XGATE PIT based bit bang fuel injection

Official FreeEMS vanilla firmware development, the heart and soul of the system!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: XGATE PIT based bit bang fuel injection

Post by Fred »

What are the two different traces?

You should perform your testing at a higher frequency so the scope updates more or less in real time too. There could be artifacts being hidden by the slow scope response.

Looks promising, though :-)

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
SleepyKeys
LQFP144 - On Top Of The Game
Posts: 549
Joined: Mon Feb 11, 2008 10:52 pm
Location: Arizona
Contact:

Re: XGATE PIT based bit bang fuel injection

Post by SleepyKeys »

Since you insist on the xgate bit bang code supporting an inferior batch fire mode I will make two versions of it. One that is slower and supports what you want to do and the other which is as lean as possible to provide the best accuracy. Maybe I can put a special flag in the event so it knows when it can treat it in a simpler/faster manor and just keep one version.

I'm not opposed or against doing this, but it really needs to be tested as it is now. It should be 100% for doing sequential injection/ignition as it is now. After it's verified and changes made if necessary I can feel good about moving on to adding this feature. Everyone knows it unwise to make changes without first knowing your base point.

Sound fair?
You snooze, you lose!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: XGATE PIT based bit bang fuel injection

Post by Fred »

Totally fair, link your test plan thread, when done, and I'll review and we can put it through its paces and decide where to go from here.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
SleepyKeys
LQFP144 - On Top Of The Game
Posts: 549
Joined: Mon Feb 11, 2008 10:52 pm
Location: Arizona
Contact:

Re: XGATE PIT based bit bang fuel injection

Post by SleepyKeys »

You know what. I could easily build two loops. Ign which has a higher priority and simpler bang code(much like it is now) and I can add your *overlap feature to the fuel loop and the small time penalty wont matter to fueling..... It seems we could both win in one shot. Thoughts ???
You snooze, you lose!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: XGATE PIT based bit bang fuel injection

Post by Fred »

Thoughts: Don't fuck with what you have now, except to tidy it up/finish it. Just work on your test "scripts" and specs and binutils and loader and rail and and and.

Wildy changing tack, again, before you get the first version really working and tested is just a plain shit idea.

Having two algorithms running at the same time sounds like a horrible dirty hack, too. It IS possible to do this right with one algorithm. Maybe you have it now, maybe you don't. It needs to see peer review, testing (unit, functional, integration & user), and miles, FIRST.
  1. Make it work (integrated, as a real solution, not hacked in)
  2. Make it right (clean, reliable, stable, repeatable)
  3. Make it fast (enough for ignition use)
In that order!

Fred.

PS, but after the loader is refactored and tested thoroughly, and binutils is packaged for xyz platforms and the fw is switched to using it.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
SleepyKeys
LQFP144 - On Top Of The Game
Posts: 549
Joined: Mon Feb 11, 2008 10:52 pm
Location: Arizona
Contact:

Re: XGATE PIT based bit bang fuel injection

Post by SleepyKeys »

Just look at how your attitude has changed............. There was once a time when I used to look forward to conversing with you about problems and soloutions. Most of what you said, goes without saying.

Your code is full of stuff like this:
/** @todo TODO move sched code to a function or functions (inline?)
* that can be unit tested such that we KNOW it performs as anticipated
* rather than just trying it out on a 400hp turbo truck engine.
*/
Yet you give me miles of shit for tossing some ideas out there. I belive you have a hard-on for trying to make me look bad, any way you can..... Why thought I ask myself....... The only thing I can come up with is you despise seeing anyone progress the Firmware, but yourself.... Maybe they are correct, this is "Freds project" "Just work on your test "scripts" and specs and binutils and loader and rail and and and"

"
Make it work (integrated, as a real solution, not hacked in)
Make it right (clean, reliable, stable, repeatable)
Make it fast (enough for ignition use)
"
The only one that is certain right now is *integration, wether it is fast enough depends on the application. "clean, reliable, stable, repeatable", well testing takes time. I ran it on my car and did some elementary testing on the bench, so I suppose that puts it on par with some of the sections of your firmware.
You snooze, you lose!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: XGATE PIT based bit bang fuel injection

Post by Fred »

https://www.youtube.com/watch?v=Xbw_BxDwdjk

Temper temper.
Sean0 wrote:Just look at how your attitude has changed...
Nope, not one little bit. I've always wanted what I detail below.
Your code is full of stuff like this:
Yes, I document the limitations and weaknesses and incrementally improve it. As opposed to leaving it looking pristine but not actually being.
/** @todo TODO move sched code to a function or functions (inline?)
* that can be unit tested such that we KNOW it performs as anticipated
* rather than just trying it out on a 400hp turbo truck engine.
That code has run all but the first FreeEMS car (which didn't have scheduling at all), it's been tested with thousands of miles of driving, as well as thousands of hours on the scope and LA. Write a suite of unit tests for the loader, then we'll discuss unit testing.
Yet you give me miles of shit for tossing some ideas out there.
No, not for that, for working on the "wrong" stuff. Honest, and rhetorical, question: Are you proud of the the state of the loader, as it sits? I know that you're proud of the binutils stuff, as you should be!
I belive you have a hard-on for trying to make me look bad, any way you can.....
LOL, not at all, you're my mate, or so I thought, at least. I'm TRYING to guide you into being in a position to look very good. I'm certainly enthusiastic about maintaining the reputation of this project as one of absolute high quality, though. If you want to be in the lime light, you've gotta do things right. Absorb that. You've got too much on your plate to be screwing around with XG stuff right now IMO. You don't need it to run your rail. You can run that rail with existing XG code and OEM or ECT ignition as you intended. You're a free man, though, so do as you please.
Why thought I ask myself....... The only thing I can come up with is you despise seeing anyone progress the Firmware, but yourself....
Absolute rubbish. I despise the idea of stuff that is sub-par ending up in the firmware. That's all. Peter recently had his work included, I was happy, ask him.
Maybe they are correct, this is "Freds project" "Just work on your test "scripts" and specs and binutils and loader and rail and and and"
Sounds like somebody has been hanging out in #preludelinux and listening to them badmouthing me. I particularly like the MTX commits with "Fr3dEMS" in them LOL. I wish the loader was stable and fast on the three required OSes, but it's not. I wish it was simple to use, powerful when needed and delivered excellent intuitive feedback to its users, but it isn't and doesn't. If you had started it 4 months ago like Mike with EMStudio, I'd cut you some slack on it, but you didn't, it's been 2+ years with little change in stability and feature set. If you step up to a task, be sure to finish it before getting side tracked with other stuff. Admittedly, in the last few weeks you've put a lot of time into it, which is GREAT, and I was really happy about it, and I told you so. The above, and comments in IRC overnight, were disappointing due to my expectation that you'd see it through to completion (to a solid state, not final completion) as you recently told me that you would.

I had this same conversation with Dan yesterday re switching power supplies.

I'm focused on a goal, not the fun in getting there. Stuff that doesn't align with that goal (solid firmware and tools for it) is noise to me, and when it steps in front of previous responsibilities (chosen responsibilities!!) then it's annoying. Dan's attitude when I brought it up was excellent. I can show you the log, if you wish.
testing takes time.
testing takes time and large amounts of organisation and attention to detail.
Fixed.
I ran it on my car and did some elementary testing on the bench, so I suppose that puts it on par with some of the sections of the firmware.
ROFL, that's the funniest thing you've ever written. Fix in italics.

It's not part of the firmware until it's part of the firmware. I very much look forward to executing the git command that some day merges this work into the master, but that's quite a way off yet. You know damn well that you'll have my absolute support and undivided attention as soon as the more important stuff (both mine and yours) is out of the way. We'll be working as a team on 0.4.0, you'll be leading the charge. Make it happen, the right way.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: XGATE PIT based bit bang fuel injection

Post by Fred »

^ last significant typing for today. Sorry everyone else.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
SleepyKeys
LQFP144 - On Top Of The Game
Posts: 549
Joined: Mon Feb 11, 2008 10:52 pm
Location: Arizona
Contact:

Re: XGATE PIT based bit bang fuel injection

Post by SleepyKeys »

By attitude, I meant more specifically your demenior.

Your project attitude and *standards are great. Hard to keep up with at times, but thats how you progress IMO.

"Yes, I document the limitations and weaknesses and incrementally improve it. As opposed to leaving it looking pristine but not actually being."

Which is the right thing to do, but also remember I don't go around saying "hey use this, its all ready". I flick videos every now and then so other people know work has been done and will resume. IMO this deturs users from thinking that it's a very far off feature. And to be 100% honest a way to toot my horn... But again remember I don't go as far as to push it on people. The *noise I create with this is very minimal.

"Sounds like somebody has been hanging out in #preludelinux and listening to them badmouthing me. I particularly like the MTX commits with "Fr3dEMS" in them LOL. "

That's the first I have heard of "Fr3dEMS"(on second though you may have send me a link with that in it), again these thoughts stem from your deminor, not repeating what others say.

Deminor sends a message, wether I am being over sensivie or not is anther matter, but it does make a difference and people formulate ideas from such *feedback.

"The above, and comments in IRC overnight, were disappointing due to my expectation that you'd see it through to completion (to a solid state, not final completion) as you recently told me that you would."

We are taking about a few week pause with loader progress. I told users if the loader is preventing them from loading, let me know and I WILL drop what I am doing. In a sense what I was saying was "Fred, dont get pissed when you see some firmware commits come though reguarding my R1 or Xgate". Where did I say I was walking away? There are several issues that have been asigned to you for resolotution verification. I know you are busy, how often do I give you a hard time about that?

"ROFL, that's the funniest thing you've ever written. Fix in italics."

Don't think I meant for that to be insulting.

"You know damn well that you'll have my absolute support and undivided attention as soon as the..."

Yes you have said that before, and I belive you. I'm not asking for that now, I'm just asking that I don't get heaps of shit with the exception of me or my work being a road block IE "The loader flat out wont load for me". In which case I would expect the worst.

IMO a *better response would have been: "That may or may not be a good idea, I dont have the time to really think about it. The fastest way to getting your R1 running would be to use what you have already done or xxx..... Not to worry we will get to it".
You snooze, you lose!
User avatar
SleepyKeys
LQFP144 - On Top Of The Game
Posts: 549
Joined: Mon Feb 11, 2008 10:52 pm
Location: Arizona
Contact:

Re: XGATE PIT based bit bang fuel injection

Post by SleepyKeys »

It seems I was able to move the xgate output scheduler into a more appropriate section, of the firmware. This was the result of sending the benchtest firmware the test.sequential.bin file.

Image

Which outputs are xgate and which ones are ECT........... :)
Last edited by SleepyKeys on Wed Oct 17, 2012 1:48 pm, edited 1 time in total.
You snooze, you lose!
Post Reply