When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
I've been thinking more about the pistons that one might want to use in a pump-gas street-driven twin-turbo 928.
The way I see the piston design for a specific application working is that one first picks the bore, stroke, and max rpm. I think that for a twin-turbo car, you want to keep the cylinder towers as thick as possible. Smaller bore will also help with pump gas. Short of dry-sleeving the block with high-quality steel sleeves that would add to strength, this means sticking with 100mm nikasil bore (or maybe 100.5mm if the existing bores have damage). The two stroke options are stock 78.9mm or the largest commonly used stroke of 95.25mm. Given how the intake and the turbo systems work, there are two rpm limits that one could consider: Option 1 which is 6200 rpm peak power and 6800 rpm redline. Option 2 which is 7000 rpm peak power and 7700 rpm redline.
Then, the max rpm and stroke gives one the ring width based on predicted flutter critical rpm. My simple math says that option 1 (6800 rpm redline) would require 1.5mm (stock width) top ring with stock 78.9mm stroke and 1.2mm top ring with 95.25mm stroke. Option 2 (7700 rpm redline) would require 1.2mm top ring with stock stroke and 1.0mm top ring with 95.25mm stroke.
The stroke and rpm give one the maximum weight for the piston assembly, which with bore size then determines the compression height, piston design, and how expensive the piston materials and manufacturing will end up being to get under the maximum weight. If for stock stroke the stock piston assembly is acceptable weight at the stock 6800 rpm GT redline, then we have 765g total piston assembly weight there, which is option 1 with stocks stroke. Option 2 (7700 rpm redline) with stock stroke would require 596g piston assembly. Option 1 (6800 rpm redline) with 95.25 stroke would require 525g piston assembly. This is already quite challenging as the Mahle Subary pistons are coming out at about 550g. Option 2 (7700 rpm redline) with 95.25mm stroke would require 409g piston assembly, which is at minimum very difficult and probably impossible with my budget and skills. So we can rule out the option of a factory-stock reliable 7700 rpm redline 95.25 stroke motor.
Thus, the three remaining options are:
A. 100mm bore, stock 78.9 stroke, 6800 rpm redline, 1.5mm top ring, and maximum 765g total piston assembly weight.
B. 100mm bore, stock 78.9 stroke, 7700 rpm redline, 1.2mm top ring, and maximum 596g total piston assembly weight.
C. 100mm bore, stock 95.25 stroke, 6800 rpm redline, 1.0mm top ring, and maximum 525g total piston assembly weight.
The power level, ring width, and top land width will give one the heat flow equations and how much additional cooling one needs to keep the top ring and top ring groove below critical temperature. The additional cooling need estimate gives one the answer to whether the piston cooling jets are required and what is the required oil flow rate (Mahle book gives the heat rejection rates as a function of oil flow rate for piston squirters). Installing squirters to any 928 block probably isn't an insurmountable problem, but if the squirters are needed, I think option B will possibly require a dry sump and depressed crankcase gas density and option C will in my opinion surely require a dry sump and depressed crankcase density to be reliable and not puke oil all over the place.
There are 100 other things to figure out, too, which I don’t really either understand or know about. The above design problem is so complex that I don’t think I could do it (no shocker there). I don’t even think that some lesser aftermarket piston manufacturers can really do it correctly, if one goes right to the edge in terms of requirements. And as a hobbyist, ordering say 16 pistons, it’s probably impossible to get attention from anyone in any company who can actually do the design work correctly. So why bother even making these computations? I think they are useful in ruling out options that are likely dead ends if one wants stock-like reliability.
That looks REALLY ghetto. All this to not put in a modern aftermarket ECU. The only reason to not do an aftermarket ECU in 2018 is if the smog test needs an ODB2 port.
That looks REALLY ghetto. All this to not put in a modern aftermarket ECU. The only reason to not do an aftermarket ECU in 2018 is if the smog test needs an ODB2 port.
Do you have any idea what’s happening there? (It’s a rhetorical question.) Which of these wires going to the dyno data acquisition do you think one would not want to have connected if one were trying to calibrate another ECU?
Do you have any idea what’s happening there? (It’s a rhetorical question.) Which of these wires going to the dyno data acquisition do you think one would not want to have connected if one were trying to calibrate another ECU?
I can extrapolate from your stories here, and from John WAY before you even existed on this forum. I also know how modern ECUs work, as well as how those looms would be wired. In addition, I have extensive knowledge of modern data acquisition.
You don't need to obfuscate. The way those wires are twisted together gives me information enough. No need for your snark. Its ghetto.
Do you have any plans to DRIVE the car? When everyone else is afraid to replace trim, let alone cross pollinating engines, etc....
I guess enjoy your science experiment. Just irks me to think people may think you and/or John actually have any actual new info or this path has not already been worn by people before you.
Last edited by BC; 07-02-2018 at 04:15 PM.
Reason: Iterate
I can extrapolate from your stories here, and from John WAY before you even existed on this forum. I also know how modern ECUs work, as well as how those looms would be wired. In addition, I have extensive knowledge of modern data acquisition.
You don't need to obfuscate. The way those wires are twisted together give me the info. Do you have any plans to DRIVE the car?
So the answer to my rhetorical question appears to be "no". No surprise there.
Let me assure you that we’re not inventing anything new about internal combustion engine in this project.
In any case, thanks for your continuing encouragement!
That looks REALLY ghetto. All this to not put in a modern aftermarket ECU. The only reason to not do an aftermarket ECU in 2018 is if the smog test needs an ODB2 port.
On your planet, in what years were any 928s ODB2 compliant?
This is a hobby science projects, and one of the objectives is to change original factory parts only when it’s proven beyond reasonable doubt that there’s a major gain from replacing them with other parts and even then trying to be cost effective about it. The reason being “because we can”.
My personal interest in this particular experiment is how consistent various load measurement methods are for this engine.
Right now, we don't have a throttle position sensor installed, just the stock throttle position switch, but we're making progress on other dimensions. Specifically, I can think of four ways to measure the load on this engine:
Hot film / hot wire mass air flow sensor (both sides), which is then divided by rpm.
Speed density calculation from pressure and temperature sensors in the intake manifold downstream of the throttle plate, combined with a volumetric efficiency vector (per rpm).
Alpha-N calculation with the pressure and temperature sensors in the intake system upstream of the throttle plate, combined with a volumetric efficiency table (per throttle position and rpm).
Turbocharger speed sensor combined with the pressure and temperature measurements upstream and downstream of the compressor (both sides), from which the mass air flow can be solved using the compressor map and then divided by rpm.
I'm very curious how consistent these load measurements are between each other, and how much noise each of them have depending on the load and rpm. Not that there's any immediate practical use for this information, but I'm just curious. Hooking it all up in this way will furthermore allow us to solve for a number of things. For example, since the compressor covers are custom, we can use these data to graph the actual compressor map with these custom covers. We can estimate the volumetric efficiency per two definitions (I already did one). We can match the 1D simulation results better to the engine. Etc.
As painful as it is for humans to work in hot climates, tuning a turbo car during a heat wave amounts to preparing for the worst case scenario. It’s right now 117F indoors there, and the car is holding all temps rock steady. This won’t be one of those boosted cars that one can’t drive during the day in Florida. Hell, it’ll climb out Death Valley on a summer afternoon, with the A/C and stereo on!
That looks REALLY ghetto. All this to not put in a modern aftermarket ECU. The only reason to not do an aftermarket ECU in 2018 is if the smog test needs an ODB2 port.
The “REALLY ghetto” #80 lbs injectors, dual MAF, and LH 2.3 runs well now even in closed loop and idle:
We used Jim Corenman’s algorithm to estimate the Siemens Deka 80 lbs injector opening time with our fuel rail to intake manifold pressure differential, and got 0.94 ms at 13.1V battery supply voltage:
I've been thinking more about the pistons that one might want to use in a pump-gas street-driven twin-turbo 928.
The way I see the piston design for a specific application working is that one first picks the bore, stroke, and max rpm. I think that for a twin-turbo car, you want to keep the cylinder towers as thick as possible. Smaller bore will also help with pump gas. Short of dry-sleeving the block with high-quality steel sleeves that would add to strength, this means sticking with 100mm nikasil bore (or maybe 100.5mm if the existing bores have damage). The two stroke options are stock 78.9mm or the largest commonly used stroke of 95.25mm. Given how the intake and the turbo systems work, there are two rpm limits that one could consider: Option 1 which is 6200 rpm peak power and 6800 rpm redline. Option 2 which is 7000 rpm peak power and 7700 rpm redline.
Then, the max rpm and stroke gives one the ring width based on predicted flutter critical rpm. My simple math says that option 1 (6800 rpm redline) would require 1.5mm (stock width) top ring with stock 78.9mm stroke and 1.2mm top ring with 95.25mm stroke. Option 2 (7700 rpm redline) would require 1.2mm top ring with stock stroke and 1.0mm top ring with 95.25mm stroke.
The stroke and rpm give one the maximum weight for the piston assembly, which with bore size then determines the compression height, piston design, and how expensive the piston materials and manufacturing will end up being to get under the maximum weight. If for stock stroke the stock piston assembly is acceptable weight at the stock 6800 rpm GT redline, then we have 765g total piston assembly weight there, which is option 1 with stocks stroke. Option 2 (7700 rpm redline) with stock stroke would require 596g piston assembly. Option 1 (6800 rpm redline) with 95.25 stroke would require 525g piston assembly. This is already quite challenging as the Mahle Subary pistons are coming out at about 550g. Option 2 (7700 rpm redline) with 95.25mm stroke would require 409g piston assembly, which is at minimum very difficult and probably impossible with my budget and skills. So we can rule out the option of a factory-stock reliable 7700 rpm redline 95.25 stroke motor.
Thus, the three remaining options are:
A. 100mm bore, stock 78.9 stroke, 6800 rpm redline, 1.5mm top ring, and maximum 765g total piston assembly weight.
B. 100mm bore, stock 78.9 stroke, 7700 rpm redline, 1.2mm top ring, and maximum 596g total piston assembly weight.
C. 100mm bore, stock 95.25 stroke, 6800 rpm redline, 1.0mm top ring, and maximum 525g total piston assembly weight.
The power level, ring width, and top land width will give one the heat flow equations and how much additional cooling one needs to keep the top ring and top ring groove below critical temperature. The additional cooling need estimate gives one the answer to whether the piston cooling jets are required and what is the required oil flow rate (Mahle book gives the heat rejection rates as a function of oil flow rate for piston squirters). Installing squirters to any 928 block probably isn't an insurmountable problem, but if the squirters are needed, I think option B will possibly require a dry sump and depressed crankcase gas density and option C will in my opinion surely require a dry sump and depressed crankcase density to be reliable and not puke oil all over the place.
There are 100 other things to figure out, too, which I don’t really either understand or know about. The above design problem is so complex that I don’t think I could do it (no shocker there). I don’t even think that some lesser aftermarket piston manufacturers can really do it correctly, if one goes right to the edge in terms of requirements. And as a hobbyist, ordering say 16 pistons, it’s probably impossible to get attention from anyone in any company who can actually do the design work correctly. So why bother even making these computations? I think they are useful in ruling out options that are likely dead ends if one wants stock-like reliability.
Continuing my monologue on piston oil squirters:
My theoretical logic for installing piston oil squirters to stroked and/or high rpm engines is the following: Suppose you want a 968/928/944 to run at higher rpms and with longer-than-stock stroke. This by my logic requires thinner top ring than the early 1.5mm 928/944 top ring and especially the late 968/928/944 1.75mm top ring. Otherwise, it’s just going to flutter and cause all kinds of problems. Say 1.2mm with stock stroke and high rpms and 1.0mm with longer stroke and high rpms. Longer stroke and/or higher rpms will also require lighter pistons. The lighter pistons have to run cooler than heavier pistons, so cooling needs to be improved. At the same time, the thinner rings will reduce cooling and make the piston run hotter. Finally, the power levels are up a lot (otherwise, what’s the point?), so the heat flow needs are generally up.
I don’t see many other solutions here than piston oil cooling jets. Do you?
I agree with the factory that at stock 928 S4 piston weight, stock ring width, stock stroke, and stock redline rpm, the squirters may hurt more than help. But this is with higher rpms, longer stroke, thinner rings, lighter pistons, and more power.
So how much oil would one need to flow thru the squirters?
I’m looking at the Mahle guidelines for piston oil squirter flow rates from their piston manual. They say that the reference implementation uses 3-5 kg of engine oil per kWh. If I’m sizing the system for peak power, that is, 3-5 kg/h per kW of peak power, perhaps going to the low end of the range due to the duty cycle.
If one wants to make say 125 rwhp per cylinder, that’s about 110 kW at the crankshaft. This means that each cylinder needs 330 kg/h of oil. This in volume flow rate using my oil is 388 l/h or 1.71 gpm per cylinder. For all 8 cylinders that’s 13.7 gpm of additional oil that would need to be supplied by the oil pump and then drained again, on time, to the oil sump! I don’t immediately see how that could work out with the stock wet sump setup...
The options seem to be to go dry sump with a big *** reservoir or use squirters that flow a lot less than what the Mahle reference implementation and peak power would suggest (for example, actually think thru the duty cycle of the engine in use and size the piston oil squirters based on maximum kW actually used in any 10 minute interval) or both. Thoughts?
The car idles well in all operating temperatures that can be tested in summertime Kentucky (day or night). Idle rpm holds steady.
On other news, my 12 year old surprised both me and the instructor by landing a 737 perfectly without any tire audio feedback on her first try in a real pilot training simulator.
07-02-2018, 10:08 AM
