Roman

A little progress

Unsolicited general Life update 

Things have been a bit frustrating at the moment.
It felt like I was right on the precipice of having the motor firing up, and on a roll after getting the sump fitted.
However seemingly simple things have been getting complicated and house related stuff has also been sapping my time/energy/life.

I dont think there's any chance of attending oldschool drags in this car now, just dont have enough time left. 
Current goal is just to try get my car moved to my new house. 
Just getting the car movable has been surprisingly time consuming.
A few years ago I was planning a fairly decent overhaul of a few things on this car. 
Swapping SW20 rear brakes for MX5 stuff, swapping booster setup for a pedal box, redoing fuel lines and front calipers etc. 
So I ended up pulling some stuff apart, figuring it'd be coming out for good.
So I didnt bother putting bolts back in the right place and all that. Which I'm usually fairly good at. 
So, I've chewed up a massive amount of time just looking for bits which is annoying. 

I managed to get the handbrake working, after lots of rummaging through boxes etc to find parts.
Still need to spanner check a lot of the suspension stuff.

New intake stuff

This has also been a lot more complicated and time consuming than expected.
I've definitely pushed my timeline back a whole lot by swapping throttles.
I ordered some fuel rail extrusions, 12mm rod, and some bearings.
I've finalized the drawings to drill new fuel rails and actuator rods and have done a few test prints before chopping anything up.
For the fuel setup, I was initially wanting to feed into one fuel rail, link to the other, then have an FPR on the back of the second rail.
However its too crowded up front to link the left and right banks together. 
So I'm going to use the Corvette fuel filter with the built in FPR, to make it a returnless setup.
That will split to feed each bank and have a pressure dampener on each rail.
I'll mount a fuel pressure sensor near the tee, to see if there are any weird things happening with having the FPR down the back.
However that's how a lot of cars do things now, so hopefully it's fine.

I've decided it'll be a good idea to have some support plates that link each bank's rods together and hold them straight etc.
As each bank tries to pull itself in towards the center of the vee when rotated against spring pressure. 
I'm just going to run a cable pull setup as it's seeming a lot more complicated to get an e-throttle setup working for now.
It might still end up a dual e-throttle setup, as this will actually make balancing the left and right banks a lot easier.
Do it digitally rather than mechanically.




I'm going to drill the fuel rails and rods on my dads manual milling machine, which has encoders on it to output X and Y coordinates. 
So it's easiest if all of my dimensions are relative to a zero point on the parts. Hopefully get some of this done tomorrow night when I'm up in Auckland.






With some printed mockup parts on the engine it's looking like this. 
My throttle pull wheel thing will go down the back of the motor instead, as it's already quite crowded up front. 
Annoyingly, for whatever reason the BMW TPS's stick out quite far off the end throttles. 
However I'll keep both of them, as it'll make it easy to diagnose if I've got any bank to bank imbalance for whatever reason.
It's been finnicky to get right. But once I've got the new rods drilled, it will be a lot easier to test etc.
As it all binds up a little currently as the linkages arent spaced apart properly.
Although it's all been a pain, the amount of room to the bonnet that I've gained, will make all of the extra trouble worth it.
The final version will have the pulley arrangement fully enclosed so there's no chance of it getting jammed up with stuff falling in somehow.
These linkage rods I've got are no good, they'll vibrate the ends loose easily by the looks.
So I might end up loctiting the threads to the exact right lengths.
Or find some better ones.




Digidash fuckery part deux 

After a lot of frustration troubleshooting this LED screen, which would only display some of the lines on the screen.
I found that my Teensy board was the actual issue. Replaced that, then it started working properly. 
It doesnt look so great on photo/video but the appearance in real life is ssweeeeeeeeet! 
And it's super bright too, and the boot time is almost instantaneous.

So I've started working on some basic graphics for it. 

Fuck knows why there's a radar looking thing there, but there ya go

What I have learned from previous efforts, is that any time spent adding some structure and scalability to things pays dividends really quickly. 

All of the stuff shown above is movable/scalable in real time. 
To generate one of the vertical bars on the screen, I just need to run this line of code and it generates the borders, level, colours etc. 

VerticalBar(6, 25, 25, 30,0,100, temp); 

Where the values are: 
VerticalBar(Xposition, Yposition, Width, Height,Min value,Max value, CAN Variable);

Then for the indicators, its just like:

IndicatorLeft(10, 7, 1);
Where the values are X position, Y position, and 1 or 0 for on/off. 

So if I have some context sensitive screens, where some of the features need to move around. It's incredibly easy to reposition them.

I could even move them around in real time, for whatever reason.

One thing that is great about how these LED screens work, compared to the Nextion LCD screens I have been using in the past.
Is that the LCD screen, there was an onboard (slow)FPGA and (slow)processor which handled graphics stuff. 
So it was cool that you could load fonts and bitmaps etc to the screen itself, and using a GUI to design things on your PC it would just load all of the graphics, fonts etc.
Then you just had to send it serial commands to update values or whatever.
However the problem was, any time you send a command (for example, update the temperature value) 
It would start that command, pause the unit while it refreshes/rewrites all of the relevant pixels. Then when finished it would send an Ack command back over serial to say it is happy to accept the next command. 
The problem is that even the smallest quickest command has quite a high overhead in write time.
So if you want to update 10 things at once on the screen, you end up waiting while the stack of commands sequentially run. 
So it ended up that whatever you could do to minimize the amount of pixels that updated, paid dividends. 
But it was annoying, and ultimately a bit too slow.
The refresh rate was too slow to have anything like a shift light or shift bar as part of the screen. 

With this LED screen instead.
The only onboard hardware is essentially a few buffers which hold the current RGB value of each pixel.
Whether you're updating anything or not, It rewrites all of the lines over and over from the buffer at a very high rate.

Instead of sending each separate command, like to update the temperature value. Then requiring a confirmation that each command has finished.
My controller board (Teensy 4.0) has a copy of the entire screens RGB values stored in RAM. So any update you make, first happens just to update these values in RAM.  
Meaning you can absolutely stuff it full of changes and there's no external communication, it's near instantly changed.   
Then when have changed everything you want, it sends the whole screen's new image over.
Whether you have made 1 pixel change, or all of them. It's exactly the same speed either way. 

So essentially, complexity is free. which is super awesome, and means I can have as many layers or stupid animations or star swipe or whatever the hell else. 
Any animations or generative graphics can be scaled and reapplied etc as much as I like and it'll have almost zero impact on performance.
If I use the Teensy 4.1 arduino board, which has an onboard SD card. I can load bitmaps or GIFs onto it to display on the screen as well. 

Basically, in conclusion, I've found a digidash option that is an animated, low res version of MSpaint.
Who'd have thought that in the year 2023 there are such glorious technologies available for those of us wanting to draw pointless low resolution graphs in real time. 
Living the dream!
 

Sump Stuff

I managed to get the sump finished off, in place and gooed up.
Since I am not exactly sure about how much oil it can hold, and it's probably a good idea to iteratively mark the dipstick as I add 1 litre at a time. 
I just poured 5 litres straight in, because, fuck it. 
This worked out reasonably well, the high tide mark is just below top cover of the pan. 
The oil pump and filter will hold a bit more oil when it's turned over, and the external oil tubes that run up to the cam covers for VVTI and oil spraying the cams. Actually have non return valves in them, so hold a bit of oil up there too. 
Then, some amount of oil stays in constant circulation while the motor runs, more so at high rpm. So hopefully that's around a good amount. 
It's not weed its oil all over the floor, it's still good. Exciting! 

Digidash Fuckery 

So my replacement LED screen turned up, aannnndddd... I just cant get it to work properly. 
It only displays the top 4 lines, then skips 4 lines, then next 4 lines, etc.
The way these work is actually really complex.
I think this particular screen is incompatible with the only available library for controlling these screens, which is a real pity. 
This doesnt photograph well, but this looks absolutely AMAZING in person. It would be absolutely incredibly awesome as a dash. 
So I think I might need to buy some of the smaller screens that are supported, and daisy chain them together. 
As they have different chips controlling the LEDs, or some crap like that. Dunno. But it's way over my head to try and troubleshoot what's going wrong here. 

However, In the meantime I can still figure out how it works, and draw some graphics on it. So long as I dont mind everything can only be 4 pixels high, haha.



Flywheel Stuff 

Dad got the ring gear machined to the right dimensions for shrinking onto the flywheel. 
However when skimming down the back of the flywheel, the tool ugga dugga-d itself into the part. or something.
Nothing catastrophic for the part or the machine, but bloody scary apparently! 
Luckily he's smarter than me, and this happened only a few mm from the top, we wasnt trying to cut the whole 10mm out at once.
He said once he found better tool speed and feed rate that suits the chromoly, it cut amazingly nicely.
So hopefully this isnt far off, anyway.
I'm excited that his mill is running. It's damn cool.



Intake Stuff 

I swapped intake setups with @Hurmeez so now I've got the V10 throttles. 
This solves some problems, but adds some new ones. 
Firstly, it solves all of my height and fuel rail issues which is a big win. I'll now have some space up top for a nicer shape trumpet, and a sweet looking airbox. Which will also help give some rigidity to each bank of throttles by tying them together at the top.



Unfortunately the factory throttle spacing is a little closer together than expected - But, these are a lot easier to space apart.
I was considering just leaving it at current spacing, so I can use the fuel rails etc - But nah. It would just nag at me that I'd half assed it.
So to space them out I just need to redrill some holes on the rails, to rebolt the linkages in a new position.
Otherwise, if it's close enough I might be able to slot or extend the linkages to just push them across a bit.
Then I'll need some new fuel rail extrusion, and drill them. But this is all easy drilling stuff that can be done no prob on a manual mill.

The center to center on these throttles is 98mm, not the 107mm I was expecting. (Could old forum information be wrong?! shock horror etc) 
Also they have a highly oval outlet, and the throttles have a 24 degree curve. 
Despite have a big entry size and quite big throttle, somewhere around 50mm. They taper down to the cross sectional area of a 42mm circle at the exit. 
However there's enough meat around the perimeter that they could easily be ported out to the area equivilent of at least a 48mm circle no problem at all. 

 

In order to get the top of the throttles standing vertically, I need a manifold that cancels out the 24 degree angle. 
Here's where I am currently at for prototype shape: 



Since can align each throttle central over each port, all of the port shapes are now identical which is nice. 
The engine they are from (S65 V10) makes 500hp from 5 litres, so around 50hp per throttle. 
So that's equivalent to 300hp using 6 throttles. So hopefully injector size, airflow etc all plenty suitable for the application.

I'll also have 4x throttles left over, so it might be interesting to see how these look against the port shape of 1NZ or something.

Engine Weight 

 @flyingbrick loaned me his scales (thanks) so I could get a ballpark weight for motor.
I've not gone quite to the level of @Hyperblade's awesome post about engine weights at this stage - as I still dont have some parts to put on. 
However, with the motor and box put together. Engine mounts on. No starter/flywheel/manifolds. Like so:



Total weight came to: 



184 kg.
A J160 is 38kg, putting the 4GR around 146kg. A little lighter than expected, given the service weight was listed as 165kg for engine. 
So this puts it at slightly heavier than a beams motor once it's fully dressed. Much heavier than a K20, and the weight of almost 2 1NZs. Ha!
So I'm not gonna win any awards for worlds lightest engine, unfortunately. (Did I ever mention how great a 1NZ is? Cant remember)
However I'm also relieved that it's not going to be massively different to the beams setup.
The fact that the weight is further back and lower down will hopefully help offset the extra podge.

For some less discouraging context - comparing to a previous generation Toyota NA 2500cc six cyl engine, the 1JZGE.
This was 207kg for the engine alone. So my motor and gearbox weighs 25kg less than just an NA 1JZ. ha!
Before @xsspeed mentions RB in my thread again, an RB20DET is supposedly around 245kg which I assume is with turbos etc mounted. 

So lots of different things going on, the most exciting thing on the to-do will be getting the flywheel in place. 
As once that's done, I can crank the motor over and get some oil circulating through everything.
I need to assess if my tiny lithium battery is going to have enough grunt to turn the motor over, before I waste a bunch of time building my wiring etc around a battery that might not be suitable. 
Hopefully it's good though, as it cranked the 1NZ way faster than the regular battery did.

Oil pan activities!

Absolute GCs @Stu and his workmate Sam found some time in their incredibly busy schedule. To spend a night programming some cutting paths and cutting a big slab of alloy down into an oil pan. 
I went along to watch and learn some things and it was awesome. 

Firstly, start out with a 22kg block of 32mm thick alloy plate. 

 

Then align this onto the bed of the router, and the it gets sorta vacuum sealed against the bed. 
Then one side was machined, making an absolutely hectic amount of mess.

Then it was flipped over, and a a template was cut to align it onto the bed again. 



Making zillions of glitter carving out the other side.



Then once it was all cut, we are down to around 1.8kg from 22. 
So there were plenty of chips for me to sweep up, haha.
we needed to look at drilling the oil galleries for the filter, that I mentioned a few posts ago.
Problem - the depth needed was too long for any available tooling and the height of the machines etc. 
So we decided to mill two slots in, and then weld a cover plate over the top. We made the slots do a bit of a zig zag, so that they would avoid one of the threaded holes for the oil filter housing. Which previously penetrated through into the oil gallery as there was no other way.

Then Stu on the tig to zip it all back up:



Then ready for a test fit. How good! 
(It still needs the welds flattened off and some threads tapped, havent done this yet)





Much to my relief, it fits up great. 





Just zip tieing the oil pan on for now, until I can tap the holes. 



and back in the car... everything good! 




So that's one of the last remaining big tasks mostly sorted. 
And a big relief that it looks like everything is going to work as expected.
Hugely grateful for the time spent by Stu and Sam to help get this done. 
I think its the most complex and coolest car related thing I've ever drawn up.

Next steps are to thread a few holes for the oil pan and oil filter, flatten off the welds, and make an oil pickup.
then I can fit it on for good, and fill the motor up with oil. Getting closer! Exciting.
 

A few bits of progress.

Dual e-throttle
This has been harder than expected because PID settings that work great on the ECU, are totally different to the PID settings that work on my controller.
If anyone wants an easy to digest explanation of PID control, then this video is absolute magic. I end up rewatching it every time I forget what each of the variables do.

I've found out the reason for the settings being way different, is that the I and D in P.I.D. are time dependent variables.
So if you are refreshing the controller at 100hz, needs totally different values to 200hz (or whatever else)
Since the H bridge hardware is different to what's in the ECU as well, It might react at a different speed. So needs different numbers too.
I managed to get it tracking fairly close to the ECU controlled throttle, but still had some issues with things like slamming the throttle really hard into the end stop, when it's trying to shut it quickly. 
I think I need some more complex code to stop integrator windup and maybe some different PID values for when it's opening, vs closing. As the return spring obviously makes it react differently one way to the other. 
These throttles act quite "sticky" compared to others I've used, like it needs heaps of current to start getting it moving. Then once it's moving, you need to rapidly reduce it towards the target.
So its been harder to get the numbers right, even with the ECU controlled bank.
I think I'll just resign to the fact that having both throttles run from matching external hardware will be the way to go here.

Sidetrack / scope creeping the intake setup 
I've also still got the nagging issues of that these throttles are closer together than I'd prefer. But I also need to try mount them close to the ports so I've got bonnet clearance - which means my manifold shape is okay but a bit yuck.
So I've still been half way considering other options while I've been plugging away at other stuff.
With some interesting timing I've been chatting with @Hurmeez who has an absolutely kick ass Escort Wagon project, with a Mazda KL V6 engine. 
Which unfortunately recently was in some flooding but hopefully keeps making good progress soon.
He's been using some of the BMW throttles from the V10. However for a few reasons he has ended up needing to run the throttles quite far from the head, in which I think is a really awesome looking setup:



However having the throttles mounted in that position presents him with some maintenance difficulties for things like changing spark plugs. 
He's had similar nagging thoughts in the back of his head, about changing his setup.
So he's been looking at options for a more compact throttle setup that can live in the vee.
As it turns out, my Triumph throttles are insanely close to the bore spacing of the Mazda V6.
The Mazda engine has 97.5mm bore spacing, and the Triumph throttles have a 94mm spacing. 
Which is pretty damn close to perfect really!

I'm not sure if this is the right gasket for his engine, but I found one online and scaled it for a quick sketch of how a manifold might look. 
Even with quite a short manifold, it looks really good as the throttle distance/size is very similar.





So I'll go take my throttles up to Auckland some time so he can have a test fit.
Probably easy enough to print a few PLA manifolds for a test fit as well.
Then in return, I'll see if I can have a borrow of the BMW throttles too.
If it looks like it'll work out well for both of us, we might end up doing a swap.
The BMW throttles can more easily be spaced apart, which suits me better.
From factory, the BMW throttles use linkages to connect to a single e-throttle motor that runs both banks. So this might get rid of another one of my headaches if I can go down that path instead.
So, on one hand I'm not overly keen to scope creep the intake setup when the current setup "works".
However eliminating the need for 2 ethrottle motors is appealing.
See how it goes!

Bellhousing swap
So far I've just had empty casings bolted to the back of the bellhousing and motor, so its lighter and easier to test fit everything. 
However, thought I'd better finally put this on the proper box. 
Well, this was a friggen headache! 

The bellhousing should fairly smoothly just slide on over the bearings and front shaft things with a few taps.
However what was happening is that it would barely engage on the bearings, then become super tight. I could put in some bolts to wind the front casing on, but then its locked in tight and nothing will rotate - and there's not enough room to get the circlips on the front bearings. 
I thought maybe the casing had pulled a little, bit from the heat of welding. But it looks like the circlip groove on the bearing had caught on the lip of the housing, and galled the bearing housing, and dragged some shit down into the hole then made stuff worse. 
So I dremeled a fillet onto the bearing holes, and knocked off the high spots. Then after a lot and I mean a LOT of trial and error I managed to get it all together succesfully. 
When you pull the front casing off a J160 box, there is a little lever arm that is only held in place by a pin that screws in after the bellhousing is on. 
So you need to fit a rubberband over it to hold it in place, and hope it is nicely aligned and doesnt pop off when you put the pin through. 
Which sure enough, it popped off a bunch of times before I managed to make it work.
Not a job I hope to repeat any time soon! 
I also cut a hole for the clutch fork, and shoved a spare clutch/flywheel assembly in there to see if there's enough travel etc. 
Everything's looking good, hopefully I'll have the modified flywheel back soonish so I can whack it all together and onto the engine, ready to go!




Digidash 
The screen and controller board thing for my LED screen turned up. 
And it doesnt necessarily show well in pictures but the LED screen looks absolutely super cool! 



It took a little while longer for the control board to turn up. Once it did, I was looking forward to loading up the demo software and assessing whether this is a viable option for the viewing distance.
I remember reading in the instruction manual a few weeks ago, that these can draw a fairly crazy amount of current. They have some grunty big power wires, then regular small wires for the control board etc.
So I hooked up a 2amp 12 volt power supply for starters, not sure if this was enough.
A few lines on the screen were flashing, but it wasnt responding to inputs. But it was crazy bright, like burn my brain out bright.
I figured it was drawing more than 2 amps, so I connected it straight to a 12v battery annnnnddddd blew it to smithereens  

So back to RTFM and yeah it turns out that it does draw lots of current... But with a 5 volt power supply, not 12 volt. Haha. Oh man.
So I've paid the stupid tax and ordered another one, but will be a few weeks away until I can see how well it will work.


Finalizing engine height 
I printed a section of the new sump design, so I can finalize and minimize my engine height. 
I thought I'd print this at double speed, because I am impatient. This turned out great:



So after throwing that in the bin, I had to end up printing another one anyway, at regular speed. Which turned out properly. Ha.
There's probably a life lesson in there somewhere.



With some plasticine measurements, eyeballing, and iterative cuts lower. I managed to get the engine around another 7mm further down. with what I'm hoping is acceptable clearance to the crossmember. 
I was considering cutting a dish section into the crossmember, but there's no point - As the steering rack is currently about level with the crossmember's max height anyway. 
There's no way I'd want to lower the steering rack, or space down the crossmember as both of these options are nasty for steering geometry.
I would also have another issue, that the bottom of the bellhousing is getting perilously close to the road. Even at the ride height the car is currently at, which is hardly mega slammed.
But, it's about ready to have the crossmember mounts welded on which is cool.


Oil pickup tube 
Since this motor usually has the oil pickup right down the back, but oil pump up the front. 
It's got a very long oil pickup tube, like so:



So I need to modify or replace this, and I've not done anything yet.
Thankfully since I have a cad drawing of both the new sump part, and the altezza pan relative to each other.
It's easy to work out the new position without lots of fiddling around to take the pan off and on etc.



So the new pickup is going to be about 70mm across, but annoyingly it will cross over the bolt hole. So I think it'll need to run it's original direction for a bit first, then have a bend to reach it's target. 
I might sit it a little further back than center of the hole, because I'm guessing acceleration will be the worst case scenario for oiling when I've only got a very shallow pan under the windage tray.
So I've at least got some semblance of a plan before I go cutting everything up.
I might print a little jig to show the position it needs to end up in relative to the pickup bolt holes. 
Might print it at regular speed this time, who knows.

Firewall fixup 
Since I'm reasonably close to having the motor all sorted, it makes sense to get the engine bay ready for it to go back in, so it can (hopefully) stay in.
So I need to make some progress on properly cutting out and then patching up the firewall.
So I'm looking for a section of panel steel, around 1000x600mm or so. 
Does anyone have any for sale?
As I'd rather not have to end up buying a 6m length (or whatever) from Steel & Tube, when I'll never end up using the rest of it.

So, some slow progress and dumb mistakes but I'm still heading in the right direction overall.

The current state of oldschool.co.nz Toyota meme engines.

In today's episode of "why the fuck are things this hard" 
Adapting a TPS to get rid of the stupid magnet based one.

Thanks once more to @flyingbrick who has printed me a test version of my TPS adapter on his resin printer. 
Initally I just needed to figure out where things would end up, to see how to tweak it.



So I ended up skinnying it up a lot and adjusting a few bits to make it align and rotate nicely.
The (hopefully) final version features a gratuitous amount of fillets, because, thats what you do. 
Hopefully once this is finished I can finally do some testing on the H bridge setup for the second ethrottle bank.