Roman

Aaahhhhh good news and bad news: Good news: Got a wideband Boss welded into exhaust (Cheers Sentra Dave!) Got exhaust manifold thread fixed (Cheers Peter!) Got everything else wired up, and running nicely at about 2000rpm with nice fuel. Bad news: Welded the wideband boss in a place where there isnt much room under the car, had to aahhh 'massage' part of the gearbox crossmember to make it fit. The engine is fine at 2000rpm... fine at 1900... 1800... 1700... then cuts out. Try it again, same thing. And again and again. And when cranking, it would sometimes backfire out the intake. Weird. So I did some datalogging Just below 1700rpm cuts out because the rev limit cuts in... Because it thinks the motor is going 32,000rpm. Lol! Checked trigger error logs... lots of crank angle sensor errors. and all at that same rpm. Ran the onboard trigger scope, showed this: (look at the wonky/shorter wave near middle of the page on top graph) A long time ago I broke a tooth off the Crank pulley (Doh!) and couldnt find a replacement so had another one welded on and filed to shape. Well it looks like this might have been causing problems ever since, you can see that the wave on the graph gets a bit unsteady at a certain point. It happens at the same place each rotation. EDIT: Looks as though there's a software fix for this in the ECU, I'll try tomorrow! (Reducing arming voltage for the trigger at 0rpm and 1000rpm)

There's a Hammy Doon Trackday on the Monday next week (Auckland Anniversary) I'm paid in but I've got a shit tonne of things to do before then haha.

Hey Peter when people fit an EGT (Exhaust Gas Temperature) sensor, they always find that when you run low ignition advance the EGTs go up. As you advance the timing more the EGTs go down. If your motor is running noticably hotter than when you had carbs it could be that the ignition needs to advance a bit more at low load / idle etc

Problem! I get ants in my pants when I've got nothing new to learn / am a bit short on hobbies currently. Also, I hate time and money. Solution! Buying a Link G4+ Xtreme and an LC2 wideband What wooouuullllddd have made sense for me to do next, considering my driving is a bit shabby and I havent done that many trackdays. Would have been to buy a set of semislicks and spend a bunch on trackday entry fees while summer is still here. However! That bloody 7400rpm rev limit was annoying the piss out of me, some parts of the 'tune' on factory ECU dont run so well on the slightly bigger injectors (like cold start) and there were still a few annoying spots where it seemed to run rough and I had no way to fix them. I was originally looking at a megasquirt, because they are cheap and looked to be able to achieve everything I wanted. But then I decided I'm probably not skilled enough to assemble one, and didnt really want to either. So then I looked at premade megasquirt kits... Bloody hell almost the price of a midrange Link anyway! So this train of thought escalated to the point where I figure that if I'm going to spend $$$ on an ECU, I dont want to buy something that then runs out of scope for what I want to achieve, and then regret not spending another couple hundred bucks. So opted for the fancy pants one with lots of inputs and outputs. I spent two nights wired up the engine loom to the Link, and double checking everything. Plugged it all back in, rather nervously... (Half expecting an engine bay fire or something) It fired up first pop! Super stoked that it could run fine, especially considering I dont have an oxygen sensor or MAP sensor connected yet. But all of the sensors register in the ECU and are working, so I'm pretty bloody well chuffed about that! I didnt have to do a single thing to get the crank or cam sensors, VVTI solenoid, coilpacks or anything working apart from wiring them in. Awesome. Tomorrows mission is to get the wideband fitted, find a 1 bar map sensor from something at Pick a Part, and tidy everything up. Oh yeah, and I'm booked in for a Trackday next Monday so I've got to hustle a bit to get this thing tuned too haha. I think I've learned enough from my time with datalogging the standard ECU to be able to give this a good shot at a decent tune myself. From my experience with the standard ECU behaviour it likes to run right to the limit of knock, both at cruising and lots of load. So I dont think I have to worry about shooting too far past minimum advance for Best Torque because knock seems to be the limiting factor to ignition advance instead. Although, I've datalogged a bunch of ignition behaviours from the standard ECU so I know what to start out with for timing under cruising / load / etc anyway. I'm just stoked that now I'm able to change everything, rather than smoke and mirrors to get the factory ECU to behave. This can also datalog 15 sensors at 40hz, rather than my factory ECU + OBD which could datalog 6 things per second in total. Huge difference! Looking forward to MOAR datalogging and most of all making some 8000rpm doort noises!

I've almost finished my loom for the link, yusss. Hopefully finished tonight with the wiring, except for wideband (need to figure out where to mount it / run wiring) and map sensor. (as I dont have one yet)\ But will hopefully be able to start on sensor checks and calibration ooooohhh exciting. So far I havent used a single thing from the 2nd ECU plug, and have a few left over on the first. So I've got a shit tonne of inputs and outputs left!

Beware that Hilux diff only has fairly average LSD options, unless you spend a fair bit. If you're planning to lock it though then go for gold! What ever you decide on though, do some thorough checks into what diff ratios and LSD options / disk brake options are available. And if these suit your car. It's a lot easier to make something work that's got factory disc brakes etc. If you do end up in the situation where you need to convert a drum rear to disc though, look for cars that have handbrakes built into the caliper to get calipers from. S13 Silvia, SW20 MR2, AE86, AE92, VW passat and a few other things are reasonable options for pinching handbraked disc calipers from, depending on how you need the cable orientation to be. Although these tend to give shitty handbrake strength so if you're wanting to E brake D1NZ all over the place then you might be better off with a diff that keeps drums on it, or has internal drum as well as disc rear.

How tricky is it to calibrate a map sensor to use with an ECU? I want a 1 bar map sensor, and there are zillions of these around (pretty much every NA car ever with MAP) Not keen to pay $200 for one if I can help it... Is it just a matter of finding voltage/pressure at each end of the scale? Link has an onboard barometer that apparently helps with MAP sensor calibration but I dont know how it works yet as I need the ECU plugged in for the tuning software thingy to tell me how the calibration works.

Regarding 02 sensor longevity - In the manual for my LC2 it says: Naturally Aspriated Daily Driver: -Calibrate before install of new sensor Calibrate new sensor again after 3 months of use Thereafter calibrate once a year or every 20,000 miles, which ever comes first Which would seem to imply it's not unreasonable to expect it to last longer than that.

Is this a factory sensor on any type of car that anyone knows? http://www.nzefi.com/product/bosch-fast-response-iat-sensor-turbo-na/

Thread bump! On new ECU there are Analogue inputs (0-5v) digital inputs, and temperature inputs. I understand analogue 0-5v and temperature ones, but what does digital input mean? Is that for something like a square wave signal from a crank angle sensor or something? I want to know what/how I could datalog some non engine parameters using those unused digital inputs, but before my imagination runs too wild I need to know what exactly it can or cannot do. I'm guessing any form of potentiometer is an Analogue signal. But anything like an ABS sensor isnt. Sound about right? If it's square wave or similar, counts as digital signal?

Today I learn: Peters car is an LFA You got the motor back together now?

Yeah it was the 8 injector drivers that were the main reason to get the Xtreme over the Storm. That and the extra inputs/outputs mean I can datalog a whole lot more about the car itself, not just the engine. When thinking about the price difference over the course of a year or two, it's not too much. But if I bought something that didnt have enough stuff for what I want, I'd probably end up wishing I just spent the extra couple hundred bucks.

Ahhh here we go... http://www.diyautotune.com/catalog/diybob-breakout-adapter-nippondenso-122-pin-p-537.html A bit spend though for what it is I guess! Maybe butchering spare ECU is the way to go haha.

Cool yep, that's pretty much exactly the plan! But you can buy the same thing from somewhere without having to yank an ECU apart. I just forget where at the moment. But in my mind it's the best option - Toyota factory wiring is good, plugs are good, etc etc. So may as well keep as much of it as I can. It means I can design and assemble and test the loom out of the car, as it will only be ~300mm or so long if that, and all just goes between one thing and the other. Any wiring faults are contained within that area, rather than spread across the whole thing... And I dont have to change a single thing in the engine bay.

Twisting wires + masking tape wont cause problems for a wideband signal right? Or does it need shielding? I've got some tinfoil here too.

My engine loom is staying as is, no need to reinvent the wheel when it's all already there. I cant remember where from, but you can get an adaptor board that has the plug shapes of the standard ECU. So I only need to make an adaptor that goes between the link ECU and that adaptor board, then the standard engine loom plugs into that. Keeps it simple / can revert to factory ECU if need be for whatever reason / wiring loom ends up very short with all connections in the same location which makes troubleshooting anything easy. There's a bunch of spare pins on the factory ECU plugs too, and I've got an Altezza loom here which has unique wire colours which I can use. So if I add any other stuff it's mostly a case of pushing already crimped connections into the plugs.

Factory ECU is pretty awesome for most things, but just cant get around the rev limit. So yeah, if going aftermarket definitely no reason not to setup launch control etc There's a basemap for the link already for my engine that sets up VVTI, coilpacks and all that other junk. I havent found anyone who's running MS on a beams motor, so dunno what to do if I get stuck. Vs lots of link installs by heaps of people already. I dont have the electronic smarts or the patience to spend ages figuring things out with an oscilliscope or whatever. If I converted my engine to an Altezza loom there's a version of the Link where you can just open up the factory ECU case, replace the guts with a link and then it plugs into factory loom. Was considering this but it doesnt have as many inputs and outputs as the 'normal' one. So went with that, as the wiring looks easy enough.

Okay so I've been giving ECU choice a bit of thought over the last while. As there are a still a few niggling issues with using the factory ECU, and then if I went to an aftermarket one I'd want to get something with a lot of extra scope for nerdy datalogging, extra features etc. So the Megasquirt V3 or whatever with heaps of inputs and outputs looked like a good option. But before diving into that I downloaded the MS tuning software on my PC, and got familiar with it. Seems all good. But then I upgraded my PC, and I cant get the fecking thing working again. Also doing a bit more reading it seems about 50/50 odds between getting things running alright with megasquirt, or having a complete nightmare of a time and being out in the cold in terms of any support. The fact that the tuning software doesnt want to work anymore didnt inspire confidence in the rest of it. I tried reinstalling JRE etc but no luck. I dont really want to assemble anything myself either, so looked into the cost of a preassembled one... $1600 or so by the time it gets here. Cant remember if that was with or without tax. So looking at the difference between that and a link, it's not that much more of a stretch to get something that's going to give me so much less of a ball ache. So decided to bite the bullet and have forked out for a G4+ Xtreme and an LC2 wideband. I cry when I think about the cost vs potential HP / rpm gain, but it's more that it's something interesting for me to setup and tinker with. Factory ECU tries to run on stoich at cruising too for sake of Catalytic converter, so will be able to lean it out a bit more and make it a bit more thrifty to drive perhaps too. It's got so many inputs and outputs, good datalogging etc which will keep my nerdy tendencies satisfied for a long time yet! Something new for me to learn about in this coming year. Should be here some time next week.

The point of coilovers is usually that they use a generic sized spring. So you can usually just find the measurements and then get generic coilover springs in whatever rate you want. A coilover setup that uses an oddball sized spring diameter probably isnt a good choice when you're likely going to tinkering with it as you go. but if you can only buy the kit with 6kg then yeah it's a more useful starting point than 8! Speak with some local spring shops and see what diameter coilover springs they can get and how much for. Then see if this matches the size used in the package you are getting.

I've still got all of the interior out so took the car. So for curiosity's sake I took it to the weighbridge with a near empty tank of gas. 970kg with all seat belts and a few other things still in. Pretty chuffed to get it under a tonne, didnt think it would be that light as it is. The new exhaust is a bit lighter than before. Still has heavy wheels, full glass, heavy drivers seat and a few other things so you could likely get a beams Carina down to low 900s or high 800s if it was a race car. But taking all the interior junk out looks to be the way to go for next time at the drags, could help a bit.

Have you played around with fitting a rear swaybar or not, on any of your cars KPR? I've got the option to fit one on mine. On one hand though the current feel of the car doesnt seem to need or want one. But on the other hand quite a few live axle cars of the era came with one. the GT Celica I got my diff from had a swaybar from factory. But all of these cars had very soft springs to start with. If you've increased the spring stiffness you get more roll resistance anyway... And the roll centre distance to CoG reduces when you lower the car a bit too. So seems like it becomes redundant anyway. I guess I should just bolt it up and see haha.

What is your starlet going to do / going to be for? 6 or 8kg is fecking stiff for a rear spring for a street car or track car of that size/weight I reckon. My car is bigger and heavier than yours, but everyone told me to start at about those rates. (TRD AE86 springs are 8kg front 6kg rear, so seemed like a good starting point) Ended up lowering spring rates 3 different times and picked up stability, and speed each time. and lowered laptimes by a big margin. Ended up about 3.5kg in the rear and probably 6 or so in the front. With the hard springs, under braking any tiny bump in the track would cause wheels to lock up because the car bounced. And cutting over ripple strips (the most fun thing about trackdays) was impossible. It was scary and not very fast! But it depends on what you're wanting to do, drift cars generally seem to run much harder springs all round... I'm guessing these are what the TRD springs etc with the high rates are designed for. Or if you're running 13s or 14s with a big sidewall on the tires then maybe it's a bit nicer.

Just in case my blathering above didnt make sense, here's some more nonsense to ruin the pictorial part of your brain also: Basically... there are 6 axes of movement rotation which a joint potentially needs to do. Rotation on XY, XZ, and YZ planes. Then movement in the X, Y, and Z directions. Rubber pretty much allows all of the above to occur. When the bushes get more worn out, they allow even more XYZ movement. Generally when replacing rubber the idea is to reduce the 'slop' (X Y and Z movement, which makes it feel like the car wobbles on the diff) Different 4 link ends restrict different axes in different ways, a Rosejoint restricts down to the three rotational axes and removes all XYZ movement. On my car, the upper 4 links (non triangulated) have cutouts in the rubber bushes to allow increased movement in the X axis but keep the Y and Z axes stiff. However the point of contention is whether or not the end of a 4 link arm needs to allow for X axis movement on a triangulated 4 link. Theoretically the answer is yes, and some peoples broken floors and worn out rosejoints say yes as well.

In my opinion / dont own a starlet / grain of salt / etc Rosejointing all links on a 4 link whether triangulated or parallel is a bad idea. Imagine that you had 3 out of your 4 links connected. These are triangulating the position of the diff to the body so that it moves through a particular arc, where ever it is moved to. If you moved the diff through different arcs and measured the length that the 4th link would need to be, it would lengthen or shorten slightly depending on the position of the diff at the time you measured it. As since the diff is already triangulated, a 4th link causes problems. Hence some race cars going to a single upper link on one side of the diff head. The rubber bushes allow movement as needed - so does urethane - but rose joints do not. Since the 4 link arms can no longer lengthen or shorten at all through travel the end result is usually a combination of cracking the floor mounts, wearing out rosejoints quickly, or causing everything to flex to account for that difference instead. I'm sure it would be possible to sketch up a cad model to show how much 4 links need to grow or shrink by, it might only be a few mm or less. But it's bad idea in theory / in my opinion. I'm not much a fan of urethane replacements either, where the factory rubber usually has a steel pin through the middle which accounts for side to side movement of the arm, urethane normally have like a big 'washer' type shape which stops movement in this direction too. Which can cause binding as well, in some cases. Out of those options I'd go for urethane one end, with the urethane trimmed down so it looks like factory rubber (no washer type shape on outside) and rosejoints at other end. Maybe with a hole or two drilled in the urethane on the upper links to allow slight growing/shrinking of length.

Awwww yeah cut dem holes