Jump to content

007

Members
  • Posts

    30
  • Joined

  • Last visited

Posts posted by 007

  1. I'm definitely now going to crunch the car on any speed breakers. LOL.

    Thanks for the information though. *If* this happens the next time and I'm in a position to have the car towed straight back to the shop, I'll try to hunt for the short.

    Just wanted to make sure that this is usual behaviour under the circumstances and the ECU is fine.

    Thanks Adam!

  2. There's a bit of background I need to establish here, so apologies if this is slightly long.

    I was hooning around one night as usual and on the way back, ran into a couple of extremely lousy speed bumps in quick succession. If sounded like the downpipe and / or oil sump bore the brunt of the blows. After the second hit, I noticed my CEL and oil pressure warning light come on, and the engine died.

    When I broke out my laptop and connected to the ECU, imagine my surprise when every single sensor was spazzing out. Even the ECU temp read -12C, which it pretty much impossible. The root cause was the 5V supply malfunctioning. The real time display of the 5V supply read under 1V. The MAP sensor readout was swinging from -ve all the way to thousands of kPa.

    I naturally assumed that the hits must have shorted out a sensor and that was causing the 5V rail to keep cycling. However, I wasn't able to find anything obviously broken or damaged. Shaking the harness wherever it had some play didn't seem to change anything either, so that ruled out a loose contact.

    After about 30 minutes of futile attempts, I was ready to leave the car where it was (pretty close to my destination), and come back with a tow strap in the morning. However, I decided to give it one last shot and cranked the car, and was rewarded with a slight cough from the engine. So I continued cranking with the pedal mashed to the floor and the engine sputtered to life, even though the MAP was still all over the place. With the pedal still mashed all the way, I kept riding the clutch to get the car to move. It died thrice in the next 100m and probably pissed off the neighbourhood, but the fourth time around, while I was still trying to ride the clutch, everything seemed to come back to life just as suddenly as it had died, and the car started to run like a champ once again as if absolutely nothing had happened 30 minutes ago.

    The next day, I lifted the car and checked for underbody damage, and found absolutely nothing. All electrical components and connections were solid too. I started the car up and shook whatever I could, and the engine didn't skip a single beat.

    I'm not sure how to make sense of any of this. How could a sharp impact have caused this issue that seemed to solve itself? Is this something I should be worried about? Obviously, besides being more aware of my surroundings, should I also be looking out for something in the ECU? From what I gather, everything in the 5V loop is automotive grade and there are no moving parts here that can short on impact only to be magically fixed with no input from my end.

    Has anybody had a similar experience that can shed some light on this?

  3. Another related question: if I've wired a fuel temp sensor to the ECU, is the density change automatically accounted for by the fuel model or do I need to set up a separate table for this?

    I ask because the block diagram in the help file did not seem to show a fuel temp input like it did for fuel pressure.

  4. Thanks Adam. I think I understand. I was equating VE with cylinder filling.

    Per your explanation, VE for the most part  - especially in boost - should be 100% or thereabouts, right?

  5. Thanks for the clarification Adam. I understand the warranty implications. Both my eThrottle outputs are free, and since you confirmed that they have active current limitation, I'll try combining the two of them.

  6. Thanks Adam. What do you think of:

    1.  Using two combined injector outpputs and configuring them identically.

    2. Using one of the eThrottle motor outputs.

    Also, from what I gather, both the above outputs are over current protected, right? Meaning I should be able to experiment without frying something in the ECU?

  7. 41 minutes ago, neil brown said:

    Injector output is limited to about 2 amps max

    That would be Ignition output.

    Injector outs can deliver a max peak current of 10A. I'm just not sure if they can hold 7A continuously.

    I reckon eThrottle outputs could support this.

     

    6 minutes ago, Confused said:

    I'm using a pump from an X5, a Pierburg CWA50, for my water to air chargecooler.

     

    I posted my settings in this thread here, hope it helps!

     

    Thanks! This looks interesting. However, it looks like the PMU is handling power in your case while the Link only provides a control signal.

    The pump I intend to use is a simple 2 pin affair that will kick on when a certain condition is true and remain that way.

  8. Hi. As the title says, I'm planning to use a G4+ Thunder to drive a water pump in a W2A system.

    The pump I plan to use is one from a BMW M2C, and per the documentation, it draws 6-7 amp.

    Now I could wire a relay and get it over with, but in the interest of keeping things clean and simple I was wondering if I could wire it up directly to the ECU.

    I reckon an INJ or IGN output could be used for this purpose. Can somebody confirm? In case 6-7 amp is too high for a single output, could I join two of them together to supply the required current?

     

    Thanks in advance.

  9. The only info from the ECU they expect are RPM and VSS. The HDS scanner also expects a repeating VIN frame to be able to connect.

    Because no other modules are touched, they remain happy and chatty as usual.

     

    There are two major issues here

    1. the cluster expects a speedo pulse from the ECU once every 10 meters to keep the odo ticking

    2. The last byte of each frame is like a checksum, which must be valid, or the cluster rejects the entire frame and everything goes haywire. I've managed to figure out enough of the logic to have the Macchina compute this byte and keep the cluster alive.

     

    Besides these, everything else is fairly straightforward.

  10. Hello. I've spent a considerable amount of time analyzing the CAN data for a Civic FD2 in order to retain the stock cluster functionality. Now I can operate it with a G4+ Thunder and without requiring the stock ECU to be present on the CAN bus.

    There are some quirks of the Honda protocol that cannot be defined using CAN Manager, so I use CAN Manager to put the bulk of the data on the bus, pick it up with a Macchina M2, do the required transformation and retransmit.

    I was wondering if you would be Keen to incorporate this logic into the Thunder directly. This will hold good for a lot of Honda's with CAN driven clusters.

  11. Thank you for the response Simon. What you say makes sense, but do you reckon that the nice and round threshold at 55C is coincidental?

    I've disabled ECU Hold Power for now by using he same switched ignition output to A5 and C2.

    Would it be possible for the team there to recreate the issue in the lab by putting a Thunder under a hot air blower?

    I don't have another unit with me at the moment, or I'd have tested that one in the bench myself.

    In case there's a suspect joint, would it be worth the effort for me to take the case off and visually inspect the board?

  12. I haven't been able to find a way to generate a high side driven PWM output. I'd like to use this to implement alternator output control. Since the entire unit is grounded to the chassis, low side driven PWM isn't an option.

    Using GP Output and setting a threshold voltage causes the output to keep thrashing on and off, and makes the idle really lumpy and the lights flicker.

  13. It's been a while, but the build is finally alive once again, and this time, I have concrete information.

    The issue is most definitely with the ECU and not the wiring.

    There are no problems below ECU temp 55*C. When the temp crosses 55, the ignition signal input (pin C2) stops registering key off, and instead reads Active. When this was happening, I even snipped the wire going to C2 clean off and probed the voltage from the ECU, and it read 3V. I left things untouched and let the ECU cool down, with C2 still snipped off and the moment the temp dropped below 55, the ignition input correctly changed to inactive and the ECU turned itself off.

     

    I'd really appreciate if if someone could look into the matter and hopefully come up with a fix. This behavior doesn't look random to me, and I suspect should be reproducible in a lab.

  14. 6 hours ago, cj said:

    Having another look at your pics, you could get a thin sheet of alloy and bolt it between the ECU and the engine, and have the front section run 45* across to behind the headlight so it blocks radiator air, kind of like a lot of people do to shield cold air intakes. You may even be able to run a ducting tube to pick up clean air and run it into the ECU side of the alloy sheild if its still getting a little bit hot.

    After everything's been installed, there's hardly any room to fit even a pinky finger around the ECU. So I'll try to completely wrap it in heat reflective tape and have a tiny opening facing the firewall. I don't think a diagonal sheet will do much good once the hot air from the fan diffuses everywhere in the engine bay. Summer temps here nudge 50*C.

     

    5 hours ago, Simon said:

    For the ECU side they are tested up to 70C, 80C will be pushing it for sure. 

    Thanks for the confirmation Simon. I must compliment you guys on this ECU. Even above 80*C, it handled itself like a champ. I had previously installed a Haltech Elite 1500 in a similar position, and that thing would start going into limp mode the moment internal temps crossed 60. I was finally forced to relocate it inside.

  15. Thank you for your response. To your points:

    13 hours ago, cj said:

    Looking at your pic you have replaced the dizzy and are running COP's, so there is no way the engine can "self idle" if the ECU is off.

    The engine isn't self idling. The ECU isn't powering down.

     

    13 hours ago, cj said:

    Therefore, your issue is more likely to be either A your main/ecu relay sticking on because of temp, although factory mounting for these is up under the dash rather than the engine bay,

    It isn't a sticky relay.I triple checked, and once when this was going on, pulled the relay clean off and the ECU still wouldn't power down.

     

    13 hours ago, cj said:

    B (more likely) you are getting some kind of back feed of power through the ECU which is allowing it to keeping it running even when the main relay feed is cut.

    Pretty much impossible because the line that goes to the ECU's ignition signal pin is the only thing aforementioned relay controls. The low power side is wired to the ignition key, and the high power side to the ECU and nothing else. When I turn off the key, everything else in the the car that is on ignition power goes dark, except the ECU. Also, when this happens, the Runtime Value for Ignition Signal remains ON.

     

    13 hours ago, cj said:

    Assuming the temperature thing is not a red herring, maybe one of the temp sensors drops to a really low resistance when hot and because of a wiring quirk, power is being drawn in through this? Have you got any sensors or solenoids being hot fed direct from the battery or are they all behind relays? Have you wired in the Main relay controls on the thunder or just connected "14v in" to an existing ECU relay-like feed on the existing car wiring?

    I've wired up the Main Relay controls, and everything on the engine draws power from one of three relays that are controlled by the ECU's Main Relay Out. If this was something to do with flywheeling diodes, this would happen all the time. Every single time this has happened has been after a crawl through traffic.

     

    13 hours ago, cj said:

    My gut feel is that its unlikely to be ECU internal temperature, but this is a logged value so you can check. Can you post a config and log of when it fails to shut down?

    I wasn't able to log unfortunately, but I was keeping an eye on ECU internal temperature while a friend was driving through some pretty thick traffic after a long drive. There was no headwind to evacuate the engine bay, and the radiator fan venting all the hot air inside the engine bay drove up the ECU temp to a really alarming 85*C by the time we reached home, and I rushed out and pulled the C connector to kill the engine, which is what I always do when the ECU fails to shut down.

    This was the point when I decided to Thermaflect sleeve the ECU to prevent it from heating up. If that does not work, cabin mount is the only thing left to do. I really hope it doesn't come to that.

  16. Hello. I've been running a Thunder on a build I did recently. This ECU is installed in the engine bay like shown in the attached image.

    For the most part, the ECU doesn't seem to mind this at all but I did notice one glitch that would repeat consistently after some crawling through traffic. The ECU would not recognize the ignition key being turned off and would continue to idle the engine even after the key was removed. This doesn't happen when the engine bay is relatively cool. I haven't been able to identify the exact threshold past which this takes place, but I'd guess around 70*C ECU temperature.

    Any ideas why this might be the case?

     

    On to the second question, I plan to do something about keeping the ECU cool. While the obvious solution is to move the entire unit inside the car, that's something I'm considering as a last resort option. The current idea is to get a Thermaflect sleeve and wrap the ECU in it. That should prevent direct blasts from the radiator fan which is what seems to be sending ECU temperatures to 80*C and above. What I'm not sure about, however, is how this will affect the ECU's ability to dissipate its own generated heat. Can somebody from Link chime in to and let me know whether or not this is a good idea?

    39676136_10217405720761396_4678614279052591104_o.jpg

  17. Hey Adam. The ABS module provides individual wheel speeds in a 8 byte frame. However, the speed displayed on the gauge and the odo tick information comes from the ECU utilizing input from the vehicle speed sensor mounted on the gearbox.

    I understand that it will take a lot of generic CAN functions to cover all possible OEM strategies, but in the meanwhile, would it not be possible to add a wheel revolution counter for either one or all four wheels, similar to the engine revolution counter that is already available? I feel that is a generic enough function for this purpose.

    I would really appreciate it if you could add this in. It is the only fly in an otherwise perfect ointment.

    Good to know that you are working on a plugin solution for this platform. I have reverse engineered a fair bit of the CAN structure to drive the gauge cluster, and would be happy to share that information if that will be of any help to you.

  18. Hi. I'm going to be building a FD2 Civic with a K24 engine and a Link G4+ Thunder. I plan to drive the gauge cluster using the User Defined CAN feature. I've managed to figure out everything except a tiny detail related to the digital odometer.

    Honda sends an "odo tick" once every 10 meters to the cluster. This is a 1 byte value that counts up every tick and cycles back to 0 after 255 and repeats. Every 10 ticks, the cluster increments the odo by 0.1. I've searched all the available data items to send in a CAN frame and haven't found a speedo pulse count or anything I could use to enable this feature. I see only speedo out (status), but that won't do. See the attached graphical representation of the CAN data.

    Am I missing something here? If not, could this feature be added in a future firmware update?

     

    Thanks in advance.

    speed.png

  19. Solved!

    Again, it was the default settings that did me in. The max and min clamps in the IACV settings were holding the commanded duty cycle back.

    The IACV works as I guessed before. Unpowered, the shutter is halfway open and that is where it returns to rest once any applied signal is removed. Grounding the IACVP pin rotates the shutter one way and grounding IACVN rotates it the other way. 50% duty cycle keeps the shutter at its default position. Anything above or below 50% either increases or decreases idle depending on how IACVP and IACVN are wired.

  20. I believe this IACV requires a PWM ground signal on both pins besides the +12V pin to move the shutter in opposite directions. Let me try to scope the stock ECU's control signals versus the Link and revert.

×
×
  • Create New...