DIY Engine Build - well mostly

[madjak]

As for compression? Ive got 12:1 motors on 25psi.

Dann

What duration cam is that with? Probably stock still I'm guessing. Say I aim for 274-284 degree cam do you think 13:1 would be ok?

[NitroDann]

225+ @ 050 and 400 lift

[madjak]

CNC Machining the Combustion Camber.

I have a small CNC machine at home which is just big enough to fit a MX5 head on the deck. The machine is a Roland MDX-650 with a 4-axis option. It's fairly slow going as the biggest tool I tend to use in the machine is a 6mm bit but thats pretty much all I need for the combustion chamber.

The first task is to build an accurate model of the existing head, then I'll model my target chamber around that. I've found the best software for this is AGI Photoscan. Basically the process is to take several rings of photos around the target object at varying angles and then pass them through the software to build a 3D model. You need to be careful how you take the photos, in that shiny object don't work very well, and you need to make sure the lighting doesn't change. An old dirty head with lots of carbon build up is perfect to scan. Also you need to keep the object in focus so ironically a good mobile phone is one of the best cameras.

here is the end outcome:



You can see the model is fairly clean. Where there is noise is due to the specular reflections in the photos. You can spray the object with matt paint or dust it with chalk, but for this purpose the model is good enough.

Next I load that up in a 3D application and scale the model to a known dimension. In this case I measured the width of the chamber and scaled it to that. The model also needs to be centered around the origin which I set as the center of the cylinder.

The 3D Modelling process if fairly easy. I use the photoscan model as a rough guide but modify it where I think I need to deshroud the valves. In this case, I've widened the chamber out to 84mm which is the size of the piston I'll be running, and moved the boundary around the valves out a few mm. This model also smooths out the factor 'ledge' where the radius of the chamber is cut by the seat insert. I need to be carefully not to take out too much material as it will effect the combustion ratio.




Here is the overlay of the original and the target:


The head then gets a first roughing pass with the depth lifted 0.5mm and inset 0.2mm so that I can check alignment.



Once I'm happy with the positioning I can drop the model closer to it's final height. The seats are very hard on the cutters, but on this head the seats are proud of the surface. This means I have to take them down a bit. Each pass at this level takes around 20 minutes.



This next photo is to check where the model is cutting the surface. Ideally I need everything machined so from here I modify the target 3D surface to push it out to machine all areas.



And here is the final finishing pass. This pass takes around 1 hour.



Here is the final surface with a little bit of 240 grit sanding. The CNC process now gets finalised and offset for each chamber. There is a roughing pass and final finishing pass which takes around two hours per chamber.



Finally, here is a video of the machine running:

[madjak]

The CNC machine leaves very fine machine marks which look nasty but they feel smooth to your hand. Similar to mowing the grass on a football pitch, the direction it's mowed pushes the grass over so that it looks different but it's still very flat. So it only takes a few minutes with some sand paper to smooth out the machining. I think I did 240 grit then 400 grit then zapped it with a buffer on a dremel.

Here is the final result. The video is useful because you can see the changing reflections on the surface.


Next I'll show you my porting. I can't fit the head into the CNC machine vertically so this part will be done by hand.

[madjak]

225+ @ 050 and 400 lift

That's around a 264 adv cam.

What do you think about 284 intake (254 @ .050) and 274 exhaust (246 @ .050). That's around 27 degrees overlap or is that going too far? I'll be able to tweak the overlap a little with adjustable cam pulleys when it's on the dyno.

With a supercharger the intake cam is the most important, plus whatever cam I run will be 10mm lift. I still want a bit of exhaust scavenging which you don't get with a turbo so larger duration cams should work for this build. If I get it wrong and have too much overlap, the boost and fuel out just be blown out the exhaust wasting power. At some point the exhaust scavenging and overlap should help pull out the exhaust gasses and suck in the clean air and fuel without any waste, but I'm not sure where that crossover point is. Hopefully around 5-6000rpm.

The Honda boys with Rotrexes run big cams so I can't see why it's not the same on a BP. The power curve does shift upwards so it's just a question of how much of a dyno queen this engine will be verses what will be fast on the track.

[StillIC]

Yeah you mentioned the 12:1 Wossner pistons before. Have you actually seen them? They list an OEM compression ratio of 11:1 which is obviously wrong so I doubt they are actuall 12:1.

Given there is very little clearance room on the peak of the 10.5:1 Wiseco piston to the head i can't see where the meat for the extra compression is going for 12:1. There is really only the area between the valves.

I think you'd need to double check the specs on the Wossners.

Yes, I use Wossners in my B6. Yes, they are domed to get the compression. Yes they do make 12:1 and are available in every 0.5mm bore size you'd want for a BP. E.g.
http://www.wossnerpistons.com/products/product-details/id/WOS-9078D200/name/wossner-piston-mazda-miata-1-8ltr-16v-85-mm-bore

[StillIC]

Also, if you can build a 400hp mechanically supercharged engine you can build a >>400hp turbocharged engine with the same basic engine parameters (fueling, air supply, compression, boost etc.), due to their higher efficiency (read lower mechanical losses).

Well, yeah and nah. A rotrex engine will never ever consume the same amount of air per cycle that an equivalent turbocharged engine can. So its likely that this build will not need to be as stout in any way from mechanical strength to tune quality that an equal power turbocharged engine would need.

As for compression? Ive got 12:1 motors on 25psi.

Dann

Let me say what I said anther way. A 400hp Rotrex engine will consume *more* air than a 400hp turbocharged engine, as the turboed engine is more efficient. I think this is the opposite of what you are saying. I therefore think your logic is flawed.

[madjak]

Yes, I use Wossners in my B6. Yes, they are domed to get the compression. Yes they do make 12:1 and are available in every 0.5mm bore size you'd want for a BP. E.g.
http://www.wossnerpistons.com/products/product-details/id/WOS-9078D200/name/wossner-piston-mazda-miata-1-8ltr-16v-85-mm-bore

Here is a orbit of the pistons. Actually you can see the peak goes right to the edges of the piston (sloped to the sides) where as the Wiseco 10.5:1 pistons have a step here at the base piston height which is where the extra compression is. The issue with the larger peak is total weight.
http://www.wossnerpistons.com/images/store/wossner-usa-inc/WOS-9078D200.swf

For me though I'm just going to stick to the pistons I know and deck the head to suit. The Wiseco's are lighter which is a big factor for a high reving engine. My current N/A race engine is somewhere around the 14 - 15:1 compression ratio with the 10.5: Wisecos and around a 2mm decked head. In this case I'm just not going to take it as far.

Wossner: 326 g
Wiseco: 293 g

[madjak]

Let me say what I said anther way. A 400hp Rotrex engine will consume *more* air than a 400hp turbocharged engine, as the turboed engine is more efficient. I think this is the opposite of what you are saying. I therefore think your logic is flawed.

It's complex though. Sure the Rotrex is sapping 50-60hp from the engine whilst the turbo is not. However the turbo is also adding back pressure to the exhaust causing less volumetric efficiency.

Also a turbo will hit it's torque limit much lower in the revs putting far more strain on engine components than the same HP Rotrex which I think it what Dann is referring too.

So a 400whp turbo engine hitting peak torque at 6,000rpm will be making 350 ft/lbs of torque. This will load rods and all the drivetrain.
A 400whp Rotrex engine hitting peak torque at 8,000rpm will be only making 263 ft/lbs of torque. Well under the limit of a 6 speed. Of course the rods will still be taking a hammering as they are loaded with power from the ignition as well as stretching from the revs. But most forged rods are well over spec.

For this to work, I need to build a very efficient engine that can breath well and flow air in and out of the cylinders efficiently. To do this it needs as much cam as possible and high compression. Then the Rotrex needs to be flowing enough air to add the required boost without running out of puff. That's which is why I have specced the larger C38 series blower.

[madjak]

This power... in an MX5 is my goal. Well maybe a bit less power

[StillIC]

It's complex though. Sure the Rotrex is sapping 50-60hp from the engine whilst the turbo is not. However the turbo is also adding back pressure to the exhaust causing less volumetric efficiency.

Also a turbo will hit it's torque limit much lower in the revs putting far more strain on engine components than the same HP Rotrex which I think it what Dann is referring too.

So a 400whp turbo engine hitting peak torque at 6,000rpm will be making 350 ft/lbs of torque. This will load rods and all the drivetrain.
A 400whp Rotrex engine hitting peak torque at 8,000rpm will be only making 263 ft/lbs of torque. Well under the limit of a 6 speed. Of course the rods will still be taking a hammering as they are loaded with power from the ignition as well as stretching from the revs. But most forged rods are well over spec.

For this to work, I need to build a very efficient engine that can breath well and flow air in and out of the cylinders efficiently. To do this it needs as much cam as possible and high compression. Then the Rotrex needs to be flowing enough air to add the required boost without running out of puff. That's which is why I have specced the larger C38 series blower.

I'd argue that the turbo doesn't reduce volumetric efficiency due to the back pressure, as the thing causing the back pressure is connected, via a shaft, to another thing producing 'front pressure'! And it is using otherwise wasted energy to do so.

But you have given a good explanation of the difference between your 400hp Rotrexed engine and a 400hp turboed engine. Well done.

[datfreak]

Your builds are always left field and interesting, I cant wait to see the results of your work- subed.
The cnc chamber work is so good - I've got a spare Na8 head with +1 os values and porting if you want another head to experiment on
Are you going to go with china rods? At one point I was hoping to organise a group buy but iirc the min order was 5 sets at about $250au/set. There are some on ebay for about $300 landed :http://www.ebay.com.au/itm/181287280067

May be a silly irreverent question but does the crank and pulley mating have enough key strength to hold 50-60 hp(plus alt) load for the supercharger?

[NitroDann]

StilliC.

The key phrase was "per cycle".

The actual force on everything from the rings to the gearbox input shaft at its peak. It's far greater in a turbo car. Hence the HUGE torque turbo engines make.

That's what makes it harder to build a 400whp turbo than rotrex.

The rotrex simply never fills the cylinder with as much explosion.

Dann

[StillIC]

StilliC.

The key phrase was "per cycle".

The actual force on everything from the rings to the gearbox input shaft at its peak. It's far greater in a turbo car. Hence the HUGE torque turbo engines make.

That's what makes it harder to build a 400whp turbo than rotrex.

The rotrex simply never fills the cylinder with as much explosion.

Dann

I accept this might be typically true, based on madjak's comparison between his planned engine and a *typical* 400hp turboed engine....more revs at peak power for his rotrexed engine therefore less air per cycle at peak power.

But there is nothing stopping anyone (happy to hear your opinion on this Dann) from building a high revving turbo engine, with an electronically controlled boost curve that imitates the rotrex boost curve and air consumption, but ultimately makes more power. That is, same boost and air used 'per cycle' at all engine speeds, but more output because there is not a ~50hp drain on spinning the rotrex. I do accept that at low revs the turbo will suffer lag and be unable to match the boost of the rotrex down low.

[NitroDann]

I accept this might be typically true, based on madjak's comparison between his planned engine and a *typical* 400hp turboed engine....more revs at peak power for his rotrexed engine therefore less air per cycle at peak power.

But there is nothing stopping anyone (happy to hear your opinion on this Dann) from building a high revving turbo engine, with an electronically controlled boost curve that imitates the rotrex boost curve and air consumption, but ultimately makes more power.

This is literally how I build 200rwkw stock motor mx5s that don't break

That is, same boost and air used 'per cycle' at all engine speeds, but more output because there is not a ~50hp drain on spinning the rotrex. I do accept that at low revs the turbo will suffer lag and be unable to match the boost of the rotrex down low.

You're actually wrong about low revs, the turbo eats the rotrex for outright torque at the bottom, its completely inarguable infact.

Good insight on your behalf, btw.