How much can backpressure effect hp and turbo spooling?
06-16-2017, 04:37 PM
#61
9 Second Club
I have no idea if this would work, but it occurred to me as a test that would be easy to setup (if the pressure test is good). Can we remove the spring from one of the gates to prevent that turbo from working, and run the engine from just the other turbo purposefully? Obviously reduce boost pressure so the motor doesn't outflow it on the map. Then you could repeat the test with the other turbo and compare the data. If one of them is showing wild out of place data like a skyrocketing IAT or big difference in backpressure or slower spool character it would point to that side as a problem.
It would be a lot more sensible and practical just to throw a pair of speed sensors on the turbos, and boost would simply blow back out the other turbo.
06-16-2017, 11:05 PM
#69
Uploading some screenshots from the OP's tune file as well as from one of his logs. Should clear up a lot of the questions about the spark, fuel, etc.
Base Fuel Table:
Base Spark Table:
Target AFR Table:
Dome Control Table:
Boost Vs Time Table:
Fuel Tuning from Datalog:
Sensors from Datalog:
Boost Tuning from Datalog:
Base Fuel Table:
Base Spark Table:
Target AFR Table:
Dome Control Table:
Boost Vs Time Table:
Fuel Tuning from Datalog:
Sensors from Datalog:
Boost Tuning from Datalog:
06-17-2017, 02:17 AM
#70
TECH Apprentice
so are you using just 1000 rpm in second (6000-7000)? staying at the same power level for two seconds? duty at 50% max? what is fuel flow? is inj. on time available? the stuff looks good. maybe put a restrictor before the exhaust sensor. is that a closed loop system?
06-17-2017, 08:47 AM
#72
06-17-2017, 11:53 AM
#76
wow stevie, I wonder if that is what is happening now? I notice the backpressure oscillates as if a turbo is 'letting go' and catching back up. I know we expect the graph to move up and down some, but that one is moving slowly, and the difference is significant in my opinion. From 28psi to 37psi what looks like 3 to 5 times per second. Normally when a sensor has a sensitivity oscillation it is something like 100 to 1000 times per second, and the unit would be more near its mean value. Can the pressure really be jumping 10psi every 0.25 seconds? That seems very out of place, and if that is 'normal' then I would say the sensor is useless. If the backpressure is not constant, doesn't that point to an issue with the turbine or engine flow rate somehow? Or does every backpressure log look like that (i've never used one, so not sure).
06-18-2017, 01:23 AM
#79
In order for me to buy that, I would need to see some oscillation represented by the number of exhaust valve events that occur in the given time interval. In other words, how many exhaust valve events occur in just 1 second? Certainly more than 3 or 5 events, which is how fast the chart is oscillating in 1 second.
If you figure out how many events occur per second at 6000rpm I guarantee it is more than the number of peaks on the chart. Each event should show a peak, right? since exhaust gas pressure goes up everytime an event occurs, theoretically anyways. Even just half the number of peaks as events would be "buyable" in my opinion. I spend a lot of time gathering data from extremely precise instrumentation and so this is just my cautious analytic side talking. The device which measures backpressure is likely sampling at a given rate, which is probably slower than the number of exhaust valve events. So I am not literally suggesting we could or should see that many peaks, I am only pointing out the obvious elephant in the room: exhaust events are occurring very fast, but the graph is cycling from high to low pressure rather slowly. If the graph is accurate, then the exhaust gas pressure is showing a serious problem, it should not bounce up and down like that in real life. Since I've never used one of those sensors, I have no way of knowing if "thats just how it always looks". If we can compare the data to other data that used the same exact sensor, and sample rate/computer, then it would be most valuable comparison data.
If you figure out how many events occur per second at 6000rpm I guarantee it is more than the number of peaks on the chart. Each event should show a peak, right? since exhaust gas pressure goes up everytime an event occurs, theoretically anyways. Even just half the number of peaks as events would be "buyable" in my opinion. I spend a lot of time gathering data from extremely precise instrumentation and so this is just my cautious analytic side talking. The device which measures backpressure is likely sampling at a given rate, which is probably slower than the number of exhaust valve events. So I am not literally suggesting we could or should see that many peaks, I am only pointing out the obvious elephant in the room: exhaust events are occurring very fast, but the graph is cycling from high to low pressure rather slowly. If the graph is accurate, then the exhaust gas pressure is showing a serious problem, it should not bounce up and down like that in real life. Since I've never used one of those sensors, I have no way of knowing if "thats just how it always looks". If we can compare the data to other data that used the same exact sensor, and sample rate/computer, then it would be most valuable comparison data.
06-18-2017, 03:19 AM
#80
9 Second Club
You dont even know what sensor he is using. How it is connected, what response time, how fast it is logging....etc..etc.
But an undamped reading from the exhaust will never be a smooth line.
But an undamped reading from the exhaust will never be a smooth line.