Twin Turbo 928 fixed and back out there terrorizing the streets!
#616
Rennlist Member
Tuomo and John are quite capable of eliminating the knocks they had during their last experiment. They have the tools to do so, but didn't think it was important to retune for those runs.
#617
Considering all the work that's been carried out on this project I'm surprised that the old Bosch ECU was not upgraded to a more modern speed density unit using closed-loop knock control/boost control and all the flexibility of a modern interface.
Was the engine harness rebuilt?
Was the engine harness rebuilt?
#618
Nordschleife Master
Thread Starter
Considering all the work that's been carried out on this project I'm surprised that the old Bosch ECU was not upgraded to a more modern speed density unit using closed-loop knock control/boost control and all the flexibility of a modern interface. Was the engine harness rebuilt?
Part of the challenge we have is to try to push the stock parts as far as we can. Projects are only fun if there are some constraints, otherwise the solution ends up being buying a modern, new car in pieces. It's the curse of no financial constraints that you have to come up with some other constraints.
It is however an interesting question where the stock ECUs fall short of the maximum potential. I think there are a couple of areas:
(1) more resolution at higher rpms. (This is being remedied by JDS.)
(2) the ignition and fuel grids to coincide. This is simply to make tuning faster.
(3) more columns and rows for the fuel and ignition maps
(4) individual cylinder fuel and ignition trim maps
(5) boost control integrated with ignition control and knock sensors
(6) traction control integrated to the ignition control and boost control
(7) throttle position switch input instead of a continuous throttle position sensor input
It appears that I am in a minority, but I like the hot-wire mass air flow load measurement. It's very flexible and self correcting. There's some flow restriction, but I think that's really minor. I wouldn't go to speed-density as long as using the stock intake manifold. With a "sheet-metal" intake, it probably would make sense to go to speed-density. By the way, JDS are working on a speed-density conversion kit for the stock ECU's.
Speaking of sheet metal intakes, I am not yet convinced that the stock intake is the most binding bottleneck in a boosted car. So far, we seem to be doing fine with it. At some point it will become the bottleneck but I am interested how far it can be pushed.
Another system that I don't think is going to be a bottleneck soon is the ignition. This is because we're running 93-octane pump gas. With alternative fuels, the ignition would probably have to changed, but for pump gas the stock ignition appears sufficient.
#619
Nordschleife Master
Thread Starter
WE are taking whatever Tuomo shares with us. I have my own theories about what he's modelling, and what variables he's anticipating changing, but they're based purely on conjecture and inference, so they're not worth posting, as I'd say I have maybe a 30% chance of being right.
Interesting - is there any electronic/electromechanical way of detecting these pre-detonation events? Thanks for the info, I'll have to do some research, although I suspect that any equipment to measure/detect knock precursors will be outside of my price bracket. If its "judgement" or "experience" based, then that could be an interesting thesis for a Mech Eng. - using machine learning and suitable inputs (sensors, microphone, etc.). If it could be packaged, you could do some crazy stuff with ECU's.
I think you are essentially correct in that you have to let the engine knock somewhat when mapping it because otherwise there is no way to learn where the knock threshold is. The question is how efficient your learning method is. How quickly can you learn the knock limit (or the knock probability function) with the fewest possible actual knocks.
One piece of equipment on my wish list is this: http://www.tfxengine.com/ It is expensive starting at $7k and if I buy it I don't want it to collect dust in a box, so I'll only pull the trigger once I know I can get it installed immediately when delivered. This equipment allows measuring irregular combustion the same way that car factories measure it, with a pressure sensor in the combustion chamber. I don't think that this equipment changes fact that you have to go to the irregular combustion region in order to learn where that region is, it'll just make it easier to learn a lot quickly once there.
Last edited by ptuomov; 05-03-2013 at 08:46 AM.
#620
Nordschleife Master
Thread Starter
The last dyno session was ended because the waste gates were cracking open. New, stronger, waste gate actuators were installed which allowed John to reach slightly higher boost levels than before. at that point, according to Tuomo, there was knock. There was also a boost leak or some other anomaly that occurred that caused them to end the testing and tuning once again. John is taking TONS of data during all of these runs but I doubt he will make 100% of it public.
Last edited by ptuomov; 05-03-2013 at 08:47 AM.
#621
I find it odd making a point in keeping the stock ECU when so manys changes have been done to the engine, and that was not "only" modifying external parts such as exhaust/turbos but also replacing inner parts such as pistons to lower the compression ratio.
I understand that moving to an aftermarket ECU and building the engine harness may represent a major investment (not just in money but also in personal free time) for the average DIYer, but using a speed density-based ECU should make turbo/intake/exhaust swaps a doddle compared with a MAF system - BTDT.
Of course we can still use a MAP signal to ponderate fuel/ignition tables in the stock ECU but that is not the right way to do it, IMO.
Anyway, thanks for the open discussion. I totally relate to the "if it's too easy it's not worth doing" stance, and am looking forward to hearing of your progress.
I understand that moving to an aftermarket ECU and building the engine harness may represent a major investment (not just in money but also in personal free time) for the average DIYer, but using a speed density-based ECU should make turbo/intake/exhaust swaps a doddle compared with a MAF system - BTDT.
Of course we can still use a MAP signal to ponderate fuel/ignition tables in the stock ECU but that is not the right way to do it, IMO.
Anyway, thanks for the open discussion. I totally relate to the "if it's too easy it's not worth doing" stance, and am looking forward to hearing of your progress.
#622
Nordschleife Master
Thread Starter
I personally find it remarkable that John can get 700hp out of the completely stock engine with reprogrammed ECUs and that the whole setup currently on the car can be replicated for about $20k.
I understand that moving to an aftermarket ECU and building the engine harness may represent a major investment (not just in money but also in personal free time) for the average DIYer, but using a speed density-based ECU should make turbo/intake/exhaust swaps a doddle compared with a MAF system - BTDT. Of course we can still use a MAP signal to ponderate fuel/ignition tables in the stock ECU but that is not the right way to do it, IMO.
When more internal engine modifications are made and as the engine starts resembling the stock engine less and less, we'll also reevaluate the question of what engine management system to use.
Last edited by ptuomov; 05-03-2013 at 09:16 AM.
#623
Attachment 724800
John is a pro and I am an amateur. However, despite being an amateur, I enjoy theory and John's happy to try out some of my theories. Not everything I suggest pans out the way I expect it to.
John is a pro and I am an amateur. However, despite being an amateur, I enjoy theory and John's happy to try out some of my theories. Not everything I suggest pans out the way I expect it to.
We don't have anything like the data acquisition machinery that Jim Morton probably has or the multi-sensor wbo2 system that Todd Tremel has shown. That said, we log data and take notes. Furthermore, we run computer simulations that we try to calibrate to match the data.
For example, the most recent set of experiments today were about figuring out how much boost we can generate at which rpms. The objective was not to find out how strong the motor was, instead it was to figure out whether the new wastegate springs solved the boost control problem we had. The answer is it didn't, but we have some leads where to look next.
Those experiments today were run on pump gas. The main reason is that it's cheaper to run them that way, rather than retune the ignition map for a race gas, and because we run them on Dynojet. Was this safe? I argue it was safe enough. It's really hard to break a 700 hp engine because of detonation on an intertia chassis dyno.
Is this because you need to know more about the wastegate and its operation? Have you spoken to the manufacturer?
#624
Considering all the work that's been carried out on this project I'm surprised that the old Bosch ECU was not upgraded to a more modern speed density unit using closed-loop knock control/boost control and all the flexibility of a modern interface.
Was the engine harness rebuilt?
Was the engine harness rebuilt?
#625
Stock ignition systems don't always work with alternative fuels and sprays BECAUSE the Mixture is MORE difficult to ignite ON PURPOSE. Its knock resistance goes up, as does its resistance to actually lighting it off properly. So what essentially you are saying, emphatically above, is that you are using an inferior ignition system because you know the fuel is inferior and is capable of disastrous levels of detonation within the load and air flow ranges you are tuning in.
#626
Nordschleife Master
Thread Starter
We tune pretty good-sized gas turbines with microphones on combustor cans, and do software analysis of the various sound patterns that combustion generates along with the combustion products. No such thing as knocks so no mechanical knock sensors. Fun stuff! There really is a difference between IC and EC engines though, so no direct correlation between what we listen for vs what you might listen for in the IC engine as a 'precursor' of real detonation. Something to think about for IC engines?
For debugging purposes, the car was run with a constant boost profile today. It's simpler than the rpm specific profiles, and eliminates one potential issue. The knock retard was on since this what not a mapping session.
For those of you who are curious, this is what the power and torque curves look like when the car is run on a constant boost profile:
The peak power is in the same ball park as the graph I posted earlier, but now we don't have that flat torque curve. Instead, we have a flat power curve, making over 600 rwhp in the 4300-6000 rpm range.
More generally, we're hitting some sort of bottleneck at about 700 rwhp. The main hypotheses right now are boost leak that shows up at 18 psi (for example something flexing), low pressure cold side restriction, or some control issue.
Hitting bottlenecks by the way is not unexpected. John designed his kit to be correctly sized, with margins of safety, for about 600 rwhp, and we're significantly over that. In a well engineered kit, everything runs out of capacity at about the same point because it's correctly sized for the application, with a margin of safety. Now we just have to identify the bottlenecks one by one and keep solving them.
This is a science project. If you just want drive a fast car, I can say that for someone looking 600 rwhp the existing twin turbo kit is a completely problem free system.
#627
Racer
Join Date: Jan 2009
Location: Adelaide South Australia'79 5spd twin turbo
Posts: 326
Likes: 0
Received 1 Like
on
1 Post
Go You Good Guys !
Keep it up Tuomo and John, There be Dragons out there, but I have no doubt you two will succeed in pushing to a limit.
Why on earth anyone would need over the 600 hp the kit was designed to deliver, is beyond me, but wanting it is no crime.
Bring it on.
Cheers Guys
Keep it up Tuomo and John, There be Dragons out there, but I have no doubt you two will succeed in pushing to a limit.
Why on earth anyone would need over the 600 hp the kit was designed to deliver, is beyond me, but wanting it is no crime.
Bring it on.
Cheers Guys
#628
Race Car
I think the bigger picture is how much can you get out of a stock set-up. If the goal was to get the the highest H.P. 928 out on the road it would be done by now. (what about our Nos enriched British friend was he not over 1000 H.P.
I am seeing a lot of resources, money and time fuled by the passion to see how close to the edge you can take it and still turn the key and drive it every day.
This project is in very capable hands the upside is all the R&D is being funded at no cost to us and will only filter down to the 928 community.
This thread seems to be like paper in the scientific journal a hypothesis is presented skeptics bring up that your methods are wrong and advise how you should be doing it. Either way how cool is this and waiting on the end number!
I am seeing a lot of resources, money and time fuled by the passion to see how close to the edge you can take it and still turn the key and drive it every day.
This project is in very capable hands the upside is all the R&D is being funded at no cost to us and will only filter down to the 928 community.
This thread seems to be like paper in the scientific journal a hypothesis is presented skeptics bring up that your methods are wrong and advise how you should be doing it. Either way how cool is this and waiting on the end number!
#629
Nordschleife Master
Thread Starter
Small news:
The car can now get the boost psi up to low 20's at high rpms while still keeping the turbo shaft speed at 130k. Logging data is slow on the roads for obvious reasons, although perhaps "slow" is not the right word ... let's rather say "difficult."
It looks like there will be one final dyno data collection session before the old engine and old turbos come out and the low compression engine and the cut-weld-machine hybrid Frankenturbos go in.
A quick triangulation gives a very rough estimate of where we are on the compressor map:
"There be dragons."
The car can now get the boost psi up to low 20's at high rpms while still keeping the turbo shaft speed at 130k. Logging data is slow on the roads for obvious reasons, although perhaps "slow" is not the right word ... let's rather say "difficult."
It looks like there will be one final dyno data collection session before the old engine and old turbos come out and the low compression engine and the cut-weld-machine hybrid Frankenturbos go in.
A quick triangulation gives a very rough estimate of where we are on the compressor map:
"There be dragons."
Last edited by ptuomov; 06-04-2013 at 01:43 AM.