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.
You might need to remove the compressor housing to get that hose end on there. Turbosmart (and others I bet) make an extension that moves the hose end away from the CHRA.
I would also mock this up a bit, you might run out of vertical space for the oil pan side return. You don't want that drain going in the wrong direction.
Would it be OK to set up the EBC in open loop mode?
I have the 35A-AAA-DDBA-1BA MAC valve.
It's just that I do not have sufficient grasp of the PID controls required for closed loop operation, and I thought it would be better to not take a risk with an untested setup.
I want to start with as little boost as possible that will operate the system.
Yeah, you can run EBC open loop. Most installs have the valve default (powered but not actuated) set to provide wastegate pressure. If yours is this way, simply set it all to 0 and you are good to go.
read the descruption for the "Algorithm" setting in the boost control settings window.
So I am developing a nested-loop pressure control system for my compound-hybrid. Like Jason mentions somewhere up there, I decided to go with 2 pressure sensors for each actuator. The flow through traditional solenoids is a function of the square root of the pressure differential across the average solenoid orafice area... so hard to make linear. This is not a trivial development, but if you decide to pursue the arduino route, start no lower than the Teensy 3.2.
Oh ****, I actually have something to contribute. If you know what you're doing, you can use the fast root from Doom. It should be close enough for this, and could be probably programmed into most ecus. There are more precise algorithms, but if speed is a concern, and 2-3% error is acceptable, that'd be rad as ****. I have lots of experience getting low level things to do hard math, so I got really stoked when I read this haha
Oh ****, I actually have something to contribute. If you know what you're doing, you can use the fast root from Doom. It should be close enough for this, and could be probably programmed into most ecus. There are more precise algorithms, but if speed is a concern, and 2-3% error is acceptable, that'd be rad as ****. I have lots of experience getting low level things to do hard math, so I got really stoked when I read this haha
can't tell you if I know what I am doing or not Speed was a concern on the atmega 328, which is why I moved to the arm core at 90MHz. Now speed isn't an issue, that being said, I will run my linear regression models with the fast root algorithm to see what kinf of performance I can get because, well, just because.
can't tell you if I know what I am doing or not Speed was a concern on the atmega 328, which is why I moved to the arm core at 90MHz. Now speed isn't an issue, that being said, I will run my linear regression models with the fast root algorithm to see what kinf of performance I can get because, well, just because.
thanks for the pointer! \end jack
If you're using an iterative solver, this could be a good first guess way to shave off the most expensive steps. If you need help with that part let me know, that was my favorite part of school. God I love math.
Would it be possible to tune EBC using an electric air pump without starting the engine?
I am trying to sort the boost control plumbing/wiring/tuning before I break in a brand new engine.
Also, if I can pull this off with a pump, I can play with and really learn the boost control system without risking anything, and that is a huge advantage.
My plan was to seal the throttle body and feed pressure to the intake manifold in small increments, monitor pressure and wastegate action, and keep playing with values on TS till I get boost numbers right.
I realize the dynamics of a running engine is far more complex and, well, dynamic, but wouldn't it at least give me a baseline idea for things like correct polarity, which way I should go with DC numbers and safe max boost levels?
Hmmm, not sure I understand your proposed method. The IM won't hold pressure...
you can characterize the solenoid by measuring the inlet pressure and the outlet pressure at various duty cycles (when connected to your WG with a representative length of hose). The WG spring is fairly linear with pressure. So on an ideal system, a 50% DC will provide 50% of the inlet pressure to the WG, and the WG will regulate about the 2x spring value. The thing that makes this messy is that 50% commanded DC is not close to 50% actual. I posted a bunch of stuff on this in my build thread.
run any tests with a solenoid voltage equal to the alternator regulated voltage, typically 14V.
turbosmart has a sensor cap which allows you to actually characterize valve displacement on some of their WGs. I have used one, it works well.
Ultimately your boost controller is controlling exhaust gasses. So without said gasses, you’re simply learning the range of valve and actuator. Which is helpful but still a long way from “boost control”
I decided to hook the EBC up as follows, with the exception of boost reference source.
I installed a nipple on the IC to TB elbow and routed the hose to the MAC valve.
The plan is simple:
Uncheck (not use) the boost control option on TS for a few days and have the system run at WG pressure.
There is a 7 pound spring in the WG as a precaution, just so I do not risk anything till I sort things out.
Once I am comfortable, I will tune the EBC to 14 or 16 psi.
First open loop to learn it properly, and after that, closed loop adventures shall begin.
Call me an idiot, but I just can't get the additional programmable outputs to work.
I was told I could use any spare ignition output for that purpose.
My understanding is those are pins 3 and 4 on the bottom connector. This is because I had used them to convert my ignition to sequential per Reverant's instructions in the past. (I reverted back to wasted spark since.)
I devised a simple plan to test programmable outputs: Active over 10% TPS. That way, I can test them without starting the engine, and keep checking the pins till I find the right one(s). This method works on the existing 2 programmable outputs available on my MS by default. (Pins 6 and 18 on the top connector)
I am using a LED test light and a voltmeter to see if there is any change in the outputs as I press the gas pedal while I check the pins. Nothing.
I enabled all of the outputs and checked every pin on both connectors.
There was one scary issue: If I enable the output port settings, I get major conflicts in the system.
But, enabling these get me nothing:
The problem is, I need 2 more programmable outputs to make my system work.
I have been working on this turbo project for 10 months now, and I really do not want to give up at this point.
Why aren't you using ignition outputs? They are the only high side drivers that will give you output voltage without a pull up.
plus you can test using IO test mode
It's 6:48 am here, and I have been at this since about 5 pm yesterday, so I am not all that together right now.
I was under the impression the ignition outputs were for used to run the coils, and that there are 2 more spare ignition outputs (if you want to run sequential) and those can be used as programmable outputs if you need them.
Which ignition outputs are you referring to please?
Also, I checked every single pin, to no avail. What am I missing here?
You can use spare ign outputs, but they do not appear in the screens you posted unless Rev has renamed them. I am not familiar enough with his hardware to know what resources you have available.
what do you need the output for? Maybe some of what you have available is open drain.
You can use spare ign outputs, but they do not appear in the screens you posted unless Rev has renamed them. I am not familiar enough with his hardware to know what resources you have available.
what do you need the output for? Maybe some of what you have available is open drain.
I need the additional outputs for vacuum solenoids to run the turbos properly.
07-26-2018, 10:00 AM
0
Speed was a concern on the atmega 328, which is why I moved to the arm core at 90MHz. Now speed isn't an issue, that being said, I will run my linear regression models with the fast root algorithm to see what kinf of performance I can get because, well, just because.
