Roman's 4GR V6 Toyota Carina

[Roman]

I've been chipping away at a bunch of tasks and tidying things up, after the excitement of first drive has faded.
But one thing on my mind which is - Man I honestly hate how it sounds at the moment.
It's not right for the car at the moment and not very pleasant overall.

This car has only had two objectives... Do a shitload of rpm, and sound cool. 
I've been fixated on sorting out the nuts and bolts of getting the motor fitted and running.
Then dealing with some "shitload of rpm" problems. 
But the current noise even just idling. Just seems so discordant to the rest of the car.
It sounds like a straight piped AU Falcon or lil johnnys RB20DE. 
Honestly, I have looked up on youtube some straight piped falcons and RB20s and it sounds just the same haha.

How to improve? 
Here's what I'm hoping: 
-Will sound better once I've got the fuel map sorted a bit better, and some cam overlap chucked in there
-Will sound better once trumpets are on (was just bare throttles currently)
-Will sound better once it's doing a full throttle nang up the road
-Will sound better with big cams in it
-Might try keeping the 2-1 pipes smaller diameter, (currently 3") then flare out to 3" 
-Will try some different muffler options 
-Might try splitting exhaust back out to 2 pipes 
-Smaller tip diameters on the muffler (currently 1x 3" in, 2x 3" out)

Ideally want to get rid of the bassy exhaust note a bit.

Given the variety of sounds that I've heard from 2GR exhausts, I dont think it's an insurmountable problem. 
If I had a broader vision for the car than the two factors above, I could forgive it for sounding a bit "meh". 
But this is the one thing that will drive me mad if it's not right. 
Currently sounds way worse than a prius motor did - Surely that's a fairly low bar to set! haha.

Still stoked to do some skids in it, but really looking forward to getting the tune more dialled in and seeing where the torque/power ends up with current setup.
Then cant wait to chuck the big cams in and see how it compares.

I will go back to listening to Alfa 155 DTM noises while my faith in V6 wavers...

[Roman]

I have started dialling in fuel a bit better with 2 or 3 full throttle runs so far.  No cam advance, just setting a baseline. 

The shape of the fuel map / power curve is absolutely hilarious when there is no cam advance.
So the peak VE (105%) is at around 5500, secondary peak 7000rpm 103% so pretty good for the dunga stock cams and no cam timing adjustments yet.
Gave it a run to 8300rpm. Good news is that as expected the bank to bank difference narrows down as rpm/throttle opening goes up.

But.

At 2000rpm full throttle, its only around 55% VE max. It can only half fill the cylinder, so it must be bouncing a big bunch of the air back out.
Still sucks until around 4500rpm then starts to wake up.

It's also quite gross coming just off idle, as I think it's gaining a huge amount of throttle area quite quickly. 
Possibly also because I've got a secondary air source that bypasses the throttles (PCV line and brake booster) so it does weird stuff where more air comes into the throttles, but then the reduced vacuum means less PCV. or something. Dunno. 
It's getting better but it's been the hardest motor to tune so far. 
I think mainly because the injector position sucks.
Will definitely look at doing some explorative drilling on my old heads to see if I can mount the injectors further down.

It will be interesting to see how the exhaust cam timing will like to be adjusted as well, as I've never played around with that before.

This graph isnt suggestive of final powerband at all. I am just posting it because it's comical to see how much this engine sucks shit with fixed cam timing.

I suspect that the home position of the cam is hugely far back, so it's working kinda like the atkinson cycle on the 1NZFXE.
So its trying to reduce pumping losses as much as possible by having the cam intake timing incredibly delayed.
People say that the Mark X gets exceptionally good fuel economy for a big barge, so this is probably one of its tricks. 
It's also probably why it cranks over fairly easily with 12:1 comp as well.
I bet it would read surprisingly low on a compression test.

The low end should jump up by heaps with even say 10-15 degrees of intake cam timing added.

[Roman]

Exhaust noise stuff 

Alright alright alright.

So, for this motor I've always envisioned it having a more high pitched racey sorta sound.
Kinda like a motorbike with more cyls. 
What I've got at the moment is like a gluggy V8 sort of sound coming out, even when it's revving out nicely.
So in order to fix this I think I need to somehow quantify what "good noise" and "bad noise" actually is. 

So any engine has a fundamental exhaust frequency that relates to the number of pulses coming out of the exhaust.
So for example. A 4 cyl engine doing 8000rpm.
What is the frequency of the exhaust pulses coming out of the engine. 

8000rpm x 4 cyls 
divided by 2 because each cyl only fires once per 720 deg. 
Then divide by 60 to convert from rpm to hz. 

8000 x 4 /2 /60 = 266hz

Then another example, a V12 engine doing 6000rpm.

6000 x 12 / 2 / 60 = 600hz 

So obviously a V12 is going to sound higher pitched, as it's got more events going on. 
So that's one thing that you cant change about the motor. 
With some excel nerding we can figure out what sort of freq a 4, 6, 8, 10, 12 cyl motor naturally makes at a variety of rpm:





There's a youtube plugin called "Youtube Musical Spectrum" which is meant for guitar nerds to reverse engineer stuff I think. 
But it's also helpful for visualizing exhaust noise.

If we start with some known examples of good sounding stuff and look at the frequency plot which now overlays at the bottom of the screen.

This 6 cyl BMW sounds quite nice:
Can see that there's a big spike around 350hz which suggests it's doing around 7000rpm at this point. 




This BMW V12 sounds good. Can see there's a big spike around 375ish hz, which suggests it's likely doing under 4000rpm at the point this video is freeze framed.



This 4AGE starlet is running just under 10k rpm, sounds nice, and sitting at around 320hz which matches the tacho.




What you will notice about the cars above, is that the bulk of the noise matches the engine's rpm frequency.
There's almost nothing going on below 200hz but I think most importantly. All of the extra frequencies are higher than the engine's natural tone, not lower.

Now, lets look at some cars that sound..... well..... not so good. 
This is a BMW straight six motor that does have a 6-1 exhaust manifold, but doesnt sound as cool as you'd expect.
300hz so its doing around 6000rpm which matches the tacho.
Difference to the above plots?
There's a messy distribution of frequencies, and a lot of them are happening below the RPM frequency.



This is the Zonda that was ruined by the crappy double exit aftermarket exhaust: 
650hz so its likely doing around 6500rpm. But look at all of the frequencies happening below that point.




This is my carina which is sounding like a puddly arse at the moment:
(Fuck you, Chicken)
400hz = 8000rpm, but then lots of frequency peaks below that point and nothing above it.



This is the awful sounding exhaust of a V16 engine that was made from 4x 4cyl motors:
You can see its absolutely howling with a 1khz rpm frequency, but its got lots of yuck stuff happening under that.




Now, lets go back to an earlier carina video where it sounded better. I just had a coby hotdog muffler stuffed into the bum of open headers:
400hz, so the motor is doing ~8000rpm. 
But there are hardly any frequencies happening lower than the engine rpm.




So this is all interesting, but what can we do with this information?
Well, I think the deeper frequencies can only exist in a few different scenarios:

-Two adjacent pressure waves collapse into a big single one, halving the frequency. (like subaru boxer unequal header noise, or v8 firing order noise)
Possible causes for this happening in the carina could be an unequal length 2-1 section, or unequal length 3-1 on each bank. But I dont think this is the issue.

-There is some length or diameter of the exhaust that can cause a standing wave at lower hz than the engine pulse frequency. So maybe a muffler + pipe length is working as a helmholtz resonator or something like that.

I think the key to the nice high pitched noise that you hear from the likes of the Zonda, LFA, etc. 
Comes from realizing that a large diameter pipe allows frequencies lower than engine rpm to resonate. 
So to get the flow they need, instead of having one big honking pipe at the exit. 
They have multiple smaller pipes. 
This seems to be the case with almost all of the other cars that sound nice too. 


Exhaust outlets of high rpm cars that sound good:
Zonda with its signature 4 exit pipe.



The LFA has 3x pipes, again not  huge:




The V12 Mercedes has only 2 tips, but seems fairly small:



This V10 Audi which sounds absolutely crazy has 4x quite small exhaust tips:

 



Based on the above my current theory is that a smaller diameter pipe must act like a low pass filter and eliminates any frequencies lower than engine rpm.

My current muffler has 2x 3" exits, and an empty chamber section in the rear part of the exhaust just before that. 
I'm guessing this is what is responsible for making it resonate at those lower frequencies in a gross way. 
If I covered up one of the exits, the tone improves straight away.

So I am going to cut the back end off my fancy muffler. Then I will make it so I can have a flanged section where I can bolt on different outlet configs. 
I think having 3 or 4 small outlet pipes (so equivilent area or slightly more, of a single 3" pipe) will end up a lot more high pitched how I want.
This way I can also test if it needs some long pipe lengths after the muffler to make it work.
Fingers crossed I manage to figure something out without 3 zillion iterations.

A single 3" pipe has a cross sectional area of 7 square inches
So to have the same cross sectional area with two pipes, they would need to be 2" each. I've got some 2" pipe here already.
To get the same area with 3 pipes, would need to be 0.87" each. I dont have any pipe that small, but I do have some 1.5" pipe that I made the extractors with. will try that.
Will be interesting to see how the pitch changes, if at all. 
Hopefully have some setups to test this coming week.

[Roman]

I havent had a chance to do much exhaust testing, apart from trying some gatling gun looking ends.

So this still has same cross sectional area as a single 3" pipe but it has longer pipe bits which protrude in, and has bellmouths on the entrances so it's hopefully not too restrictive.
Surprisingly, it does in fact sound a bit better even at idle.
God bless PLA and duct tape.
Changing the tips is not getting the sound I need/want (Didnt think it would be that simple, but best to rule out the easiest possibilities first)
I cant show some nerdy stuff about wavelengths etc because it was incredibly windy at the time.



Some ethrottle junk turned up in the post.
I gave it a test and it works great!
A single motor could pull open both banks, no problem.
So having 1 motor per bank and it should be super snappy.



I've got an ethrottle pedal from a Toyota Aqua wired in, and have the ECU running one bank through the onboard ethrottle. 
For the second bank my dashboard now doubles as an ethrottle controller, and runs an H bridge to power the 2nd motor. 
So a high priority CAN message gets sent at 200hz containing ethrottle target (throttle %) and actual throttle %.
Then it applies a PID routine to output a frequency to the H bridge to give it some juice.
It's all wired up and just needs PID settings locked in with some testing once it's operational.

Since I currently need to run ethrottle from the same microcontroller that does dash stuff. 
I need to make sure I'm not hogging the processor for half a second (or whatever) with screen updates or something. 
Currently my code has just been running in a loop, updating the screen every time it can after receiving some CAN data.
I never checked how quick this is, but its zillions fast. 
I've now cut it down to a fixed refresh rate of 50hz for screen refresh, 200hz for ethrottle PID routine. Then the rest of the time it just sits there collecting CAN frames. 
It's managing to catch 20+ lots of can frames in between each PID routine so there shouldnt be any issue getting the relevant ethrottle data from the ECU in time.
I've added some other bits and pieces to the code to make it easier to draw features, change color scheme and so on. 
It's still not quite the aesthetic I want but it will do for now.
I've ordered a Teensy 4.1 (rather than my 4.0) as this has an onboard SD car slot. Meaning I can load gifs and bitmaps onto the SD card so it can display them on the screen. 
Once that turns up I'll have the 4.0 as a dedicated ethrottle unit, and 4.1 just for the dash. 

I've drawn up some new pulleys that can sit underneath the fuel rails without bonking into anything.
New one looks like this. No holes or anything, as I dont want to have any chance of snagging on something or jamming while it's further down in the guts of the intake setup.


So all of this stuff can go up on the wall of shame, and make way for an airbox.





Alsoooooo some Kelford cams turned up! 
Exciting.
Big cams are the make or break for this project, so it's exciting to finally have them.
Am I completely full of shit about this motor revving out good? Lets see. Haha.
Lots to do at the moment but very busy.
The plan from here is to finish the ethrottle setup. Then get a decent full throttle fuel map / VVTI map sorted on the factory cams.
So I've got a good baseline for comparison.
Its incredibly annoying how much of the motor you need to take apart to swap cams.
If it was easier, I'd probably just do it straight away. Haha.

[Roman]

I havent really driven the car anymore, because I've had a clusterfuck of small inter-related issues which are causing things to not be fun. 

So firstly the bank to bank imbalance was a pain - So I went back to setting up double e-throttle.
After a bunch of hours, and admittedly was very close to having it running good.
I decided it was quite un-fun to try make my own ethrottle controller, so I am using a borrowed G4+ Xtreme just as a 2nd throttle controller. Ha.
It's not doing anything except for actuating the throttle, and receiving throttle position target over canbus.
So once this was all going. I tried adjusting each bank individually to try get lambda to match bank to bank. 
Still sucks, even when the motor is at steady state.
Still got some other issues somehow.

So I pulled the fuel rails off and printed a test rig to double check injector data. 
Yep it was all good, no issues. 

Sooooo what else could it be? Maybe individual throttle balancing. 

Ages ago I tried to print a little throttle balancer thingy, but my printer was acting up.  



But went back up the priority list.
It's a little tube that I can plonk on top of each throttle, and it'll give me an airflow reading through a MAF that I've stuffed in there.

Like so:



This worked GREAT! 

My previous guesses with the screwdriver were massively wrong. 
On one bank, all 3 throttles were being fed air by only 1, back through the balance tubes. The other 2 throttles werent flowing any air at all through the throttle plate. 
It only took about 10 minutes to get everything balanced up, and now it's all matching a lot better, and idle is now really nice.

I've just printed a version 2.0 that has a smaller internal pipe size so it pushes the voltage readings a bit further up the scale, so its a bit more sensitive. It works well enough as is, but not making much use of the range available.
I will also add a capacitor to the output so it smooths the signal a bit.

My ethrottle pedal that was from the Echo, has a weird pivot point height and pedal height when fitted to Carina.
So it's quite awkward. But it's working at least.
However I think I will try adapt the sensor onto the normal pedal instead. 

Hopefully now it'll start feeling like the car responds and acts a bit better when trying to dial in the fuel.

 

Edit: some paddock tuning

 

 

[Roman]

Airbox activities.

I wasnt sure how to start on this or what might fit with bonnet yet. So just start with making a base plate that clears the rails etc.

Then the goal is to have a filter, or at least the air inlet. Running over radiator.

Dipping into my bucket-o-filters i found this one which fills the gap pretty well.

A few more top part tests then printed one with a flange on the front. So i can test the shape and start dipping it down.

Mainly just need to work around the radiator cap area.

It looks like i will have a pinch point to the bonnet about where the flange is. So decided to start making it wider as soon as the fuel rails finish.

This is current scheme thats just started on the printer.

Im not sure what the lid will look like yet, but will try use as much head height as possible.

Im not 100% sure how I will make the final version yet. Possibly thin nylon print with carbon smooshed over the top.

 

[Roman]

Found an easier option I think.

Rather than running the super long wide filter, and having it directly over radiator. I guess it's easier to put it a little close to the motor, and then any flexi join etc that's needed, isnt critical to sealing filtered air in.

I found this other filter in my bucket, which looks like a good size and means everything fits within my printer's width:

 

So now I'm printing the front part to hold it on and dip it down a little. 
It looks like I've got more room around the radiator cap than I first thought, so will un-scoop some of that area. 

Might add some ridges that run front to back to add rigidity to the whole thing. But for now, make sure it all fits. 
Then figure out how much of a roof I can put on it, and how to hold it all together.
This has been fun to make some fairly quick progress on for once.

[Roman]

"This has been fun to make some fairly quick progress on for once."
I just had to say it, didnt I? 
Okay as per chat thread, there's a bit of a change of plans to make everything work betterer. 
I dont have nearly as much bonnet clearance as I would have hoped.
I've decided that remaking the fuel rails will help remedy this, but also solve a bunch of other problems at the same time.
The existing fuel rails sit high, meaning the airbox has to be quite narrow. So it shrouds the trumpets a bit:


So there are a few considerations for a version 2 version 3 fuel rail setup.
-Last ones looked kinda cool when I modelled them, and I was excited to print some gratuitously bendy looking shapes. But ended up weird looking. I will keep it simpler now.
-This time I will be more considerate for the person that needs to machine the fuel rail later. This means having a flat base that is perpendicular to the direction of drilling needed, so it's easy to clamp down.
-The fuel feed fittings will now be on the same plane as the injector holes. Not located in the long end which makes it very difficult to secure to drill/tap.
-I need to keep the fuel rail high enough that I have clearance for my cable pulleys to be able to move at least 90 degrees.
-I need to keep the top of the fuel rail flush with the underside of a wider flatter airbox.
-Make sure that fuel dampers can fit this time. I suspect this might solve some weird behavior with fuel at particular rpm/load combos.
-I will add some fixing points so the airbox can mount on the rails. So all of the intake stuff locks together a bit and makes the whole assembly a lot more rigid.

After some farting around juggling the above criteria back and forth. has ended up like this.

New fuel rail shape is a bit simpler. It has some fixing points that I will fit studs to, for holding the airbox on.




 

 

 

 

The throttle pulleys are a tight fit underneath, but will be able to rotate more than 90 degrees.








It is kind of annoying to be redoing things, again. However this solves a whole bunch of issues so I'm happy to do it.
Thankfully none of this impedes other progress. As there's lots I can carry on with while still having no airbox and using the current fuel rails.
 

[Roman]

I borrowed some proper corner scales.
I was kinda thinking lots of little extra bits would add up.
Even though the V6 is technically about sameish weight as a 3SGE. 

I was super stoked to see these numbers! 
This is with no passenger seat, otherwise full interior.
Not sure on gas tank level.
Heavy 15" wheels with semi slicks on. 
Full glass. Lithium battery.

There are still some easy wins in the weight department yet. But super happy for it to be this already.




For some context, this weighs about same as a gen 1 MX5.

[Roman]

Another motor?  Be rude not to

A friend had a customer's car come in with a seized 2GRFSE engine.
It was deemed uneconomical to try and repair, so they replaced it.
Being the local V6 weirdo, I was asked if I would like the seized engine for V6 activities. 
Why, yes I would! Thanks.



I've mostly stripped it down, and found nothing particularly wrong so far. 
It turns over smoothly except for a small portion of rotation where it "sticks" for some reason. I found a broken chain tensioner, but this looks to be a symptom rather than cause.
With the chain fully off, it still does the same thing.
With a camera stuck down the spark plug holes, all of the bores are looking good as best I can tell. 
I am guessing either it's a main bearing that is stuffed, or maybe a piston has cracked a ringland, so it jams when it first starts going up (or down)
But it's looking though its still going to be a useable engine with a minor fix hopefully.

So, what to do with this motor? 
Well, it looks like a good candidate for destroking with the 5GR or 2MZ crank. 
Hopefully the heads and the bores are 100% okay on this motor. 
As the crank and rods get binned anyway.
Once I've confirmed its all okay, I'll just put it mostly back together and park it in the corner of the garage for now.
The only "downside" of this motor is that the ports on intake and exhaust are incompatible with my existing manifolds.
However it'll be no stress to make some new ones later on to test.

Throttle Pulley problems

It seems I messed up my measurements for cable length for the dual ethrottle setup.
So it's now working amazingly, and both banks are getting to target and really quickly. However I can only reach about 70% throttle before it runs out of cable pull.  
So my pulleys are off with Dad to cut them down a bit with the CNC machine, so I can reach full travel. 
Ahhhh well shit happens.

Airbox stuff

Height of airbox to the bonnet is still the main issue I'm contending with. Rhys had a suggestion to try trumpets that are parallel to the roof of the airbox, to gain a bit more clearance. Good idea!
So just printing some of these out.






Thermal camera side quest

I last had a thermal camera about 10 years ago. Technology has come leaps and bounds since then. 
My last one was very low resolution, and limited to either 5hz or 10hz refresh rate.
It could only read up to 200ish degrees, which wasnt hugely useful for things like brakes or exhaust stuff.

My new one is a little device the clips into the bottom of my phone. 
It can record video at 25hz, has 256*192 resolution, and scales up to 600 degrees C. 
Considerably cheaper than last time too. Bloody awesome! 
The thing about it that will be absolutely magnificent though.
Is that I've confirmed it works with a decent quality USB-C extension cable.
So this means I can mount the tiny device somewhere in my engine bay, wheel area, brake area, or whatever else. 
Then record and view results from in the cabin. Gonna be awesome.
But obviously I need to have all of the bread and butter parts of the car sorted before this will be much use.

Although 256*192 resolution seems low for a modern type of camera. 
It's awesome when you think about what it's replacing. 
If you put a temp sensor into something, you only have 2 datapoints. temperature and time. 
This has over 50,000 datapoints instead of 1.
So instead of "radiator is hot at this particular spot" you can now see the temperature gradient across the whole thing. 

Interesting to see the temp gradient on this SW20 radiator, which has both its inlet and outlet mounted at the top of the radiator.



It's cool to be able to see the exhaust pipes without maxxing out the sensor instantly. 
The bright areas are because of the different emissivity of the material, not because my welds are bad. That's my story, and I'm sticking to it.