V (batt) -12V (switched) -GND -Inj 1, 2 ,3 &4 -Inj common (+12V) -Timo

I too had the same problem with the fan on MY99 WRX. I had to invert the polarity on Fan 1. ECU Settings -> Auxiliary Outputs - Ign -> Ignition Drive 3 - Engine Fan 1 -> Polarity = Low (was set to High as default)

Thank you for the fast reply. That's great. I was going to wire the +12 to a switched supply anyway since I'm going to leave OBD2 reader connected to the port. Br, Timo

I just a bought a PnP Link G4+ ECU for my 1999 WRX. I have a question about OBD2 wiring. Should I wire the OBD2 plug like this: Or can I use just use the " CAN2/OBD" connector on the ECU board which has the "+14V" output? I just want to use an ELM327 bluetooth OBD2 reader on the port. Is the ECU able to power the device or have I completely misunderstood purpose of the +14V pin? Thank you in advance. Br, Timo