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.
Continuing the testing of the JDS analog box. The car is getting there on low boost. John is at about 604 rwhp with 11.5 psi, which is pretty damn good but not 100% there yet. The data are also being accumulated for various boost profile adjustments, etc.
In in the meanwhile, I’m roughing it out here in Italy. We’re right now anchored in front of the old city of Monopoli. Octopus taste better grilled than raw, by the way:
This is a pump gas run with 16 psi boost. The peak power was 715 rwhp, which indicates good scaling with boost. Recall that my SWAG for the engine without boost is 340 rwhp and (16+14.7)/(0+14.7)*340 = 710 so it’s scaling nicely. With my simple formula for crank ho, I’d say we’re close to 820 at the crank. John’s goal on pump gas is 750 rwhp / 860 crank up, so boost still needs to go up.
John’s mapping process is inching up boost and rpm, little by little. The car is now at 730 rwhp at 6450 rpm, and this may be close to pump gas limits.
My highly personalized and super unofficial transmission loss formula would SWAG the engine power at 730/.88+10*(6450/6000)^2 = 841 crankshaft hp, where I’m taking torque losses at proportional 12% and transmission wind age losses at 10 hp at 6000 rpm. The engine power guesstimates are only relevant to trying to understand what is happening inside the engine anyway, so there’s no inherent meaning in that crank hp number anyway.
In the meanwhile, I was in a cooking class of all the places:
Stunning results from the dyno, but need more Italian food pictures. The wine is probably no good either, right?
Some of the wine is great or very good, and in the rare cases when it’s merely ok you can make it by drinking more of it.
In terms of tuning, one of the reasons why the car is getting tuned in 97F weather is that this is likely to give a safe tune on pump gas. The car will not want to make more than 730-750 rwhp in this weather on 93 pump fuel. I’m thinking that estimated 850 hp at the crank is enough for this iteration on pump gas.
Making more power on pump gas would require either much higher rpms, lower compression, or larger displacement. Larger displacement is counterproductive from the standpoint of running the 928 five speed transmission and wet sump oiling at high rpms. Lower compression is counterproductive from the perspective of off-boost torque and transmission longevity. Finally, higher rpms are counterproductive from the perspective of wet sump oiling. So it’s kind of where it is.
At each rpm, the maximum torque is constrained by knock probability on the one hand and exhaust gas temperature on the other hand. As the boost goes up, the tuning window between earlier ignition advance that increases the knock probability and later ignition retard that increases the exhaust gas temperatures gets narrower and narrower. There’s more fine tuning to be done at each 500 rpm band on this for sure.
The turbos, injectors, and piping are running at approximately 70% of the peak capacity now. There’s headroom for another 300 or so rwhp on high octane race gas from the system capacity perspective. On the one hand, this sort of multi use dyno mule engine isn’t ever going to be absolutely perfectly matched for any use. On the other hand, it’s an interesting little personal science project to play with.
If someone wanted a 750 rwhp turbo car that needs to run on pump gas but doesn’t need headroom for future race gas experiments, in my opinion I’d replicate the current setup in the car with 0.3 points lower compression (so about 8.2:1 instead of 8.5:1), the same or slightly smaller cams, 60 lbs instead of 80 lbs injectors, and one size smaller turbo hot side. This would get the turbine to spool earlier while still running within its natural capacity at peak power.
In Lecce, we visited the house of a friend. To drink wine, of course, but also to see the house. They excavated and made a display from the ancient soap factory under the house. The floor has windows and an access hatch to the archeological dig under it! It was unbelievably cool.
Tuomo, for more power why are you not considering water methanol injection?
Åke
More added complexity and another fluid reservoir, not a fan.
The idea of using the windshield wiper fluid and only when combustion is abnormal as a safety mechanism is an intriguing possibility, though.
I’m on a train in Finland and have some time to write about tuning under boost. There is a single graphical figure that (kind of) explains the whole turbo tuning issue, at least on pump gas.
With turbos, one has in most cases an unlimited amount of high rpm air mass flow available. The only case in which one is air-flow limited is the compressor flow being chocked as the compressor wheel tips travel at sonic speed. At compressor choke, further shaft speed increases no longer increase mass flow thru the compressor. We are currently only at about 70% of the max mass flow (if my memory servers me right) so none of the tuning issues are really about more air flow. More air flow can be had simply by a turn of a screw.
There are limits, however. The two limiting factors are:
(1) the knock limit, where knock means a combustion that starts normally but then temperature and pressure lead to abnormal combustion in the form of knock or detonation. This can be detected by the ancient EZ-K computer in its stock form.
(2) the pre-ignition limit, where pre-ignition means the charge igniting before the spark. This is not detectable by EZ-K, unless it also leads to detonation and even then it may be attributed to the wrong cylinder. The main proxy variable that is easily measurable indicator of high steady-state likelihood of pre-ignition is the exhaust gas temperature. Conditions that lead to pre-ignition are associated with high exhaust gas (and cylinder head) temperatures. When pre-ignition starts, the EGT may suddenly drop, complicating the issue. The conditions associated with pre-ignition are also associated with burned exhaust valves and turbine blades.
Given the fuel, rpm, and air-fuel mixture conditions, the tuning window amounts to something like this Mahle graph:
The black BMEP curves denote constant torque, so more torque means moving to upper right to another, higher BMEP curve. The x-axis is the spark advance and the y-axis is the boost pressure. 100 kPa is about one atmosphere. As one would expect, more spark advance and more manifold pressure lead to higher torque. In this Mahle test engine run on pump gas and within its tuning window, four degrees of additional spark advance amount to as much torque as 1.5 psi of additional boost does.
The tuning window lives between a rock and the hard place. The rock on the upper right is the detonation limit. Additional boost and/or additional spark timing will move the engine to the knock area in the graph, leading to detonation. The hard place on the upper left is the pre-ignition limit. Too much ignition retard will lead to higher EGT which will cause (or at least be associated with) pre-ignition and burned exhaust valves.
As the boost is increased, the rock and the hard place will converge, leaving no safe way to run the engine at that or higher boost.
If one is using a knock management system that only adjusts the spark timing, one can safely run less boost than if one runs a knock management system that adjusts both spark timing and boost. This is because a knock management system that only adjusts spark timing may inadvertently retard the spark timing to the pre-ignition region of the graph. Therefore, with this kind of knock management system, the maximum safe boost is on the horizontal line in the graph at the level where the distance between the pre-ignition limit and the knock limit is a greater number of spark advance degrees than the maximum retard of the knock management system.
Contrast that with the knock management system that adjusts both boost and spark advance. Now, the operating point moves diagonally to down-left on the graph, avoiding the pre-ignition limit. Consequently, a turbo engine with such a system can be run much closer to the “tip of the spear”, i.e., the area where the pre-ignition and knock constraints converge.
The EZ-K system in the 928 S4 has a former kind of spark-retard-only knock management system. For one to maximize a pump-gas 928 turbo engine, one would need to capture the knock indicator from the EZ-K in real time and feed that as a signal to the boost controller. Alternatively, one could adapt a boost control system such as SAAB’s APC with its independent simple knock detection system that would independently detect knock and reduce boost.
Currently, I’m not aware of practical ways to detect pre-ignition in real time. (Measuring the size of the hole in the piston has a lag that is too long for real-time control logic.) If the conditions that are associated with the onset of pre-ignition, most importantly either abnormally high or abnormally low EGT, can be measured, one could rig a water injection system to selectively turn on in high risk situations. However, this is really not a great method, because the EGT sensors react slowly to changes in EGT. The pre-ignition detection and management is largely an unresolved problem in “downsizing” engines with high pressure turbocharging.
Primary load map being calibrated for higher boost levels. The same basic transfer function can be used for race gas and pump gas:
(The other picture is of the girls ready go swimming in Iceland.)
The primary load map is an early stage LH map that maps the raw MAF signal to stoichiometric AFR. Usually one doesn’t have to touch it, but John (with JDS’s help) is knee deep into it because of the dual MAF conversion. The cruise map is then added to this map to deviate from the stoichiometric AFR. Because at high loads, the car can’t actually be run at 14.7 AFR, contrary to earlier promises, “there will be some math”.
08-02-2019, 06:19 AM
John’s tuning, I’m roughing it out in Italy
