ajg193 Posted October 11, 2017 Author Share Posted October 11, 2017 I did some fuel consumption calculations today. The MicroSquirt/TunerStudioMS seems to over estimate fuel consumption by about 10% - My belief is that this is due to my injector dead times being larger than estimated in the software. Then again, this could also be due to my data being collected from an ill-conditioned system as I only based these results off of about 35 km of driving. For the fuel consumption tests, I had the headlights on and held the car at a steady(ish) speed on a flat road to take the readings. For the fuel figures obtained for anything other than 5th gear, there was a fair bit of noise in the consumption signal as I had difficulty writing down the numbers while driving (a passenger would be useful next time). Also the samples were taken over a very short time period, with slight undulations in the road etc. For the 35 km trip, the computer estimated a fuel consumption of 2.26 L. The actual fuel consumption was 2.04 L. This gives a fuel consumption of about 5.9 L/100 km, which is very good considering I wasn't being nice to the car during this time. It's actually on par with the best result I ever had with the carburetor driving like a granny. Fuel consumption at 80 km/h in 5th gear is about 4.2 L/100 km. I think the 4th gear value is unreasonably high and I will have to take the reading again at some point. The same goes for the 100 km/h 4th gear result as earlier indications today suggested it was about 10% worse than in 5th gear. The values in the graph have been corrected by 10% to allow for computer error. Judging by these numbers, if I drive 70 km/h constantly (at sea level, without wind, in the dry) I should be able to make it to Auckland on a single tank of fuel. Conclusion: Car is great on fuel now, more testing is needed. 7 Quote Link to comment Share on other sites More sharing options...
ajg193 Posted October 13, 2017 Author Share Posted October 13, 2017 With the help of Jane we managed to add some more data points to the graph. Roads were kind of all over the place though so the data is still noisy. One day we should have enough data to make smooth graphs. 3 Quote Link to comment Share on other sites More sharing options...
Popular Post ajg193 Posted December 24, 2017 Author Popular Post Share Posted December 24, 2017 I finally decided to tackle the rust hiding under the hatch glass seal. I was hoping for surface rust but found a few holes. My brother cut it out and welded in some new steel for me and we then repainted the whole door using epoxy primer and 2 pot paint. The colour is pretty much spot on and the surface is pretty good, there is a little orange peel but not much different to factory paint. I still need to put the sprint badge back on (I'll do this when I get back to Christchurch as I have a better one down there) and I need to repaint the TOYOTASTARLET badge on the boot. I'll probably do a white outline around the letters or something. Factory is silver in the STARLET part and surrounding the TOYOTA part, with the rest of the badge black. Also as a bonus: I've now well and truly more than doubled the km on the car. Oh and I put a brand new radiator in the other day as the old one was getting to a point where the fins were crumbling at the slightest touch / every time a bug hit it the engine would warm up a bit more. 11 Quote Link to comment Share on other sites More sharing options...
ajg193 Posted January 2, 2018 Author Share Posted January 2, 2018 Anyone know where I can track down some brand new rear springs for the Starlet? One side is sitting about 20 mm lower than the other, I think it may be a result of hauling reasonably heavy loads in the car for long distances a couple of times. Looking for factory height and preferably factory stiffness or slightly stiffer. /justcompresstheothersideby20mmandyoullbefinemayte 2 Quote Link to comment Share on other sites More sharing options...
ajg193 Posted January 12, 2018 Author Share Posted January 12, 2018 Thanks guys for the suggestions. Tried oldscool but they reckoned they'd come from Chch anyway so I just bought a pair from Chamberlains down here. New springs are in. Look at that lush height! Before: After: New springs (bit of a bugger they don't come in black though): Old springs: Next to eachother: Now sitting nice and level again: The car seems more flat around the corners now, but I have only driven it like 1 km since I put the springs in. It was an easier job than I expected it to be, just undo the top nut on the shocks, jack the car up and pull the old springs out and put the new ones in. I marked the old ones left/right and top/bottom so I can get them back into their original spots if I ever need to. I can now crawl under my car without having to jack it up so that's a bonus. 7 Quote Link to comment Share on other sites More sharing options...
ajg193 Posted March 3, 2018 Author Share Posted March 3, 2018 I changed some settings in the tune, I found that at 4 squirts per cycle I was hitting areas where the pulsewidth was just too low to accurately deliver fuel (like 2500 rpm with no load on the engine). I changed it to 2 squirts per cycle so that the pulsewidth is now essentially doubled. I think that this has helped with that problem, and doubling the PW at idle is probably also reasonable as well as at 1.5 ms before I reckon it wouldn't have been far from the nonlinear region. I took the car on a good thrash through the hills to retune the VE map and whatnot. Shown in Figure 1 from the datalogs. Figure 1, VE map Don't worry about it being greater than 100 (colourbar is the VE value), these numbers are essentially a nondimensional scaling factor, that coupled with the fact that I am running lower fuel pressure than the injectors are rated at (I'm running 3 bar above manifold pressure). You can see above 5000 rpm it really starts to drop off significantly, but it peaks around 3500 rpm (I'd as that this is around the max torque point) I also used the Duty cycle values from the injectors to calculate approximate engine power, assuming a 100% duty cycle flow rate of 210 cc per injector, an energy density of 34.2 MJ/litre of fuel and a thermal efficiency of 25% across the board (which is a huge assumption, but I'd say is fairly typical for a petrol engine). The results of this calculation are shown in Figure 2 Figure 2, power curve The colourbar is manifold pressure (in % of barometer). This would give a peak power of around about 58 kW (79 ps/hp), and the curve pretty much perfectly overlaps the one given in the JDM Starlet brochure for the 4K-E. I'd say that the power is probably over estimated by about 5-10% but isn't too far off. I've also found that the car can now get up Hackthorne Road in 4th gear quite happily where I would have needed to stay in 3rd gear with the carb setup. Figure 3 shows a map of the pulsewidth vs RPM. The colourbar is the manifold pressure value. The line of points along the RPM axis/0 PW is where fuel cuts were active on deceleration. Figure 3, Pulse width map Figure 4 shows the observed AFR of the engine, mapped to manifold pressure and RPM. I'm generally holding around the 13.0 mark at full throttle and cruising at around 16.5. My idle AFR is currently a little high at around 14.5, I'll probably drop this down to around 13 as that seems to minimize any effects of heat soak if the engine has been sitting for 10 minutes or so. I probably also need to adjust the VE values on the trace route back down to idle as occasionally it will miss the idle and stall after a hot start since I changed from 4 squirts to 2 squirts (used to be good with my old map). I'm currently not running any EGO correction, I may enable it at some point in the future once I am convinced that the O2 controller starts reliably (I've been having issues recently where it sometimes doesn't output anything to the ECU, but I seem to have gotten it to work reliably again). Figure 4, Observed AFR What is interesting to see from the VE map/PW map is the general operating range of the engine, where it's pretty much all used below 3000 rpm but above that I am generally either fully on the gas or fully off the gas. Then there are some areas which are never hit, such as low RPM/low manifold pressure, or just about anything below 1000 rpm in general. 1 1 Quote Link to comment Share on other sites More sharing options...
Popular Post ajg193 Posted October 8, 2018 Author Popular Post Share Posted October 8, 2018 Got a WoF for the Starlet today, it went through without any hassles. After the WoF I hit up Butler Auto Mart and found they had a set of NOS headlight covers for slant front Starlets in stock. They'd been on the shelf for a pretty long time (4 years in the computerised stock system) and the cashier threatened to charge rent on them. They're stupid and ugly but I just had to have them. Who knows, they might even offer improved fuel economy from reduced drag - it's practically a Prius. 17 Quote Link to comment Share on other sites More sharing options...
ajg193 Posted March 7, 2019 Author Share Posted March 7, 2019 I've finally gotten around to making a return fuel port for the tank, I'll put it on the car either tomorrow or Saturday. I should probably zinc plate it first, or do you guys have any paint recommendations? It has 3/8 feed and using the original 1/4 feed as return 1 Quote Link to comment Share on other sites More sharing options...
Popular Post ajg193 Posted March 30, 2019 Author Popular Post Share Posted March 30, 2019 Recently I've been developing the ignition system for this car. Ever since I put a brand new starter motor in, I was experiencing interference with the ignition system during cranking so I decided to finish building up the EFI distributor that started life in a 7K-E. The 7K-E distributor differs from a 4K distributor in that it has no advance mechanisms, is about 10 mm longer, has a bigger cap diameter and different rotor. Usually a 7K-E distributor cap is of the type that you clip the leads onto instead of pushing them on. I found a 22R distributor cap at repco that pretty much fits onto the 7K-E distributor, only needing the mounting holes drilled out slightly. This allowed me to use the original push on 4K leads. I also had to make up a spacer plate to put between the distributor and the engine block, this was relatively simple to just make out of some steel. The 7K-E distributor uses the same rotor as a 4A-FE and I'd assume 22R. This rotor worked perfectly with the cap. Having only 4 points for the VR pickup, the distributor wasn't exactly suited to a megasquirt based EFI system. Toyota used a system known as VAST, which basically used the distributor for base timing and would fire a spark when a signal came from the distributor unless the ECU overrides the module. I had no VAST module, so this setup wasn't exactly ideal as the megasquirt would just have to guess at the base timing (10 deg BTDC) based on the average speed of the engine. If using a basic 4 point distributor trigger setup you need to phase the distributor so the trigger comes in earlier than the most advanced timing you are going to use (ie 50 deg BTDC). Anyway, I did try to use the distributor with the factory 4 point setup (this was before I changed to the new starter), but I had problems with getting the engine to start. If the engine started, it worked somewhat well but the timing as no very precise. It was pretty clear that more data points (teeth) were needed. After that failed attempt I changed back to fuel only EFI and just kept using the car. Then the original starter motor stopped, the solenoid contacts were a bit burnt. I rebuilt the solenoid but I didn't trust it to work again so I bought a new aftermarket starter. The new starter was interfering with the ignition during cranking, causing tach spikes. The starter would also crank at a randomly selected speed every time I turned the key. This was making it somewhat annoying to start the car as it wouldn't always start on the first try. So I decided to go back to ECU controlled ignition and see if I could get it to work properly. I was considering using the flywheel of the engine as a VR trigger wheel, but I think they have 105 teeth, which isn't divisible by 4 or 2 so it wouldn't have worked with the megasquirt. My Barry lathe came complete with a broken quick change gearbox, and it just so happened to have a 20 tooth gear with an outside diameter of 35 mm - perfect for the distributor. I clocked up the distributor shaft in the lathe and turned the toothed portion down to 19 mm so the gear would fit over it. I welded the gear on, removed some of the thickness and knocked a tooth off to make it a 20-1 wheel. This first attempt actually worked quite well, the car ran great and the timing was accurate but I still had issues getting sync while cranking. I figured it was due to the small diameter of 35 mm not giving enough tip velocity at cranking speed to get a good signal from the VR pickup and decided to redesign the system. I used this setup on the dyno, and tuned in a good ignition map. The next attempt was to make a trigger wheel that went over and around the VR sensor, almost doubling the diameter of the wheel (so almost double the signal). The VR sensor needed to be modified slightly to sense from the outside. This modification consisted of welding a bit of gauge plate to the existing sensor so it would stick out more and removing some material from the VR sensor housing. I epoxied the sensor together after, to reduce any risks of bits falling out over time. I made a 30-1 wheel for this, but as soon as I put the details into TunerStudio, the ECU said "no". I somehow missed the fact that 30 isn't divisible by 4 and the ECU wasn't having a bar of it. The hole in the wheel is mainly to allow for screw driver access for assembling/disassembling the distributor, but it should also help (or make worse) the balancing from the missing tooth. In a bit of a huff, I ripped it all out and went back to fuel only mode for a few days before I built up the motivation to cut all of the pretty teeth off the wheel and try again. I used my Barry lathe to remove all of the teeth and then used a mill with a DRO to put 23 holes on a 61 mm PCD (set to 24 equispaced holes). These holes were 2.4 mm. I then cut up a 2.4 mm diameter mild steel TIG rod to use as teeth and welded them all in. After making the wheel, it was a simple case of heating it up, slipping it over the distributor shaft and hoping it would stay put - it did! Then I faced off the welded surface and reassembled the distributor. At this point, the new starter motor started to smoke and quite often not even start turning. I put the original starter back in the car and it worked perfectly (turning twice as fast with half the power). The car started up pretty much first try. After adjusting the distributor phasing in TunerStudio it now works perfectly and starts first time, every time. Looking back, the 20-1 setup probably would have worked fine using the original starter motor but this new wheel should be slightly more robust to signal noise. The car just runs generally slightly better all around. It now uses about 15% less fuel at idle, about 3.5 L/100 km at 50 km/h and 5 L/100 km at 85 km/h so I can't really complain. I also bodged up a cover for the J121 igniter module, so it looks a little more factory. Hopefully I will get around to making it a little more permanent one day but this will probably last 20 years. So basically I achieved my goal of having an OEM looking ignition setup that uses the ECU to control the timing. 12 Quote Link to comment Share on other sites More sharing options...
Popular Post ajg193 Posted March 30, 2019 Author Popular Post Share Posted March 30, 2019 The car is also running the full return fuel pickup now too Zinc plating was only $5 10 Quote Link to comment Share on other sites More sharing options...
ajg193 Posted November 29, 2019 Author Share Posted November 29, 2019 You guys should all subscribe to this welsh hippy on youtube: He also did a video on @h4nd's car 7 Quote Link to comment Share on other sites More sharing options...
Popular Post ajg193 Posted June 3, 2020 Author Popular Post Share Posted June 3, 2020 Finally all legal-schmegal I guess this means I can now fit a T50 and massively work my engine without a WoF guy getting grumpy (not that I'm going to) An item of note is that the high pressure fuel hose from SuperCheap isn't EFI rated, so don't buy it. It's also not really rated for being in engine bay. Just get some good fuel hose from BNT, it doesn't even cost that much more. (Must be SAE J30R9 instead of J30R6) Hopefully WoF guys don't get grumpy about the red springs in the back of my car as they aren't technically OE... 13 Quote Link to comment Share on other sites More sharing options...
ajg193 Posted June 18, 2020 Author Share Posted June 18, 2020 Car was making a bit of a belty noise the other day so I pulled off the alternator belt. Pulley on the alternator had a little slop in it so I stripped it down. The front bearing seemed to have a bit more play than I'd like so I threw a new one in and vapour blasted the front half of the alternator (illegally borrowing @ProZac's blaster). I went to put it back together but my OCD couldn't cope So I pulled off the back half and cleaned it too Anywho, it looks a lot better now and doesn't seem to have any unreasonable play. I also put a new belt on, last one was about 6 years old. Nice and quiet now. $22 fix 9 Quote Link to comment Share on other sites More sharing options...
ajg193 Posted June 18, 2020 Author Share Posted June 18, 2020 Before and after for back half 3 Quote Link to comment Share on other sites More sharing options...
ajg193 Posted June 18, 2020 Author Share Posted June 18, 2020 As requested by @Roman, here are recent fuel consumption values, night time, in the rain, heater on, reasonably level roads/motorway 5 1 Quote Link to comment Share on other sites More sharing options...
Popular Post ajg193 Posted August 8, 2020 Author Popular Post Share Posted August 8, 2020 So I decided "right, I've had enough of that really slow leak from the sump gasket. I'll whip off the gearbox, pull the sump from the engine and sort that stuff out" Upon removing the sump I found some cake that looked like bearing material. Connecting rod #1 bearing didn't look the most flash so I decided "out with the engine" Upon removing the engine, most of the bearings weren't too bad but definitely worn. Next task was pulling the pistons Two of the pistons are cracked. A little googling indicates this is a pretty common thing on 4Ks. So the plan now is for the following: -port match intakes on head to manifold/gasket -drop cam and lifters off for a grind, something pretty mild -drop block off for a hone (maybe bore, but the bores look pretty good), new rings and pistons, block may potentially need to be decked if using aftermarket pistons -???? -profit* *probably huge loss, but maybe a gain in power The stupid thing is the car was running absolutely perfectly. I probably shouldn't have opened the can of worms, but I'm kind of glad I did. After this is all done I should hopefully be earning enough money/retired by the next time it needs to be done that I can just pay someone else. 11 Quote Link to comment Share on other sites More sharing options...
Popular Post ajg193 Posted August 12, 2020 Author Popular Post Share Posted August 12, 2020 Cylinder head rebuilt, intakes port matched to manifold. Exhaust ports tidied up near valve seats to remove steps. All ports somewhat cleaned up for casting irregularities. Another 0.12 mm skimmed from head to clean it up. Next task will be to make a vacuum testing tool to ensure the valves still seal after I lapped them. Also collected the cam and lifters from Kennelly cams, m6-7 grind - whatever that means. Dyno tests have shown it to take a bone stock 4k from 64 to 76 HP or something with no other modifications. I believe the correct technical classification is "3/16th BSP*" *Barry spec performance 16 Quote Link to comment Share on other sites More sharing options...
Popular Post ajg193 Posted August 15, 2020 Author Popular Post Share Posted August 15, 2020 Cylinder block tested up fine, so is just getting hone and rings. I bought some second hand pistons from Kennelly cams - inspecting and crack testing each one from his pile to find the four best. I made a valve leak testing tool yesterday and used it this morning on the head The threaded nubs go into the big block, allowing for it to be connected to a vacuum gauge or sensor and to a vacuum pump (smaller block just goes into the vacuum pump's port). An o-ring is glued into that groove on the big block. All valves sealed perfectly with springs on, one wasn't sealing with the spring removed though - there was a bit of finger nail stuck between the valve and the seat. Removed that and now they all seem to seal perfectly. 13 Quote Link to comment Share on other sites More sharing options...
Popular Post ajg193 Posted August 21, 2020 Author Popular Post Share Posted August 21, 2020 Engine is all back together and running now. Broke it in for about 100 km and did a quick power run, running a bit lean at the top end (I manually adjusted the VE table after, will have to let auto tune hit that region one day) so started misfiring. It gained more than 10 hp though and didn't even hit peak by red line. Once it is fully run in I might consider lifting the rev limit to 6500. On the other end, it lost a fair bit below 3000 rpm, down 17% at 2000 rpm compared to before. Fuel consumption when cruising on open road looks to be about the same as before: 60 cc/min at 80 km/h (4.5 L/100 km) 90 cc/min at 100 km/h (5.5 L/100 km) The idle is a little bit screwy with the new cam, but could be worse. I kind of feel like I have ruined the car's soul by doing the cam though... I also had a little issue with the idle oil pressure getting low enough to make the light flicker when the engine is fully hot, threw a gauge on it and confirmed it makes /some/ pressure, pressure rises quickly above idle. Peter at Kennelly cams reckons that it's pretty common on 4Ks, he suggested going to a thicker oil. Now running 20W-60 and no more lights on, does seem stupidly thick though... I don't really have any new pictures, so here's the one I posted in the spam thread 8 2 Quote Link to comment Share on other sites More sharing options...
Popular Post ajg193 Posted September 4, 2020 Author Popular Post Share Posted September 4, 2020 Low oil pressure was keeping me awake at night and I just couldn't cope So I ripped the engine out this afternoon and dropped it off for crank grind and bearings. He's also going to throw in an actually mild cam this time. At least one of the lifters wasn't rotating so it was probably good that I pulled it apart Less than 2 hours to get it out though so that's a win I guess 10 1 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.