7/8th Inch Oil Drain Line

One annoying thing about The Dog since i started going turbo was the occasional smoke from exhaust. I've tried restricting the oil supply line with custom banjo nipple bolt, installing an oil catch tank and connecting it to the turbo inlet in hopes to reduce some crankcase pressure and replacing the oil drain flange on the sump with bigger pipe with no significant success. Those steps have managed to eliminate the smoke during deceleration but that's it. Most annoying smoke problem is during idling after the engine has warmed up such as traffic light stop. I had to blip the throttle every few seconds to avoid having white thick smoke coming out from the exhaust. No fun for sure.

One last thing i could do before blaming the turbo oil seals is to enlarge the drain line and that was what i did today. My drain line initially was 5/8 inch (~16mm) and i decided to try with 7/8 (~23mm). Since finding a stainless steel braided hose with such diameter is nearly impossible in this small town of Bandar Lampung, a radiator hose was chosen as an alternative. Here's a comparison shot, left is my old steel braided line and the other is the 7/8th ID radiator hose.

I also decided to ditch the oil sump drain flange and weld the drain pipe directly to the sump. No particular reason, just because.

Here's the finished work.

Yes, that looks dangerously close to the exhaust manifold. We'll see if that radiator hose sans stainless steel braids would handle the heat.

Probably too early to tell, but after it's done, i let the car idled for 15 minutes, drove for another 15 and not a hint of smoke can be seen from the exhaust even during idling. Fingers crossed!

Proper Oil Catch Tank

The Dog has been having occasional heavy white smoke coming out of the exhaust since the turbo conversion. There could be some causes for this but i ruled out most of them already like worn piston rings or valve stem seals. What's left is the probability of inadequate oil drain line, bad turbo seals or clogged PCV line.

Since the smoking only happens occasionally and mostly during idling, bad turbo seals are unlikely as this usually happens even when cruising. What's left is now either inadequate oil drain line or clogged PCV line.

Being cheap, i used an old AC dryer bottle as my oil catch tank. Since i didn't plumb it back to the intake line, it's possible that this isn't enough to relieve the crankcase positive pressure and causing the oil not draining properly from the turbo and finally seeping out of the seals and smoking. Fortunately, i found a used Greddy oil catch tank replica at a bargain. I bought the tank and install it myself. Here it is in action.

It's not uncommon that PCV lines are run incorrectly after a turbo conversion. Most forgot that now their intake manifold will also see boost pressure and they didn't change the original PCV line. As far as i know, stock 4A-GE PCV line don't have any check valve. It just goes straight from the valve cover to the intake plenum. After a turbo conversion, you'd need at least a check valve there. You don't want any positive pressure from the boost to leak into your crankcase. This will blow your seals prematurely and, as i suspect it earlier, could cause improper oil drainage from the turbo.

If you don't want any check valve, at least you'll need to move the line going to the intake plenum to further upstream. This line should only see, at maximum, atmospheric pressure and this, on a turbocharged engine, can only be found at the compressor inlet. Thus my line goes from the valve cover straight to the catch tank and from there it goes to the compressor inlet, routing the line above the intercooler and no check valve necessary.

Did it solve the problem? Not completely but it's better now. I still have some smoke every now and then but it's noticeably less frequent. All there's left to do now is enlarge the oil drain line. It's annoying that i have to take off the oil pan just to do this but i guess there's no other way if i want to get rid off the smoke completely..

Megasquirt Boost Control

Having realized on the dyno that manual boost controls are not the most reliable way of managing boost levels, I decided to see if i could utilize the boost control feature in the Megasquirt. The conversion work seems minimal as there were only a handful of electronic parts to add and, in my case, no dissecting the original MS box necessary. The most important part, however, was not the easiest to find. I had to go back to Jakarta to find the solenoid valve, not even sure if i could find one for 12VDC. The horsepower gods must have smiled at me and let me get back home with this.

This being my first time fiddling with solenoid valves, i had no idea if the one i got would be applicable for PWM drive as that is how electronic boost controlling are done. For the price, however, i decided to just try it. Better than to get back empty handed, right?

Remember this?

That was my junction box for the Megasquirt, a similar idea to the official DIYautotune's relay board. The idea was eventually ditched as the terminals became gooey and melted from being placed too close to the turbo (which i have lately realized probably due to my AFRs being too low which increase exhaust gas temps, going 12.5-13 works surprisingly well in significantly reducing under hood temps). I decided to use the junction box to house the boost control electronics and, having learned my lesson, place it away from any heat source.

The drive circuit on MSExtra hardware manual page is using IRLZ44 for the PWM MOSFET. I couldn't find one locally but got IRFZ44 instead. Tried to google the differences and i couldn't find anyone mentioning that they shouldn't be equivalent for my application. I don't really understand MOSFETs as well as the old BJTs but i do know they're voltage controlled, unlike BJTs that are current controlled. Realizing this i was worried that MS CPU output being only 5V logic level isn't enough to drive the IRFZ as it was the IRLZ that was designed for logic controls. How did it go then? Read on.

Just a few notes on the transistor: a heatsink was not really necessary as Rds on the MOSFET was very low and with the current from the solenoid being less than 1 Amp, the dissipated heat was minimum. I add the heatsink just because there was enough room for one. Also, if you can't find the MOSFET, use TIP120/121/122 instead. The Darlington power BJT is being used to drive the idle PWM solenoid at 150Hz. Boost control PWM setting on the MS are only 39Hz max so it should be up to the task. You need a heatsink for this one though.

I was lucky the nipples from the manual boost controllers matched the solenoid threads.

The ugly strut bar serves me more than it was designed to, apparently. It was the best place i could find to mount the solenoid. Also, the world would crumble and die if zip tie isn't invented yet.

As mentioned, the box is now placed away from any heat source. It may seem a waste to house just four components inside the box but i decided to have the extra space as room for improvement later. Who knows later i'll use the water injection or knock sensing feature?

The five wires coming out of the box, from left to right, are designated as the following:

1. To MS pin 36 (known as X4/JS2 output)

2. To Solenoid Valve (-)

3. To Solenoid Valve (+)

4. Ground

5. Ignition Power Supply

Once the hardwares wired, the following need to be set on MS.

1. Set X4 (JS2) function under Codebase and Outputs function to Boost Control

Note: you can test your wiring by temporarily set this to Output1 and somehow play with the setting so that Output1 is triggered by TPS. This way, you can check if the solenoid clicks by playing with the throttle with the engine off. A click on the solenoid tells you sure that the wiring is correct.

2. Set the Boost Controller Parameters

The numbers shown works in my case. You need to find one that works for your solenoid. Focus on getting the right PWM rate first. Typical ambient pressure can only be set if you disable Baro Correction on the Constants window. You need to set this to the value of the MAP sensor reading when the engine is off. This is the threshold to activate the Boost Control which is basically the turn point between vacuum and boost.

3. Put low numbers on Boost KPA Target table

This is deliberately set so that it defaults to open loop boost control all the time. MS1 CPU is said to be iffy on closed loop boost control as it tends to overshoot the target which is very dangerous to the engine. That's why it's better to use open loop which is what most electronic boost controllers are using. To make sure you get open loop all the time, set Closed loop kpa limit (kPa) on number 2 above to low numbers as well. The way i set it, MS already goes to open loop mode at 80 kPa (70 + 10) and even at this reading, the boost control is not active yet as it's still below the Typical Ambient Pressure value!

Also, since you're doing open loop all the time, the Proportional and Differential gains are now irrelevant.

4. Start with low numbers on the Duty Cycle table

This is the open loop table. The higher the Duty Cycle, in my case, the higher the boost as the solenoid leaks more air out from the wastegate actuator line. You need to start with low numbers and rise slowly until you reach your target kPa. Once you find the relation between DC and boost level, you can alter the DC table so that the turbo spools faster. You can do this by setting the DC to 100 just before the RPM where your turbo spools.

I haven't really tuned my Duty Cycle table yet but from a test ride yesterday, it seems the conversion was a success. I managed to reach higher boost level than the stock wastegate setting. All there's left to do is to find a time to tune the Duty Cycle safely at night. Choosing Megasquirt was really one of the best decisions i made on this project!

Today I Learned: Turbo'd Carb & Dizzy Relocation

On the last post, i mentioned about Nico, a friend i met in Lampung. Nico rides a KE20 but this one is not your ordinary "humpback" Corolla.

Here's why:

That's a 4A-GE (rough guessing from the spark plug leads, bigports) converted to carb and turbo. This being the my first time to see a running carb turbo engine in person, to say i'm amazed was an understatement. I have always wondered how can carburetor work with positive air pressure coming from a turbo and now i get the chance to learn how it's tackled.

After seeing Nico's engine in detail, i learned that the key in allowing turbo carb is this (see arrow):

That is called a rising rate fuel pressure regulator. This is what's stopping you from pushing pressurized air into the carb jets and subsequently pushing those fuels out of the float chamber. This works by sensing the air pressure on the charge pipe and via some membrane and spring system, ensures that the fuel pressure from the pump is always above the air pressure coming to the carb. Typically it's set to 3-4 psi above the air pressure. Here's a short yet good writing about rising rate FPRs.

That's not the only thing i learned from Nico's 4A-GE. Notice where the dizzy is on the engine? Since the engine's converted to carb, the stock distributor is useless since it's designed to work with the ECU. One way to solve this is to use a point-and-condenser dizzy. I don't know what engine's distributor is usually used but i think this one's from a K-series engine. Correct me if i'm wrong. Anyway, that's not my focus now. I'm more interested on where the dizzy is.

Although usually dizzy relocation is more of a 20V RWD conversion thingy as this will keep you from ruining the firewall, it's a good idea also for turbo application as you're now moving the melt-prone dizzy cap away from the turbo heat. You'll also be spared from leaky distributor oil seal.

Seeing it in person, it looks simple. The most complicated work i believe is modding the dizzy shaft so that it'd match the slot you that you machined on the cam sprocket bolt. I think Nico's slot was too wide as i still see a small gap that would allow a small play on the dizzy. This is not ideal as your ignition timing won't be firm but since you're already using a mechanical point dizzy, i guess precision ignition timing is not really what you're after, now is it?

So, imagining myself doing the same work, here's, in order, how i would do it :

1. Make a slot the cam sprocket bolt.

2. Mock up the placement of the dizzy mounting plate. This is so that you can measure how much you need to cut the dizzy shaft.

3. Once you shorten the dizzy shaft, make the key on the shaft that match the slot on the sprocket bolt.

4. Mount the dizzy, done. Make sure the mounting plate is thick enough and firm to withstand the vibration.

Last but not least, Nico also took me to my new playground here in Lampung. I've been looking for a junkyard here and finally i found one! Here's one view of the junkyard..

This is a great find for me.. Why? See the pic above. Hint: this will solve my gearbox problem. Got it? Well, this junkyard have some W58 gearboxes! Now i don't have to hunt one in Jakarta!

Budget Oil Catcher

Ghetto el-cheapo oil catch tank using old AC drier.

Manual Boost Controller

I have no idea what is the spring rate on my turbo actuator but since this is so cheap, less than 12 bucks to my door steps, bought it anyway. If the actuator spring's 0.8 Bar, then this is useless. If it's less, then it's good.

Oh, here's a little tip just to brighten up this boring post.

The Dog, Doped

Kicked myself out of bed early this morning and got small progress done on the car. Small yet quite a milestone as today is the first day ever The Dog's intake manifold sees higher pressure than atmospheric level. Yes, today the turbocharger finally functioned properly.

First thing done was modifying the Megasquirt's VE and Spark table for forced induction. Here's my current VE table. Never take my table seriously as it's not dyno-tuned yet, not even on the non-turbo region.

Next thing done was replacing the stock MAP sensor with 4A-GZE one. Aside from a minor confusing problem, i had to increase all values on the Megasquirt VE Table by 30% which should be unnecessary if both MAP sensors are mapped correctly to Megasquirt, the work was done quickly and i managed to get The Dog running on decent AFR with the 4A-GZE MAP sensor. Megasquirt's now all ready for forced induction!

Next thing i got done was putting the Greddy Blow Off Valve on its place and hooking the vacuum line to the intake manifold. With the last hole on the charge pipe now covered by the BOV, there was no more boost leak source and all those air from the turbo has no where else to go but to the cylinders et Voila! Positive intake air pressure!

Yeah i know, i need tidy up the under hood wiring. Eventually i'll do it but not now.

Flicked the ignition key, hit the throttle a few times on neutral, checked for any leak and smoke and off to the street we go. Put in first gear, clutch off, engine rev steadily rising, and just above 3000 RPM she launches! Albeit difficult, I keep my ears open for any knocking and my eyes on the AFR reading which is somewhere on low 11s. Everything seems fine.. Took right foot off the throttle and there it is.. the most beautiful sound for turbo junkies, BOV scream! Shift second gear, repeat BOV scream. Damn it sounds sweet. I didn't dare to go to high RPMs just yet cause obviously it's not yet tuned properly and there is still one important step missing: replacing the injectors. I don't know if the stocks are sufficient for forced induction but i'm not willing to find out the hard way.

Intercooler & Fender Bar Part 3

Things move a bit slow since work is overwhelming lately. I managed to squeeze a bit time to finalize the intercooler piping and here it is.

Turbo to Intercooler Charge Pipe

Clamped and ready..

Bumper space. Also this shows that the vertical bar where the grill mounts which goes from the latch to lower front chassis can no longer be use.

I was lucky to be able to use the stock bumper brackets without any mod.

Intercooler & Fender Bar Part 2

Turbo --> 2 to 2.5 inch coupler (not included in the Ebay intercooler kit) --> L-knee pipe, shortened about 50mm on the turbo side --> straight coupler --> about 160mm section of straight pipe --> 2.5 inch 90 deg silicone elbow --> intercooler inlet

Intercooler outlet --> 2.5 inch 90 deg silicone elbow --> L-knee pipe with BOV flange prewelded --> 2.5 inch 90 deg silicone elbow (not yet delivered, kit only has two elbow hoses) --> TB inlet.

Intercooler and Fender Bar Installation

Just learned that the Ebay intercooler is too big. Just for the record, the intercooler is 31" in length (inlet to outlet, not core size), 11.5" wide and 3" inlet diameter. It's not impossible to be installed but it would take more effort to put in than smaller ones. I learned from d1davey's blog that the best size should be around 510mm in length but i believe this is core size, not inlet to outlet length. I bought one of 520 x 240mm core size with the actual total length being around 680mm and i believe it's a perfect fit. Tonight we'll start the actual installation. I'll update this post with how it goes.

Meanwhile, the Ultra Racing AE86 fender bars are in! They fit nicely with only minor modification required (enlarging the bolt holes a bit and some light hammering on the frame). For the front mount, the nuts had to be welded from the inside of the frame. Just be careful when installing the door hinge part, you don't want to have to realign the doors. This really confirms how similar AE86 frame to Charmant.

The workshop i got this done has some interesting line of cars there. Ford Capri coupe, VW Dakota bus and this stunning Holden Belmont Ute. Talk about slammed!

Nearly There

Got the engine running again finally. Since pictures worth more than thousands of words (or i'm too lazy to write), here are some progress shots with minimum amount of words in between.

ARP flywheel bolts

Turbo oil drain flange on sump

ARP head studs and 4A-GZE 8.9 CR slugs

ARP head stud protrusion. This is why some need to shave one of the studs as it fouls the distributor shaft. Obviously not this one as i took photo of the wrong side.

Considered leaving the timing belt cover off but put it back on to be safe.

Last shot before turbo conversion the next day. Notice the oil drain line is only ziptied near the strut tower.

The next day, after running the engine for a day and ensuring the rebuild was successful, turbo conversion was started. Oil sandwich adapter went in first. Only one out of four outlets was used.

HKS copy cast iron manifold with gasket cut out from a soda can. Cheap and refreshing, indeed.

The very minimum space at the back of the turbo with downpipe flange off. Gotta move some wiring to protect them from the heat.

Oil lines plugged in and this is the final shot of the day.

The next day which was yesterday, exhaust was fully redone. The initial plan was only to fabricate the downpipe and connect it to the existing exhaust but after quick reconsider, decided to ditch the whole system and put 2-inch pipes front to rear. Only 2-inch? Well, the turbo downpipe flange output is that diameter, i'm merely following. No use in going bigger, yes? The flexible joint was welded to the downpipe and only one tiny muffler was used. The rest of it was 2-inch piping and elbow joints to allow going over-axle. Oxygen sensor bung was installed on the same location as the original piping. Didn't take many photos that day, sorry.

Finally, just today, i installed an SPST switch on the bootloader jumper header on my Megasquirt. I did this to ease flashing firmwares later without having to open the MS case anymore. I'm currently focusing on finding the correct values to go into the kpafactor and barofactor inc files as i'm going to use the 4A-GZE map sensor which is not directly supported by the original MS1-Extra firmware i'm using.