Potential Bulk Buy for FreeEMS first edition PCB

[AbeFM]

In general, I agree. I think
1) You should do a small run just to see how it comes out unless there's some real big savings in doing them all at once. Even so, I'd rather buy a board that will work.

2) Stepping back a bit, isn't this more of a alpha-ish-beta-ish build anyway? I would expect a second rev, with a lot of (mostly minor) changes to follow a short time after.

I would hesitate to call this first build a "product" so much as a beta, and I think people buying one for every car at this stage are making a mistake. Buy one that's mainly for testing. The design is open, so if it's that awesome, buy 16 more of them after you have one working.

Am I way off on this?

[Fred]

1) You should do a small run just to see how it comes out unless there's some real big savings in doing them all at once. Even so, I'd rather buy a board that will work. :-)

I think that can be achieved with suitable peer review, conservatism, and cautious procession of work.

2) Stepping back a bit, isn't this more of a alpha-ish-beta-ish build anyway? I would expect a second rev, with a lot of (mostly minor) changes to follow a short time after.

Yeah, probably it is, but what's the harm in trying to make it right first time? Aim for the stars, fall in the clouds ;-)

I would hesitate to call this first build a "product" so much as a beta

I agree, thought the same thing, and forgot to post. My exact thought was "wtf is it 1.0 for and not 0.0.1??" :-) Let's not mislead the public if possible. I think it's possible to edit sourceforge releases, so we could revise the versioning even now. Just a few edits in the forum to make it clear what is going on.

and I think people buying one for every car at this stage are making a mistake. Buy one that's mainly for testing.

Agreed, but I think everyone who wants to be a part of it should buy one or two if they want a bench unit or testing on two cars (perhaps different setups). Buying a stack is a bad idea, but everyone buying one is a good idea, provided the firmware is up to scratch when the order is made. If I were you, I'd be reluctant to place the order till you see me posting that I think it could run an engine properly with tuning etc. That won't be for a little while yet.

Am I way off on this?

Nope, pretty close to the mark as usual I'd say :-)

Fred.

[comrade904]

This probably isn't the place cause this seems off topic, but you said something that made me think. So theoretical you could populate a board for minimum running reqs, and then upgrade later to a much more complicated system? i.e. Batch fire at first, and then rebuild the board for seq?

If so THAT IS GENIUS. freaking awesome.

[jharvey]

That's the idea, you can stack modules together much like the PC104 platform. It's a modular design.

[AbeFM]

We should absolutely conform to the PC104 platform. It's a modular design.

Nicely put. It only took mild editing. Then again, those boards are kinds small, but I could see all injectors on one, etc... If they were longer, I have all sorts of things here at work to hold them.

[thebigmacd]

There isn't necessarily much of an advantage to PC104 form factor other than there are lots of the standard headers out there. The reason being PC104 is just a parallel data bus with address lines. Why not just use the AdaptS12 header form factor instead of pretending to be PC104 (while not being compatible at all).

[davebmw]

I really worry about using overly complex connector systems in an automotive environment.
I'm not quite sure if you guys are talking about a multi board stacking design where you add the optional boards on-top of the base CPU board or whether you just want to build in 104 ways for the system to fail.

Bear in mind automotive designs take into account the harshest environmental considerations imaginable add on top of that the vibration, the fact that at any time you and your families life could depend on your car starting first time every time and getting you out of the sticky situation where you want a twitch of the right foot to equal instant results.

There are a numerous reasons for not wanting such a risky design on my car, I can see that there are obvious benefits to such a design where you add the bits you want as and when you want.

I keep saying to myself would BMW/Mercedes/Honda/Toyota/Mazda use it?? if the answer is no then for me its totally out of the question.
Remember that automotive designers have decades of proven tried and tested design under their belts, instead of trying to re-invent the wheel with compact desktop/domestic grade technologies we should lake into consideration what this multi-billion Dollar/Pound/Euro research and experience has yielded.

Sorry to piss on anyones parade here but I have a contract in work with stacking board design and its a total lemon, we have more problems with it because of the design and multiple connectors than any other contract. I don't want to see this happen to FreeEMS.

Leave the citric acid in the capable hands of the food industry! it has no place in your cars guys.

[Fred]

There isn't necessarily much of an advantage to PC104 form factor other than there are lots of the standard headers out there. The reason being PC104 is just a parallel data bus with address lines. Why not just use the AdaptS12 header form factor instead of pretending to be PC104 (while not being compatible at all).

I agree!

I think Dave is 100% right on a 200,000km/mile non-tinkering OEM warrantability basis. I think reality is that we do tinker (read, keep an eye on it) and the wiring and other aspects are just as likely or more likely to fail than plugging the TA board into the main board. Plus, it has mounting holes and will be physically supported so it will then only come down to can the pins carry 5mA each reliably for 2 - 10 years at a time with lower than average mileage, and the answer I would think is a solid yes.

[jharvey]

Don't worry about fretting folks. I want freeEMS to win, not because someone said so, but because it fits the bill. Your right the connectors may not make it to 100,000 miles (or Km) but they allow flexibility increasing the chances it can be used as a starting point.

I see the concern with extra connectors, however there are many connectors that work very well, and for many reasons. A small amount of dielectric grease can go a long way.

I just posted in the articles forum a nice pdf from Tyco. I think I've posted that somewhere, but I couldn't find it, so I posted it in the article section where I would tend to look for it.

I think the intended audience here isn't the OEM, but the experimental engine. These units should not be installed on a daily driver, they are intended for experiments on free standing engines, or vehicles that run on closed stages or at the track. Not to say quality isn't a very high concern.

I think the maintained version of freeEMS should have a way to allow for add-ons. I think the connector modeled after the TA card is a great way to go for this. If there are other solutions, I'd certainly like to look them over, but right now I think this is the best we can do with the constraints in hand.

I also certainly support spin offs, that remove some of the potential issues and are focused at a specific application. I just think the primary needs to be very flexible.

[AbeFM]

I've personally taken PCI-104 connectors through shock, vibration, and thermal stress tests (not to mention vacuum, radiation, and other things you'd never see in a car) which I can GUARANTEE you would destroy my OEM computer.

The connectors, by the way, aren't all that fancy or exoctic. Sure, the header on the ADAPT board is fine. Just something which is more or less "bus-like" is what I'm hoping to see. As much as we can do that, everything can have access to the signals, as long as we're smart about how much load each thing has to drive.

Certainly stacking boards is an easy way to continue to add functionality.

Getting ideas from the OEM is really smart, using what they have provided development wise is too. But if you were so happy with exactly how the OEM does things, why are you taking out that wonderful computer of theirs?