Updated: Oct 25
This will be the first post in series that will cover development and production of an all-aluminum cylinder block to replace DOHC VG30 blocks in Z32 cars. This is going to be the most challenging project that I've attempted to execute to date. I have to give credit where it is due and say that the direction of this project is heavily inspired by work that Electramotive Engineering did back in the 90s. I have studied their work pretty heavily over the past couple of years and have always wanted to produce a cast aluminum cylinder block. It is important to note that my budget for this is nowhere near that of an actual race team that is sponsored by a car manufacturer. About a decade ago, I would consider this nothing but a dream. However, in the past decade, a lot of new manufacturing technologies have been developed and we are now also outfitted with modern CAD software that is here to help. At this point, I no longer consider this a dream, but more of a goal for first half of 2024.
I have been around Z32 cars for almost 20 years now and have seen lots of various builds. While there are countless cars in this community that make very respectable power figures, some of these builds come up short and end in a not-so-great way. Engine failures become more common as we approach the 1000WHP mark. This is not to say that DOHC VG30 is an incapable engine. If you think that more popular engines like JZ and RB are any better, you'd be wrong. They too require significant modifications to reach 4-digit power levels. The only difference is that JZ and RB (as an example) had endless hours of research and development invested in them by race team engineers to get them to this level. VG30, unfortunately, has been overlooked. To an extent, I am actually happy that bringing this engine to its glory is still up for grabs. I have studied countless posts and pictures of failed VG30 engines and became familiar with a lot of literature and publications discussing the matter. Earlier this year I decided that it was time to act.
As with all my other projects, it is important to understand why the project is even necessary. Given the fact that there are now billet blocks available left and right, why even bother with casting? Well, billet blocks are strong, but they have drawbacks that a customer with a street driven car would not really be able to live with. Billet 6061 and 7075 aluminum alloys have thermal expansion rates much higher than those of steel engine internals and cast aluminum heads. That makes setting bearing clearances much harder for a street car and getting your head gasket to seal is not an easy feat either. Billet is great for dedicated drag racing applications where engine run time is short, but for a street driven car they are not an ideal choice. Billet blocks are not really friendly on your pocket either and a good block from a reputable manufacturer will cost in 5-digit dollar amounts.
With a cast aluminum block, we are: -Able to fix all inherent stress areas in DOHC VG30 block
-Keep stock like operation and drivability
-Keep all parts in the moving assembly happy with similar thermal expansion rates
-Significantly reduce weight
-By modifying the bottom section of the block, convert it to receive 6-bolt main caps
-Install cylinder sleeves that would allow us to over and under bore the cylinder
-Increase cooling system efficiency due to inherent thermal properties of aluminum
-Use premium grade cast aluminum alloys that are stronger and more elastic than factory iron
-And the list goes on...
Building an engine block is not an easy task by any means. It is not surprising that this is not attempted as often as other automotive projects. There are numerous steps that need to be taken before work even begins. For starters, a very accurate 3D scan of a healthy cylinder block needs to be made and processed. If you think this is doable with a scanner cheaper than $15-20K, don't even bother starting. This particular scan was done with a professional grade scanner specifically made for reverse engineering projects. As a result, we were able to make a very clean and accurate scan and to be honest I expected nothing less of a ~$38K scanner. You can even see that it picked up slightly uneven desk surface (yellow around cylinder 1), which I was able to confirm with tools.
Scanning external surfaces is a good start, but it does not give a full picture. Actually, it does not even give half the picture because there is no way to have a usable cylinder block without a water jacket (for street application). The next step was to reverse engineer a factory water jacket. This was done by doing some x-ray work and well as physically cutting up, 3D scanning the cavities and then overlapping all data in CAD software. This is all, of course, extremely oversimplified, but what matters is that it has been done. External scan, x-ray data and scanned exposed water jacket cavities were all combined, aligned and after a few hours of processing yielded a combined point cloud.
Point cloud is nice and all, but it is not usable in its raw form. A solid model needs to be built from millions of points in space. We need to build all surfaces which eventually make up the structure of our cylinder block. We had some critical dimensions of this block available to us, so applying that data to scanner cloud we were able to build and confirm the rest of the model. Again, this is all oversimplified, but what matters is that it is done and after a few weeks to work, a solid model of DOHC VG30 block was built.
This was a picture of a very important milestone in solid model development. At this point all critical features of the cylinder block were built and we started analyzing the rest of scanned data to see which features we want to keep and what we need to clean up. Brown model is scanned data and yellow is solid model that is being built. This block was used in cars with different drivetrain configurations and since we only care about using this in Z32 platform, all unused features and miscellaneous things were eliminated.
Fast forward through immense amount of suffering in front of computer a finished, cleaned up solid model was made. The pictures do not do justice to the amount of human hours that were spent of this model, but doing anything less, or by taking shortcuts with anything would be completely useless for future development work.
As far as I know, this is the most detailed DOHC VG30 model that was developed and this is nothing short of what is needed to bring an aluminum engine block into reality. You can see in the image abode that all critical features of the block were modeled including the mains, exact replica of block's crank case, all accessory mounting points and of course, the water jacket.
In the gallery above, you can see most factory features that were removed (blue). Those are unused accessory mounting points, unnecessarily tool access points for machining, dipstick mounts/holes and other block features that are not going to be used in our final model. One of the inherent issues with these blocks in their factory form is the amount of casting flash that is left in coolant galleys. You can note the amount of casting flash was removed from main coolant galley (blue). There will be some further port optimization taking place around the water pump outlet, but even at this point it is night and day difference.
Another interesting view is a cross section of the block that shows us how main coolant galley tapers down towards the back part of the block. As you can see, coolant ports also get progressively smaller. Here we see coolant ports for cylinders 2, 4 and 6. This is a cleaned-up model of the block, so coolant galley looks more or less exactly like engineers designed it, without a pound of casting flash.
I do not plan to do any significant changes to main coolant galley aside from further cleaning up casting flash and smoothing out excessively sharp corners/protrusions.
Simply switching over to aluminum, cooling system efficiency will get a huge boost because aluminum conducts heat much better than cast iron. This is a huge perk of this conversion since factory water pump and thermostat will be more than sufficient even at higher power levels. You will still need an adequate radiator with a capable fan, but I really hope not having to explain that to anyone who is considering to run this block.
You can see factory webbing around main on the right side of the screen and V1 webbing modification on the left. Material has been added into V1 webbing to create a sufficient anchoring location for an additional stud (top left) and a wider fillet connects the main to side rail. The goal here is to spread the load and avoid any stress points. I do have the luxury of lower density for aluminum, but everything will still be optimized to keep weight down to minimum.
The factory ribs (marked with question marks) will likely stay. Their fate will be determined as we run FEA as well as CFD. Reducing crank case windage is one of the goals of this project.
This is just me trying to visualize the direction that will be taken in this conversion. This is not final version and is absolutely useless without FEA. On top of that, side rails will be extended in further revisions.
I am currently waiting for FEA to be completed on factory block model based on original material. That should be coming back to me for evaluation in a few days. It will give a good benchmark and indicate most critical areas that need to be addressed.
Lots of interesting things can be visualized once we start dissecting the block. It is also much nicer when you can do that by clicking a mouse as opposed to cutting iron with a saw.
When the VG block was scanned, deck thickness was ranging from 9-10mm. On this model, it was built at uniform 10mm. Designing proper deck thickness is a balancing act since you want to make sure that you have water jacket close enough to the top of the cylinder to evacuate the heat, while still keeping deck true and square. Thick deck is strong, but will run hotter and more likely to knock. Thinner deck will distort, but is easier to cool.
Again, by switching to aluminum, we’ll be able to address this peasantry and kill two birds with one stone. I am (at this point in development) inclined to run the deck of 18-19mm in thickness. Even by almost doubling the amount of material, we will still be able to draw heat away as well, or even better than factory cast iron setup. There are still tests left to run, but a good amount of high-performance aftermarket blocks are in that general range and after preliminary calculations I can see why.
Also, a cool thing to see here is how much more material is present around main 1 and 4, while mains 2 and 3 are lacking a good attachment to cylinder bores. Coincidentally, most flexing happens in the middle of the block so those mains 2 and 3 were not the best places to skip out on metal. Yes, there had to be a compromise in those locations to fit machining tools into and bore bottoms of cylinders, but we will be using better tools and are not limited by those constraints.
To be continued...