Status, help offered?

[hamster]

Howdy,

I'm about to get started on putting FI on my motorcycle now that the drag racing season is almost done. As such, I have until March of next year or so to get it done or get pretty close to it.

I was going to go with MegaSquirt, but... I need to build my own board to fit and be water tight for my application, and B&G made it clear to me in an email that while I can make a board, I'm forbidden to release my schematics or gerbers, and that MS is not open source, blah blah blah. Well, that really rubs me the wrong way. I'd rather build my own from scratch at this point.

Of course, I managed to find this project, and I'm slowly reading and trying to get my head around where things are going and what needs to be done. I'm assuming there is active development and some help needed? I'm an EE and I have a complete lab of equipment here to help with, plus I'm a pretty decent C programmer. I'm more than happy to buy the devel stuff I see is required and jump into this over sending more money to MS at this point in time. My question is, how can I best be of use? Ultimately, I will need to design a board that is compact and able to mounted in a waterproof container, and I want everything I do to be open source, and as such I'd much rather help with a project like this than go off and do my own thing.

Lastly, I'm also a sysadmin, and I note that your forum sends emails as www-data@localhost, which will cause a lot of them to never get delivered as the domain part does not resolve. You should consider changing that to a real domain name in the forum software I had to explicitly allow @localhost on my mail server so I could get my account activated, hah.

[Fred]

Oh, how embarassing! I just setup the email stuff quickly when I swapped the board from nodaddy to my vps... I'll have a look at that soon!!!

Status is that I ran a car fuel only in the USA a year ago. I never published the story because I've been kinda busy in various ways. I got recruited to build a closed source ems for a yank, did that, from scratch, in 6 or so weeks of dev time with full control, timing, fuel, and even drive by wire. Currently we have good control of 6 pins, the hardware pins. We need bit bang control of many more, and that will be done with XGATE which is currently being worked on by myself and SeanK. Other than bit bang control, the main chunk missing is scheduling of events, and various solid reliable wheel decoders. These are closely related. Scheduling is more complex than what I did in the US/China, but I have a fair idea of how to approach it. Hardware wise, there is some inconsistency between different pins on reset. That needs research before we get a solid pin assignment (ie, before anyone designs/builds any boards). Tuning software needs to be brought upto and kept upto date too.

In summary, there is a lot left to do, but it's fairly close to being a solid base system.

Where can you best help? I don't know. Possibly researching those pin behaviours in your fancy lab?

As for your hardware design and keeping it open source. If you use free tools, it'll be more useful to others. And in any case, you should use github to publish your work as you go.

I'm going to check the unactivated accounts now to see if there are many not getting through that are trying... I'll email each personally if there are some.

Thank you for posting!

Fred.

[Fred]

Also, where in the world are you?

[hamster]

SLC, Utah.

I currently use the gEDA suite for all my professional schematics/boards, and publish my stuff using a subversion repo that is tied to a Trac interface for the wiki portion, but I'll play with the git stuff and see how it goes.

I'll go ahead and order a dev board and see what I can get going with it. It's been a while since I messed with HC12 stuff. I have a couple of rather ancient HC12 dev boards laying around, hah. For the most part, I usually do my stuff with Atmel AVRs, but I'm sure I can remember the HC12 stuff I've actually got a pretty spiffy build system working for the AVR stuff using a gcc cross compiler through Xcode on my Mac, with a USB ISP programmer too no less.

Looks like I have some studying to do!

[Fred]

If you have a public SVN that you're happy with, stick with it :-) The nice thing about git is that it makes it dead simple for others to grab a copy and start making commits. Not the case with SVN where you would need to give them access to your server. Hence git/mercurial >> svn for FOSS stuff like this :-)

I've started looking at the mail config, but it looks like I have no config! IE, defaults = what you saw. I hope it doesn't take me too long to configure that stuff...

Fred.

[hamster]

Well, if you own the box, what you probably need to do is set the 'my orgin' to be a domain instead of the default, ie, if you ran postfix to deliver mail, you'd want to set the line in /etc/postfix/main.cf to say 'myorgin = diyefi.org' and/or 'myhostname = server.diyefi.org'.

Basically, your script is delivering the mail as the apache user, so when that hits the mail system, it is coming from a local user (www-data) and it's appending the default domain as the sender domain...

[hamster]

Alright, devel board and jimstim thingy ordered. I actually have the basic stim board, as well as a mostly assembled v3.57 MS board and MS-II CPU now..

Now just to get the build environment setup and re-learn some HC12, wheee.

[jharvey]

Have you found freeems.org? If so have you found the wiki posted there? I believe that is fairly up-to-date.

Glad to hear you use gEDA, that should mean KICAD will be a friendly change.

Is your setup going to need ITB's?

[hamster]

Hah, I've found gschem/PCB to be pretty user friendly as it is once you get over the admittedly steep learning curve. Wonder if there is some interchange between the two, as I have a good library built up...

ITB? Individual throttle bodies? If so, yes, I have 4 of them, but one tuning step is syncing them all to have the same vacuum pull, so we can assume the signal from one is the same as another. The bigger problem is volume, as the runners are tiny and the vacuum signal all over the place. Not a huge deal, it can be damped reasonably with an orifice.

[Fred]

Basically, your script is delivering the mail as the apache user, so when that hits the mail system, it is coming from a local user (www-data) and it's appending the default domain as the sender domain...

I did some experiments and used the CLI client "mail" and set header fields etc and mailed my gmail account. I got the mails, all three, and Sender and From are fine, as I set them, but Return-Path stays set fred@localhost no matter what I say. Like my option was ignored, so I added another one to set IntendedRT and it worked. I have exim4 running, and grepped for similar options and didn't find anything. I did find something about not messing with fields from local senders (php included) and those seemed set correctly. Thoughts? As for the forum software, it seems to be set correctly. Email settings are as they always were and show admin (at) diyefi dot org

I'm unsure how to proceed next.

Good on you for buying the board, I hope we can work together smoothly!

Fred.