Ned

So the ECU has 6 outputs good enough to do fancy timing for spark or injection. 2 injection and 2 spark and 2 that go to low side switches. Those low side switches can be used to drive a coil if need be, but need some extra hardware (just like a microsquirt) and they may not actually work very well for old style spark? they may have to be logic input like COP or newer LS coils but not sure, havent looked into it that much. The board is being made, i expect that to get shipped today/tomorrow so hope to have it early next week. The parts have already been shipped, so they will arrive monday first thing new motor has been found and am picking that up tomorrow afternoon, and the car is already at Cams where we'll be utilizing his new garage toys this weekend for engine swap Then next week will be make and test board and then do wiring etc weekend after... gonna be fun!

Has 2 ignition and 2 fuel with 2 more fuel outputs that need extra drive circuitry to use it. Its really made for 4 cyl batch injection and wasted spark but can do sequential injection if you really want to

cheers guys, and no, you just take a punt and hope you're up to the task really. you CAN probably model bits of the circuit in some weird program, but thats only part of it as the layout often has a lot to do with it all and you cant really. Plus i've never used the modeling software so wouldnt even know where to begin haha. you just design a schematic (kinda like a wiring diagram) and then you do the PCB layout and route all the tracks etc. The software will tell you if your PCB matches the schematic and can perform a design rule check, where you can setup things like "i want my tracks to be no closer than 0.15mm from each other" and it will run through and tell you if that all passed.

Hey, guys, i did stuff! //oldschool.co.nz/index.php?/topic/36476-neds-7k-drop-hatch-starlet/?p=1609820

wow... been a while! Used the car as my daily for 6 months earlier in the year and went fine. Now it's time to actually EFI it though... before drag day (i hope). So i figured, the best way to do that is the hardest way, and thats by designing my own ECU... All i have to design is the hardware, i'll be using FreeEMS to actually run it all, but their current hardware leaves a lot to be desired in my eyes. so let me introduce MicroEMS! a play on words between MicroSquirt and FreeEMS because i made it use the same connector and pinout as the microsquirt, so that if FreeEMS software or my hardware all turns to custard, i can just buy a microsquirt and plug it in a 3D render from the board design program; So i have actually just ordered 10 boards, and am trying to get my digikey account un-frozen so i can order the components for it as well and start building it in a week and a bits time I've been wanting to design an ECU for a LONG time and never really pushed through, but over the last few weeks i've put a lot of hours into designing something and going to finally give it a go. I shall keep you all updated

so the funny thing about china, is that they get mega mega good pricing because they make a zillion of 1 product. Meet digikey. This is where a very large portion of hobbyists and professionals get their parts from. Once you get into big volume manufacturing, you would go directly to the manufacturer of each part, but if you have a product that you're making a few hundred or even thousand of, you would buy your stuff from digikey. and this is the digikey link to the micro on that board https://www.digikey.co.nz/product-detail/en/ATMEGA2560-16AU/ATMEGA2560-16AU-ND/735455 spoiler, it's $30 NZ and you can buy a whole arduino for $12 china will sell you the bare chip as well though, $25US for 5 of them, so still $8NZ each http://www.aliexpress.com/item/Free-Shipping-5pcs-lot-ATMEGA2560-16AU-ATMEGA2560-MEGA2560-TQFP-100-IC/1927760663.html?spm=2114.01020208.3.31.jZFqGP&ws_ab_test=searchweb201556_2_71_72_73_74_75,searchweb201527_4,searchweb201560_9 It's crazy at how cheap they churn those out. If you were to make them in NZ (or any other place in the world that isnt china) then just the cost of putting the components on the PCB would be more than what china sells the whole thing for. And thats not even buying the bare PCBs or the parts for the board. Thats just the cost for a machine to place them on the PCB and solder them on. It's crazy how they can do it, it really is...

i'm sure the ones that care are actually on the speeduino forum, but the man did some actual tests on timing accuracy (on the bench) and here are the results, they are pretty good (for a low burget DIY EMS) and a lot better than most documentation on the web makes it out to be (which is where my initial info came from) http://speeduino.com/forum/viewtopic.php?f=5&t=223&p=2955#p2955

So micro is short for microcontroller, which is the processor. (sometimes can get referred to as uC to shorten it some more) Arduino has lots of different boards with different micros on it depending on how many pins you need, or what size memory etc etc. The one used for speeduino is the biggest, baddest one they have. All of them, share the same arrangement for how the connectors come out of them, and you can plug in things called shields. A shield is just some extra hardware you can simply plug in, and sometimes even stack lots of them together. So if you want to add ethernet to an arduino, you plug in the ethernet shield. If you want to control some servo motors, you add in that correct shield etc. Speeduino is essentially just an engine management shield that you plug in to give it the hardware it needs to run an engine. Nice and easy for development, not very robust for real life applications. People use arduinos for 2 reasons really. 1) arduinos have a very simplified coding overlay provided for them. So hobbyists dont need to learn an aweful lot about software and hardware because it's all standardised and simplified. You can just about say "turn on pin 1" and in the background it goes and figures out what pin is pin 1, find the correct register for that ports data direction pin, set it as an out put and then find the right register that controls the state of the pin and then set that high. It's like driving an auto vs a manual car kinda. You dont have to use the simplified software, so people like me will use an arduino as a set of building blocks. I will get an arduino, find a shield with some bit of kit i want to use in a project and then i can just plug it in and evaluate my design, get testing etc and if i like it, i can go and build some dedicated hardware where everything is just on one board. This is what speeduino is kinda doing, but they arent making the leap to dedicated hardware yet and i'm not 100% sure why they havent yet. They might have a good reason i havent found yet.

Whats wrong with this? https://www.autosportlabs.com/product/racecapturepro/

Fred? He doesnt like tunerstudio and wants to do his own thing, but that hasnt happened yet. Tuner studio has some limitations etc and apparently isnt unlikely for it to do the wrong thing. I have no idea and dont know anything about that side. The more and more i'm thinking about it, the more and more keen i am to just redo my board for speeduino haha, but a dedicated board, none of this arduino shield stuff. I mean those are the micros i cut my teeth on at work, and know them like the back of my hand so software development would be much easier for me and i could actually tune my car if i changed haha

Now look what you've done Alex! I havent made them yet, finishing touches ATM and sending off this weekend so will have them in a week or 2 probably. In time to run the KP for drag day i hope and i've jumped on the forum, i had a look around at the schematics and they seem to be very stongly influenced by the FreeEMS ones. A lot of similarities there, which is cool. Means that other than the micro, there isnt much different hardware wise between my board and theirs because mine is also strongly influenced by FreeEMS (because it has to be) so if my thing works, it would be easy to spin up a slightly different board with the same interfacing hardware, just a different micro/brain so that i can run his code, as it seems more complete than FreeEMS...

im sad i only just found out about this

It's not as exciting as it looks in fairness. I need an ECU for the starlet, and with the exchange rate, buying a microsquirt is too spendy, so for the price of getting just the loom shipped to my door, i can build a whole ECU. It's based on the FreeEMS stuff Fred wrote/is writing but is pin compatible with microsquirt, so if i ever get sick of it or if my design is rubbish, i can just plug a microsquirt in and use that haha. So same as microsquirt, 2 fuel, 2 spark, 4 spare low side drives. Some analogue inputs, 2 VR/HAL inputs. Has USB on board too, which is the only real difference functionality wise. FreeEMS is definitely not ready for this to be a thing lots of people use. It currently has no good way to tune it, which is the main down side of it all but i'm happy to make do with whats there (for now anyway) for me its as much an exercise in designing the thing as it is having an end product so even if its shit, i'm still happy.

I love the idea, for sure, dont get me wrong and it sure does have its up sides, and i see it started as a fun little project for small engines etc etc and it just grew bigger into what it is, and i appreciate that. It's a cool thing for wanting to tinker and thats what this thread is all about, but they could have build something better for similar money. In saying that, i'm opting with putting a toyota 7K in my car when for less money i can buy a 4AGE that makes 3 x the power so i guess i'm all about polishing that turd until it shines like a diamond in a goats ass!

It does look super cool, keen to hear how you get on as I do like the project

sorry, i'll elaborate, 100uS is not enough for 10 000 RMP though. It will run, but it's not good enough (if you ask me) because 10K RPM = 166 2/3 RMS. 166 2/3 x 0.0001S = 0.0166667 rotations per loop. 0.01666667 * 360 degrees = 6 degrees timing accuracy. personally i think it should be accurate to at least 1 degree, but would really expect 0.5 deg. Not to mention the fact that the micro on there is not meant for automotive use, neither are any of the other parts, and parts start behaving REALLY weird when you have them in a car with elevated temperatures... plus that micro is just not the micro of choice. You can get a better micro, for less money. (like a 32 bit Cortex M0 as a minimum, potentially with a floating point unit for super fast math). It's a cool exercise in "hmm, i wonder if i can make this atmega run a car?". I've been wanting to do it myself actually, to see what the smallest micro is i could use to run a car, but thats all it is, a fun experiment. I wouldnt run an actual car on it that i wanted to be able to get in and drive every day without question. but thats just me. Most of the code looks quite cool, and i sure do like the project and hope they improve upon it because i'm pretty sure they can improve upon those loop times.

this is pretty rubbish though and is nice as an academic thing if you just wanna learn about what happens and play around with a shitty engine on a stand, but shit if you actually want to run your car on it. Just buy a microsquirt for ~twice the money

what if i deliver 2 more hands tonight and tomorrow night?

[click me]

Fantastic work man. Love it!

I think this should be a thing in the next 5 weeks... I'm near Vegas for a month, and keen to have a meet given I'll miss 2 of the Auckland ones while I'm here

Missing this one, and probably the next. See you all in November probably

yeah. Building shit is expensive

Or buy the finished one on trademe

A lot of people were talking about going at burger boes. Should be a good one