Jump to content


  • Content Count

  • Joined

  • Last visited

  • Days Won


Everything posted by cj

  1. In a lot of cars one restistor is in the ECU and the other in the dash as this is commonly the low>high speed can gateway. Sounds like you are replacing both items so you likely need to put a second resistor somewhere + the one in the link. To be sure, connect the link and everything else, and measure resistance between the 2x canbus wires pretty much anywhere. If you see 120 ohm you need to add in a second one, if you see 60 ohm, you've already got 2x in there somewhere so leave it alone.
  2. cj

    Using 4D fuel table?

    How about testing the "extreme" option first to see if there is actually anything to be gained from your time & effort. As Adam said, the whole fuel wetting thing will likely prevent you getting this perfect, but if you just set overrun cut timer to 0, and accept the problems from this for a day or so, you will see in your logs whether this problem is actually fixable with your engine. If this test behaves excatly as you expect, or you at least see a lot smaller rich spike then there is likely a combination of settings to behave the way you want and still perform well. If this test still shows a very similar rich spike, then you know that no amount of shortening the timers before cut or using tables to do the same thing will actually help. On a more theoretical note, all modern OEM emmissions strategies oscillate a bit rich & lean regularly because it has some sort of emmisions benefit vs holding exactly lambda1.0 all the time. Because of this, short rich spikes in your tune are probably "good" for your cat if there are similar lean spikes a certain loads, or at least the will benefit the emissions purposes that it's there for.
  3. cj

    Injector Timing

    As Adam said, you have to change the table to 3d before it will let you adjust the axis values. Then set up the axes with "none" on Y axis As long as you are using traditional you dont need to set the injector size directly, you effectively just tell it directly how long to turn on for, and you tune these number until the car runs right. The size of the injector and the dead times mostly just get hidden in the tune. The numbers in Kmanager are similar enough between 270/210/360 that in traditional mode it wont really matter which you use - assuming hondata has pulled OEM numbers from an ECU. I've never actually confirmed where they got them but no one seems to have any better numbers to use. The downside to this is if you change injectors you need to re-tune your fuel table.
  4. cj

    GPS Speed Sensor setup

    It sounds like the output is very similar to what you'd see from a traditional gearbox sensor. connect it to any free DI, the configure the DI as LR wheel speed frequency. Try with pullup Off for a start, and if you see 0 all the time, turn it on a re-test. Then set the calibration value to about 500 - I think that will roughly match a 16000ppm output. Then set up logging, go for a drive at 2x constant speeds (50kph & 100kph is usually good), hold the speedo at this speed for say 10 seconds each. If you have a friend, update the calibration value until the ECU speed matches. If you dont, look at the log after and calculate the % difference between the ECU speed and the GPS speed, and adjust calibration value by this %. [edit], at 16000 ppm you will need to connect to one of the higher frequency DI's. Officially the lower ones max out at 500Hz, which will be about 200kph for you. Depending on a few other things, you may have it work above this frequency, but given you have a thunder, just use a high frequnecy input.
  5. cj

    Injector Timing

    Are there any numbers on the injectors? if not, what is the full code/model of the engine? Are you planning to run modelled mode or traditional? From memory the dead times in Kmanager are very similar between the 310's and these. If you are going to run traditional mode anyway, it probbaly wont matter which you actually have, as your master value will just be 1ms higher or something.
  6. cj

    Using 4D fuel table?

    Sounds like you just need to drop the activation delay for fuel cutoff you can also just set the *very* low values of your fuel table to be really low numbers - Assuming a MAP based table, the 10-20kpa lines at >3k rpm are probably only ever hit on overrun.
  7. cj

    Injector Timing

    All stock K20 & K24 injectors are 310's. B & D series were 240 or 270 depending on model. Are there any numbers on them, or do you know for certain which engine/chassis they came out of?
  8. cj

    link pressure sensors

    Am I understanding this right... all sensors are plugged in while you are testing, and you are just backprobing the pin/joint in the wire? [pin numbers may be wrong below] pin 1) from 5v ecu ref, shows 5v at the sensor pin 2) from ecu signal ground, show 0v/ground at the sensor pin3 - the signal wire) show 0.2v at the sensor and 0v in the ECU? Here are the things I would check: 1) make sure you have ground pin2 and not a floating pin - unplug the sensor and with the ecu on, check voltage with a multimeter ebtween pin 1 & 2. This verifies the wiring to both. 2) have you got anything else connected to that sensor wire? a factory dash gauge? another sensor? anything at all? with the sensor unplugged does the ECU see any voltage? You should also check the resistance between the signal ground and this signal wire, and compare this to the value of signal ground to any other unused anVolt wire. The values dont matter, but you want to look for differences that suggest there is another path to ground via a gauge/short/broken wire/something else other than the known good path through the ECU internals. 3) double check that you arent reading the connector pinout from the wrong side (eg male vs female). From memory, if you wire these sensors up to the wrong pins, you can get ~0.2 volts showing on what should be the ground pin or something like that. 4) check the part number on your sensor - some are rated in current output instead of voltage, or at different scales than the usual 0.5-4.5. Unlikely if you bought it from a link dealer but easy to verify. 5) if its easy enough, disconnect the signal wire near the gauge, you can then measure the output voltage directly from the sensor without any chance of the ECU/other wiring pulling the voltage up or down. You only want the 5v & ground connected for this test. 3x AA batteries stuck together can usually create a close enough to 5 ish volts to test a sensor if you cant get into the wiring easily. Remember to probe sensor output with + and the battery negative with - on your meter. Chassis ground isnt connected if you are running from a battery. As per the calibration, these sensors output only in the range 0.5 to 4.5v. At 0 pressure, your should still have 0.5v. Anything less than this means you have a wiring problem or a sensor fault.
  9. cj

    link pressure sensors

    Are you sure the pinout is correct, and is it a genuine TI sensor? There are some chinese clones where the pinouts are 1 pin off. Best way to check if its a wiring problem is to plug it in, turn the ecu/key/etc to ON, then press f12 on the laptop to bring up actual voltages. Click on the "analog" tab and check if you are seeing ~0.5v on whichever AnVolt input you have it wired to. If you see 0v, 5v, or something else obviously wrong you have a wiring issue. If you see ~0.5v at atmospheric pressure, then apply some pressure to the system (turn on fuel pump or whatever), and check that the voltage jumps up some reasonable amount. The calibration for these is really simple 0.5v = 0psi & 4.5v = rated psi (in your case 150psi). Set a custom calibration to these numbers, then assign this calibration to your anVolt.
  10. Can you post your config and a log of this happening? could be a lot of things
  11. same as k20 coils
  12. Try this one. You had switched it from traditional mode to modelled mode, but the fuel tables for these are completely different. All I've changed here is: swapped out the fuel table to one from a modelled fuel base map - it should be close enough to start but will still need to be tuned Disabled IAT correction - not usually needed for modelled fuel put in the ID1000 deadtimes and short pulse adder times. I would also suggest you run through some sanity checks for all your sensors before you start it. Looking at the last seen runtime values, the MAP sensor was reading 14kpa higher than atmospheric even with the engine off. The IAT & MAF IAT are differnent by nearly 10 degrees, and you TPS is reading 61% open. These all might be just quirks of the way you are testing it so far, but better to be safe than sorry. 2063677987_SubaruWRXV9G4XtremePlugin-model.pclr
  13. What injectors are you using and where did you get the deadtime + short pw adder values for them? There might be something bad around your o2 sensor as others are suggesting which would throw off any feedback-based tuning, but looking at your injector numbers, there are some pretty savage numbers in there and if they arent correct, it would be throwing your AFR around a fair bit at low load as the ECU will have completely the wrong idea about how much fuel its putting in. The short pulse width table especally i'd want to verify - its all negative, which seems really unlikely and as an example, at the ~1.4ms effective shown in most of your log, its removing ~0.3ms from the commanded injector time (over 20%), and based on your pw adder table, you are well into the non linear range at this point, so if that table is wrong, its going to be making it behave REALLY wrong.
  14. Hi Guys, Regarding heat soak, you can add some values into your IAT table to compensate for this. Even on modelled fuel mode where IAT correction isnt normally needed you can enable it, set one axis to TPS and the other to IAT, and then just add in some ~+5 to +10% values in the 30C+ ranges, but only at <10% TPS. Leave the rest of the table zero'd out as you arent looking to use this for IAT correction during normal load.
  15. cj

    Trigger1 err

    sometimes the trigger scope will show some noise or signal degradation, or signals wired backwards etc, even when its behaving normally, and from this we can see that if it was just a little bit worse you would have a problem. I'd suggest leave your sensors as they are (your last instance of this problem was with the same sensors you have now right?), warming the engine right up, then hold it at 5500-5700rpm and run a scope capture.
  16. cj

    Trigger1 err

    Sounds like you should run a trigger scope capture while holding rpm at just over 5500, maybe we will see some noise in the trigger signals
  17. This here says its a compatbile replacement part and expects 3.6ms - i'm guessing at 14v https://moneashop.com/electric-system/136045-coil-IC1197.html
  18. cj

    Impreza v3-4 Wheel speed

    You probbaly dont have ABS sensors on that car to tap into so you wont get individual wheel speeds. Your speedo will be driven by a gearbox sensor. connect the vehicle speed sensor in the gearbox up to any digital input, and configure this as LR wheel speed. Leave the wire connected to the speedo as well, dont just cut it. then under chassis & body, set driven wheel speed source to this DI. Also set non-drive wheel source to the same DI.
  19. depending on the coils you use you can either fire them direct from the ECU or you may need an external igniter. The coils from the more modern k20's fit onto B16/18's pretty well and dont require any igniters as an example.
  20. There are a few things here that look a bit wrong here 1) your dwell table is 9-10ms across the board. I'm not familiar with the coils on this engine but 9ms seems really high (typical would be 1.5-4 ish on most coils these days @ 14v) 2) your ignition timing is quite low. At low rpm (say <2500) there is that low speed pre-ignition issue on many DI's, but above that, I would think you should be more in the 10-15 range than your current 2-4deg (even before you lose some to knock control as pointed out by Tim D). As long as you have some knock detection happening outside of the ecu to keep it safe (ie whatever detection tool+headphone the dyno operator has), i'd figure there is power to be gained by adding more timing. Have a look at the stock map for a rough guide on reasonable numbers.
  21. the MAP limit table values are linearly interpolated between the entreies you define. So at 103 degrees you are 30% between your 100ECT value (260kpa MAP) and your 110ECT value (150kpa). This gives you an effective MAP limit of 227kpa at this temp. Minus 15 for the start of limit engagement number and you are at 212kpa for the limit to kick in, which is exactly what you see. I suspect that ECT is recorded in whole numbers internally, so you are either at 227 kpa limit, or at 238kpa limit (102ECT), and so it seems to come and go randomly, because its only 1 deg ECT difference. Quick fix is add in a row at 105ECT which still allows 260 MAP. Also, whats up with your IAT sensor, it seems to read a constant 75C for that entire last run - it doesnt change with boost level even 1 deg, but 75C is kind of high to be ambient? If you want to make the limits a little easier to tell apart without changing what you are logging, change one of them to -3 deg instead of -2, so its more obvious in a log. Really you should just change your ECU logging to include more parameters if you know there is an issue hiding in there somewhere though.
  22. DI's dont have calibrations, they are just on/off/frequency. If you mean analog inputs, what else would your input value be other than ohms (resistance) or volts? do you have a sensor that output varying mA? you would have to either replace the sensor or set up a shunt resistor to convert it to a voltage input. Your output value (ie the pressure/temp/angle you are measuring with the sensor) can be pretty much anything.
  23. cj

    COP on a B16A

    This is based on the Motec suggested numbers for the denso part number and works well for me on an n/a k20. You can also calculate numbers from the hondata base maps which come out about 20% higher in the 12-16v rows so if its turbo you may want to use 10-20% higher.
  24. Can you clarify this please? Are you saying that it you unplug the cam phase solenoids, the ECU doesnt power up at all? Your trigger1 looks like its not picking up properly. Your trace only goes up to 3.5ish volts, and its very "spiky", compared to Adam's reference which is a nice square wave. This means the falling side of each tooth is going to be a few degrees off where it should be, and this is the edge you are using for trigger1.... hall effect sensors shouldnt ever do partial voltages like that either, they contain an internal filter of sorts that should make them on or off, and on should be 5v. the sensor might be playing up or it might have a power feed issue given what I think you are saying. You havent logged any internal voltages in that log, but the last runtime values in that config show the ECU was seeing 83degC coolant (so it was running very recently), but only 11v battery input, and 10.3 volt internal ECU voltage. What does your power and ground setup look like?
  25. cj

    Stalling Issue

    There can be reasons you would want this to be different (such as trying to get revs to drop quickly to X then drop slowly from there down to idle), but its probably a good idea to start with them at the same value so you can tell if you are dealing with all idle values or all normal values. Remember that ignition idle rpm cutout is an absolute value, and throttle idle rpm cutout is value + target idle value.
  • Create New...