Computerized ignition (DIY ECU)

[Guest]

Hey guys. Just recently picked up an old gs650 has some mechanical and aesthetic problems that I'm currently learning about and working on. There were also some issues with the ignitor and R/R. The R/R I plan on simply replacing(let me know if you know a good place) but for both cost and fun I plan on replacing the whole ignition system with one I plan to code and construct. I also plan to use this system to control other systems(lights, turn signals, etc). Solution for the sake of time, I plan on using an arduino along with some fairly simple perf board circuits. Anyone have any tips or tricks before I get really going on the project?

[posplayr]

Draw a 1 wire diagram and a schematic along with your software design document

[dantodd]

For the R/R get an SH775 or a compufire. The oem were shunt based systems and killed many a stator. Also, check your stator health before getting too far along as a bad/weak stator can give you the fits.

[roeme]

Not sure if an arduino programmed above assembler is fast enough for an ignition.

but for both cost and fun I plan on replacing the whole ignition system with one I plan to code and construct.

I don't think it'll be cheaper than a points ignition. More fun, possibly.

Re. R/R; as said, definitively get a non-shunt. Sooner or later you have to anyway.

[posplayr]

Not sure if an arduino programmed above assembler is fast enough for an ignition.



I don't think it'll be cheaper than a points ignition. More fun, possibly.

The UNO/ATMega328 using floating point is pretty slow, despite the relatively high clock rate (due to software floating point emulation). Regardless with some care and using a interrupt based Timer Input and PWM output and fixed point calculations it is very doable to do the Ignition timing assuming he does not figure out a way to chew up a lot of clock cycles controlling lights.

The bigger issue is making sure he gets the driver circuits correct with sufficient voltage range tolerance to not now the blow the drivers from the inductive loads. Then again if this is just an experiment, then he can get something that is very robust albeit not very small (GM HEI modules) and be up and running with little fuss on analog electronics design.

[Tesla]

It is definitely more for the fun of it then anything. I have a lot of experience with electronics and coding and would to put them to a fun use. The ignition will be interrupt based. Hopefully everything else fits in between interrupt calls. If not I have some faster arm controllers I could use, they just take more time to fully set up. Thanks for the tip on the R/R and stator, I will definitely give those a look. As for the analog part, there will definitely be some experimenting to not blow everything up when turning on those coils. I guess we'll see how it goes! Thanks for the input

[posplayr]

It is definitely more for the fun of it then anything. I have a lot of experience with electronics and coding and would to put them to a fun use. The ignition will be interrupt based. Hopefully everything else fits in between interrupt calls. If not I have some faster arm controllers I could use, they just take more time to fully set up. Thanks for the tip on the R/R and stator, I will definitely give those a look. As for the analog part, there will definitely be some experimenting to not blow everything up when turning on those coils. I guess we'll see how it goes! Thanks for the input

Not that you need it , but I was at a seminar on STM32Cube two weeks ago. They are making it Soooooooooooooooo eazy now.

Figure on at least 250V ratings for the drivers.


EDIT:
After reading this thread you might think twice but, going bare metal is definitely more painful than having something to start with.

http://www.eevblog.com/forum/microcontrollers/st's-(stm32cube)-software-ecosystem-is-terrible-how-can-we-fix-it/

[JTGS850GL]

The Nano would have all the UNO capabilities in a much more compact design. Agreed that the ATMega328 should be fast enough to support ignition timing needs as long as the code is kept compact. No need for floating point math that I can see. A simple lookup table for timing curve should be all that's needed. There are two interrupts that can be utilized.

[posplayr]

The Nano would have all the UNO capabilities in a much more compact design. Agreed that the ATMega328 should be fast enough to support ignition timing needs as long as the code is kept compact. No need for floating point math that I can see. A simple lookup table for timing curve should be all that's needed. There are two interrupts that can be utilized.

Yea I was thinking Nano not Uno or there are some smaller but OP says he is doing perfboard.

[JTGS850GL]

I think he's doing perfboard for the peripheral components and using the pre-canned micro controller to drive it. Kind of similar to prototyping projects I've done in the past.

I like the Nano since it has all of the hardware capability of the UNO but so much smaller. Many of the smaller micro controllers don't have all the IO or interrupts that the Nano has. Adafruit also has the Trinket PRO in both 3.3V and 5V versions but are a little priceyer. I just picked up 10 of the Nano's from China for under $2 each on a "Make Offer". May take a month to get them but it's worth it if you're not in a hurry.

[dantodd]

I like the NANO too but if it's your first time with the Chinese cheap ones you will need a special driver. For the Macs it can be difficult to set up.

[JonnyC]

What year is your 650? Do you have a mechanical advance? I was having some ignition issues, or so I thought. The test on my ignitor didn't match up with how the manual said it should function, so I just replaced it with a Dyna DS3-2 since I still had the mechanical advance. For $133, it was a nobrainer for me to eliminate possible headaches.

And yes, SH775 R/R with a wiring kit from Eastern Beaver (or maybe I got the connectors off of Ebay, I can't remember).

I started playing around with building a tachometer using a Nano. Pretty basic, but I only tested it using a signal generator as input. Ended up buying a Trail Tech so I said screw it and scrapped the project.

[posplayr]

I like the NANO too but if it's your first time with the Chinese cheap ones you will need a special driver. For the Macs it can be difficult to set up.

USB hell............ I have some RS485 to USB convertors I bought from ebay and they seem to work except........ They drop a bunch of bytes at 115K baud and the FTDI chip is apparently grey market as it is not programmable using the FTDI utilities. I'm going to probably buy some replacement FTDI parts and just swap them out. FT232RL. The Ebay nanos I bought a while back seemed to connect with the IDE, but you have to be careful with it; just because it works at 9600 doesn't mean it will work at 115K and above.

With real FTDI chips and latest drivers I have gotten up to the limits (clock limits) of my ARM processor at 230K baud under Win7. It will probably go all the way to 960K (windows drivers) but I have not been able to push them that far.

I have had a lot of problems with Prolific driver mainly running on grey market USB devices. If you want to run above 9600 without headaches, but authentic USB parts/products.

[JTGS850GL]

Unusual I don't go above 9600 since I typically only use the USB port for sketch uploads and debugging slower events. With that said, I've had no issues with the Chinese grey market units do date. No special driver needed for the ones I've purchased. All have worked fine with the Arduino IDE. The only thing I've noticed is the USB port number assigned is different on the Chinese units. I don't use a MAC so can't say on that topic. I will say that I've been in the habit of using the same vendors for components. Once I've purchased from them and all works I will reuse them for future purchases.

[posplayr]

Unusual I don't go above 9600 since I typically only use the USB port for sketch uploads and debugging slower events. With that said, I've had no issues with the Chinese grey market units do date. No special driver needed for the ones I've purchased. All have worked fine with the Arduino IDE. The only thing I've noticed is the USB port number assigned is different on the Chinese units. I don't use a MAC so can't say on that topic. I will say that I've been in the habit of using the same vendors for components. Once I've purchased from them and all works I will reuse them for future purchases.

I have been consolidating all low speed serial communications in my UAV system including RS232/USB/RS485 and bluetooth into a full C++ stack mirroring the OSI standard. I have been able to push data at close to 95% baud utilization of a master slave- cmd response link over USB at 230 baud and would surely work at double that if I would get the ARM processor there.

9600 is so slow that virtually anything works. Depending on what you have the stuff will start to fail at 38K or have other really funny "features". USB three stuff runs at much faster speeds so I'm taking primarily about Virtual Comm port USB.