gemellocattivo.com

Which means "Evil Twin". Lets see your projects where you change boring into fun or create the fun from scratch.
It is currently Thu Mar 28, 2024 5:58 pm

All times are UTC - 5 hours [ DST ]




Post new topic Reply to topic  [ 13 posts ]  Go to page 1, 2  Next
Author Message
PostPosted: Thu Jan 24, 2019 12:45 pm 
Offline

Joined: Wed Jan 23, 2019 11:13 am
Posts: 27
All right, I don't expect hand-holding or someone to do this for me, unless it's already in there. I'm early on in digging into software features so I don't know that it's not.

I would like, in an ideal world, a way to tweak various tables during operation by another program. I'm not trying to come up with a whole new version of EL Console - rather, one of my ideal use cases is to have an engine dyno running LabView (I'm rolling my own dyno control) be able to talk to the ECU in order to have an engine under test and automatically rough out a tune.

I can see a couple ways of doing that (I have the thought of some CAN functions to do so for example) but if the communication protocol is documented or open enough that would be even easier.


Top
 Profile Send private message  
 
PostPosted: Thu Jan 24, 2019 1:18 pm 
Offline

Joined: Thu Jan 01, 2015 6:47 pm
Posts: 4251
That's going to be a Jim question....and he's going to say "no problem, just connect the dyno to the EL HW and let the ecu run both so there is no communication issue to think about!"


Top
 Profile Send private message  
 
PostPosted: Thu Jan 24, 2019 1:40 pm 
Offline

Joined: Wed Jan 23, 2019 11:13 am
Posts: 27
Ha! I have a thing about bugging professionals before they have my money but I may try sending him an email.

More seriously, for most purposes I probably could run the dyno on EL hardware... but one of the things I want to do is both do combustion pressure sensing and then tune from it. The hardware I have will do over a million samples per second analog input which is probably beyond the hardware on the EL box.

I accidentally ended up with a lot of irons in the fire so none of this is going as fast as I'd like. Hopefully wife's having a kid in a few months - those don't take time away from building stuff, right? :lol:


Top
 Profile Send private message  
 
PostPosted: Thu Jan 24, 2019 2:02 pm 
Offline

Joined: Thu Jan 01, 2015 6:47 pm
Posts: 4251
mekilljoydammit wrote:

I accidentally ended up with a lot of irons in the fire so none of this is going as fast as I'd like. Hopefully wife's having a kid in a few months - those don't take time away from building stuff, right? :lol:


See you about 10 years ;)

Congrats.


Top
 Profile Send private message  
 
PostPosted: Thu Jan 24, 2019 3:13 pm 
Offline

Joined: Wed Jan 23, 2019 11:13 am
Posts: 27
Not just that kind of irons in the fire - I already mentioned the engine dyno, but really it's all in support of road racing. I'm getting close to satisfied with the chassis design for an SCCA GT-3 car or, for those of you not in the know, basically a tube chassis silhouette car (RX-7 in my case) with around 280-300hp.

I might go do a thread in the chassis section reposting some of my progress because it's kinda fun. Between that and the home metal casting... well, the dyno probably has top dollar priority amongst car projects.


Top
 Profile Send private message  
 
PostPosted: Thu Jan 24, 2019 3:38 pm 
Offline

Joined: Thu Jan 01, 2015 6:47 pm
Posts: 4251
Threads about cool projects are always good :)

do not underestimate the time a child requires.....I certainly did.

On the dyno auto tuning the car....man....I've never liked any kind of auto turning so I'm not sure that would be a very happy marriage. I always like to make the run, then take a few minutes to review all the logged data then make choices about what to change. I KNOW it could be automated.....I've just never seen that work well because getting the last little bit is pretty hard....getting the car and dyno bother working well and working well together just feels like a 10X difficulty factor to both the car and dyno project....maybe I'm wrong.....


Top
 Profile Send private message  
 
PostPosted: Thu Jan 24, 2019 4:02 pm 
Offline

Joined: Wed Jan 23, 2019 11:13 am
Posts: 27
I'm not really underestimating time a kid takes... I'm prepared to scale back to "I don't get any work done on hobby stuff"; if more gets done than "none" that's nice. Maybe they'll be interested maybe not. In the mean time, figuring out this stuff keeps me sane.

As for the "autotune"... it's a bit different when it's an engine dyno you can hold things at any load/MAP point indefinitely vs trying to go back and figure out what you saw on datalogs. The way I see it, for steady state running you have a VE table and an AFR table to determine desired fuel quantity injected, and a spark advance table for ignition timing. You can directly measure everything to determine what the VE table is, cell by cell, and just be done with that part for the most part. Spark advance should be about whatever it takes to get to peak pressure at a given crank angle (ballpark 15 ATDC but may vary) and with some methods of measuring cylinder pressure you can also see when you're sneaking up on detonation / preignition. So in theory once you measure the approximate VE table at some arbitrary AFR, you can have the algorithm play with target AFR in each cell, while having a control loop adjust spark and watch for det issues. The "best" target AFR being whatever produces the most torque at that MAP/RPM cell... or most torque with least fuel, or lowest misfire rate, or whatever. Doing that sort of thing should, eventually, with enough development, get close enough to quickly fine tune the last bit.

... which will then give you a map that won't run right on the road because there's no transient tuning whatsoever! :lol: The trick though is that then you can do transient tuning alone without worrying about trying to sort out what's transient effects and what are steady state.

Of course, it's all theory, but I don't see why it's impossible. That may just mean I don't know enough, but I have done similar stuff professionally for emissions development of small engines for generators.


Top
 Profile Send private message  
 
PostPosted: Thu Jan 24, 2019 4:34 pm 
Offline

Joined: Thu Jan 01, 2015 6:47 pm
Posts: 4251
Yeah, that's how tuning works and how most aftermarket ECUs are setup to be tuned, my model included. The only "well...." is we usually look directly at torque and set the spark to max toque, adjust fuel, reset spark, reset fuel, until nothing helps then move to the next rpm point. No need to guess at what point you might want max cylinder pressure when you can measure what works directly. That's WOT. at less than WOT it get a bit more tricky....the whole point of closing the throttle is to NOT make max power so normally you want to transition form "all the fuel it can handle!!!" to "how much do I really need?" as there's no point is wasting...this is where its hard to automate. my thoughts anyway, your result may vary :)

There is a pretty fancy accel/deccel function in my model, with an enable/disable control. tune steady state then tune transitions as you say. I've not tested that code on an engine...it seems right in the simulation mode....but we'll see. Same for start up enrichment (prime). The cooland enrighment is tested and working right as are the gobal and cylinder trim functions.


Top
 Profile Send private message  
 
PostPosted: Thu Jan 24, 2019 7:33 pm 
Offline

Joined: Wed Jan 23, 2019 11:13 am
Posts: 27
Well, on one hand... eventually I can figure out what drives best for the light/medium load portions of the map and get an algorithm to do that to new engines. On the other... the big thing I'm thinking of coming in handy is for rapidly doing WOT tuning. Like, imagine having a pile of different parts on test, and knowing that the dyno controller can do a pretty decent WOT tune for each combination before doing a power pull. Header primary lengths for example - set up a dyno header with slip joints and quickly iterate on length knowing that VE/etc changes from different harmonics will get tuned for without much/any user intervention. There speed in getting a pretty decent tune would come in handy - I've seen a bunch of attempts at doing A/B tests where things were quite possibly affected by the tune not being adjusted at all between.

From what I've seen, on the relatively simple stuff I'm doing the model you have come up with so far seems feature complete on the fuel/ignition side assuming it all works as intended (no offense intended, just I know about theory vs practice) but I'll keep digging. One of the things I want to do eventually is not only incorporate antilag functions but use calculated airflow from speed/density vs MAF measured airflow to figure out how much air is being bypassed around the engine into the exhaust tract and have closed loop turbo control based on shaft speed - IE really weird stuff completely irrelevant to most people's purposes. That really needs a test engine I'm not afraid to destroy though.


Top
 Profile Send private message  
 
PostPosted: Fri Jan 25, 2019 8:13 am 
Offline

Joined: Thu Jan 01, 2015 6:47 pm
Posts: 4251
mekilljoydammit wrote:
I've seen a bunch of attempts at doing A/B tests where things were quite possibly affected by the tune not being adjusted at all between.

Make sure the tests are A/B/A at least. A/B shows you changed something, A/B/A shows you are in control of the change :)

mekilljoydammit wrote:
From what I've seen, on the relatively simple stuff I'm doing the model you have come up with so far seems feature complete on the fuel/ignition side assuming it all works as intended (no offense intended, just I know about theory vs practice) but I'll keep digging. One of the things I want to do eventually is not only incorporate antilag functions but use calculated airflow from speed/density vs MAF measured airflow to figure out how much air is being bypassed around the engine into the exhaust tract and have closed loop turbo control based on shaft speed - IE really weird stuff completely irrelevant to most people's purposes. That really needs a test engine I'm not afraid to destroy though.


I tried to add everything I though was needed.....but so for the testing is limited so yeah....if it all works is exactly right. I did look at other ECUs and do my best to back out what must be going on to require the inputs they're asking for...but I'm the typo king so.....testing required so a second person using and testing it would be AWESOME!

Making the measurement of bypass air seems like just a couple lines in the model....there are virtual channels in the tunner that I think can take data and display the result so you see it when you're connected ot reviewing logs but the ECU doesn't waste any time calculating it....I think that's how the virtual channels work anyway.....


Top
 Profile Send private message  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 13 posts ]  Go to page 1, 2  Next

All times are UTC - 5 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 3 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
Powered by phpBB® Forum Software © phpBB Group