Jump to content

Davidv

Members
  • Content Count

    311
  • Joined

  • Last visited

  • Days Won

    27

Everything posted by Davidv

  1. Do you have the other half of it connected to 12v? I've done the same where I removed the factory stopper switch and put another brake switch in. But the problem was that I really needed the switch to activate on/off closer to the engagement/bite point of the clutch, rather than right at the end of travel. So it didnt work so great for launch control etc. So I'm going to redo it with an adjustable mount somehow.
  2. So you've physically wired your tacho to the aux output that you're wanting to use as a tacho? Once you've done that, download PClink for G4+ and plug into your ECU. You'll see a list down the left hand side, listing the aux outputs and what they currently do. If you click on one you'll be able to reassign it in the box that pops up.
  3. What's been interesting is that a lot of weight removal so far has come from things I guess I have always considered to be a "fixed" weight. Like, a different alternator and bracket are 4kg lighter. I've never even considered looking at an alternator in this manner before. A lot of people fuss over having XYZ engine because it's 20kg lighter than ABC engine. But then you see people make the benefit null and void by using 10mm thick steel for engine mounts and stuff like that haha. I think a carbon fibre bonnet will make a lot of sense, but the front guards are paper thin and dont weigh much. Doors definitely have a bit of weight. Interestingly the Coupe version of this car weighs more than the sedan, because of the gigantic slab of glass in the rear hatch. So getting rid of that would be good, but, like you say not all of these things are a good compromise for a road driven car. I'll be pretty happy to get a fair bunch of weight out of the heavy end of the car. I might make moulds of all of the panels though, just because it's impossible to get replacements and the existing pool of parts slowly rusts away more with each passing day haha.
  4. In other news, I've ended up moving house a few times and various other annoying life things have meant not much time for car stuff. But I've switched over to a dual VVTI motor now. While doing so, I realized how stupidly far forward my engine sits, so I scope creeped this incredibly basic minor upgrade into a full blown overhaul haha. I decided I want to try get the car down to mid/low 900kg without sacraficing its road manners or legality or comfort etc etc. Which is so far looking like it'll be easier than expected. New engine mounts I've made are 6kg lighter than before. A 4-1 manifold is 10kg lighter than my 4-2-1. But both will be able to fit. Hoping to come up with a lighter intake situation. Moving to a lithium battery and getting rid of battery location in the boot. Saves about 15kg and should help crank the motor over a bit better. And makes my boot a bit more usable again. Then moving to a dual master pedal box, downsizing rear brakes and a few things like this too. Just lots of little things that add up. I'm Consolidating all of the 3/4 empty fuse boxes around the place, into a single Bussmann fuse/relay box in the center console. And the new engine has ended up about 100mm further back which is crazy. Some small chops to firewall but my engine bay has taken a hammering over the years so time for a repaint anyway. So I'm trying to make things a bit tidier this time around rather than just "functional" haha. But there are some remarkable improvements in coolant routing and so on which come from moving to an engine that's actually designed to be in an FR layout. Who knew
  5. I've been doing some work on making a crude engine simulator, so if I'm using something else that I want to test that has Canbus I can just hook them up to each other and expect similarish results that my engine would give. So it needs an airflow model. So far I'm just using these simple factors to determine how much airflow the engine could potentially be consuming RPM How much air per cyl at 100% VE How max VE tapers each side of max torque RPM based friction losses Potential horsepower based on mass flow Then I've modelled a throttle body too, which has an exponentially increasing/decreasing flow rate based on angle. So if the angle allows you to flow more air than what the engine is trying to consume, MAP stays at 101kpa and the engine can reach its full throttle VE value. But if the throttle is closed further than this, then it shows the number as dictated by throttle angle but also generates a map sensor value. Which then generates a pumping losses number based on MAP and RPM. So the idea here is that it can simulate the situation where the engine is still able to flow 100% of its potential power even at lower throttle, if the engines airflow demand is low enough. Even if he airflow number is fixed the actual power number changes a little through the RPM sweep due to differing losses. So to add more features I basically add more things which potentially take away from max power if they are non optimal. So I still need to make an AFR model, so that you'll make less power as you go further away from peak power AFR. Then make an ignition model which has an ideal ignition number (in milliseconds to 15 ATDC so it generates a fairly sane number across the rpm range) But then has variables to account for different amount of ignition needed when you run richer or leaner, or less engine load. Then from there, I'll make a road load simulator where the car is requiring a varying load as it travels along ups and downs. So the simulation of cruise control is a bit more accurate as in real life the test conditions go out of bounds a lot if the road gradient changes too much across the course of a single test. So obviously this wont be a 100% accurate simulation of my engine or any other engine. But it will react to changes in hopefully a similar way a real engine does, and be capable of being tuned like a real engine does. The eventual idea is that I'll be able to test and troubleshoot my cruise control ignition trim system using this instead of needing to drive the car every time I need to test something. So this virtual motor will have an ideal ignition table which I currently dont even know what it is. But it will be my other devices job to run iterative tests to figure that out. This will be able to whizz through 1000s of iterations much faster than how it can process in real time when you're waiting for fuel samples to generate and conditions to stabilize. Which should help for troubleshooting and bug finding/fixing to no end. So hopefully when I get my car running again it'll be fairly stable. What will also be good is if a G4X ECU is in the budget at some point, I'll be able to drastically cut down the number of calculations etc I'm having my dash do and send back to the ECU. As the math channel feature will allow the majority of things to be done all onboard which is exciting.
  6. Davidv

    Injector Timing

    Remember that injector timing only really makes a difference when the injector is open for a very small amount of time. Such as idle or cruising. Because once you're up to say 70% or higher duty cycle, your injector is firing for so long that it makes little difference. I've found that around 400deg is good but on cold starts bringing closer to 250deg (from memory ) meant I could reduce cold start enrichments.
  7. I've wired one up fine, 82mm though. Just using the Altezza e-throttle settings for now, havent fine tuned. But first things I would check, are if you move the throttle plate by hand, does the TPS signal move? When it does the calibration, does the throttle plate move on its own, or not? I'd say the situations you need to look for, are TPS signals not working or maybe the polarity of the motor control wires are back to front.
  8. I remember there was a thread a while back about a particular video card driver that causes these symptoms. I went from a slow old laptop that worked fine, to a flash new one that ran like a dog. But there was a fix for it by changing the video card driver. Are you definitely using the latest PClink version? As I think they also applied a fix for this inside PClink but cant remember.
  9. I've been thinking about this lately, essentially you're trying to build maximum horsepower to your cars/wheels traction curve. So you dont necessarily want more boost in 3rd gear at the same speed as second, you want the same amount of horsepower at that given speed. So I thought controlling the wastegate based on speed vs injector duty cycle would be a decent simulation of horsepower level. If you dont have enough boost in gear X to get to 30% DC on the injectors at 55kph. Here, have some more haha. (and vice versa)
  10. Davidv

    Knock trouble

    I havent had a look in your logs files, but just a thought - your knock sensor is just a microphone, so anything else that happens around same frequency as knock will get picked up like knock. I drove myself mad trying to solve a similar issue, with knock only at certain load and rpm combinations... gave up - months later I found that the springs were rattly and loose in the clutch plate. So if the clutch was loaded up one way or the other it was fine, if it was "coasting" it would rattle and show as knock! Do you think there's possibly anything else going on in the motor/gearbox/etc etc that could cause something to happen at that rpm?
  11. So the % of zone means, how close it needs to be to the center of a cell to count the value example - if half way between one cell at the next, ignored.
  12. If you do that, arent you liable to get e-throttle errors if it over runs past what it thinks is the maximum/minimum allowable voltages? As realistically it's going to overshoot at some times.
  13. One thing that's unique about E-throttle compared to a cable setup, is that its possible to smash the throttle into the end stops and damage the internals of it. (I've got a pile of busted e-throttle gears here somewhere...) So generally the solution is to limit the targeted travel to say 2% and 98% of the actual measured sweep of the throttle. But, its kinda weird looking through the logs when you see it's at 98% when really its 100% as far open as it should be. Or when it's at 2% that's really its closed position. So I know it sounds nitpicky but it would be cool if it could span from say -5% throught to 105% Or maybe on the TPS calibration you tell it to stay XYZ amount off the end stops. Then it stores that trimmed value as 0-100 but has tolerance for going say 2% or 5% over that, without triggering errors.
  14. You can already do that too. You just open the log file manager tab on the side, and then click in the box for the one you'd like to overlay. Then it shows two lots of graph times down the bottom, you can move one back and forth if you want to try make them line up.
  15. Imagine you had your bare engine block, and you hit it with a big hammer. And it went doooooonnnnggggggggggggggggggg like a bell. That's basically what knock does, and it's what you're listening for. The size and shape of your engine block changes the noise that it will make. (Primarily bore size) So it's not a case of setting it up based on what sensor you're using - it's about matching the sensor to your engine. As a knock sensor is basically just a type of microphone. There's a good article on knock sensing here: https://www.hpacademy.com/forum/webinar-questions/show/083-understanding-knock-1
  16. Hi, just putting together some math channel stuff where the numbers might get quite big. I am guessing that the largest number you can go to would be a signed 32 bit integer? 2,147,483,647 maximum? Or are they a 16 bit signed int? Or a float? something else. Also, what happens when you overflow this number? Will it stay at the max value, or bounce back to zero or a negative value? Thanks
  17. Davidv

    Water Pump PWM

    Ahhh thanks, excited about this now! Will put this back higher up the priority list haha. Does the signal ground wire need to go to the signal ground wire on ECU? Or both to same place? EDIT: Ahhhh no I get it now, feed it 12v constant then PWM on the earth pin back to ECU.
  18. Davidv

    Water Pump PWM

    Hi Richard, Do you mean it's as simple as connect 12v pin, earth pin, and then 3rd wire to to an aux output set to high side direct to ECU? Sounds too good to be true! I was anticipating needing to make a pump controller or using SSR to trigger that 12v signal pin. What do you do with the 4th pin on the pump, I was under impression this is essentially a pump speed tacho signal. So you can wire it to a DI. Is this correct?
  19. Davidv

    Modelled fuel

    Definitely put in the tested value. If you want really good accuracy though then it's good to have a fuel pressure sensor so it can work out the fuel pressure differential accurately. As mechanical FPR can fluctuate a little. Your deadtimes being accurate will be a big make or break for accuracy of your fuel consumption though.
  20. Hey Jenova, I've run a redtop 3SGE for a long time with a G4+. Great engine! You can use the Altezza 3SGE basemap, and then just turn off the exhaust side VVTI options. I can give you some decent maps for cam timing / ignition / etc but will need to know what sort of load axis you are using. (MAP sensor?) If you PM me your email address, I'll send you a redtop specific map that you can start from. But keep in mind there are a lot of variables that you'll still need to fine tune yourself for your particular engine. I would reccomend that you switch to a set of Altezza injectors, if you can source them cheap locally where you are. Because I've got good accurate injector data for those, which helps a heap.
  21. Hi, Just had a thought for an idea while thinking about a situation with an NA rally car. Being a very dirty environment in rally it's prone to clogging air filters and reducing performance. I had an idea that a set of virtual AUX could trigger a light as a "filters clogged" message. Like so: So it would trigger a light to say the filter needs changing. However - the ECU doesnt have any outputs left for a light. So I thought it might be cool if you could trigger a check engine light with virtual aux instead. You could use the label on the virtual aux as the message for why CEL has been triggered. Or, maybe as per OEM ECUs CEL you could have a list of 10 error lights. Then they will flash say 5 flashes in a row then long pause. For error 5. Maybe you could have a drop down "warning status" with the options off, trigger CEL, turn off CEL. So you could have a set of conditions contrary to above, on a second virtual aux which turns the CEL back off if conditions are met.
  22. Davidv

    G4x xtreme

    Ahhh thats a lot earlier than I was expecting! Awesome, cant wait.
×
×
  • Create New...