Ghetto sequential injection for 99/00 Miatas

[Tara]

I have a design for taking in a 5v injection pulse, timing that pulse and repeating it once for each trigger event. It differs from the MS sequencer design in that they use a microprocessor and time and repeat the pulse in SW, where I'm using 2 Xilinx XC95108 CPLDs (16bit counters) and a 1Mhz crystal. Also, my version is specific to my car model. Still, I'm using the same MOSFETs to drive the injectors as Microsquirt and the same 5v supply as the v3 PCB.

So, I would really like some help with the design but I'm a little afraid of getting sued. Do you think this is close enough to the B&G design to be a problem?

Thanks,

[Fred]

It doesn't sound remotely close to me. It sounds like an add on module for your car that could interface with either OEM ecu or MS or anything else.

I feel the need to present one downside for you though :

A big part of the advantage of sequential in my eyes is the frequency of updates of the PW information. You lose this entirely by sampling like that as you only get one pulse per engine cycle to measure.

In a true sequential system, each pulse starts offset from the rest with potentially fresh and updated PW based on the changing conditions during the cycle. What this has the largest effect on is transient "enrichment" or rather accuracy of injection during a transient.

It might be possible to fake that with a digital signal that is polled once per event, but not by sampling an existing signal unless it is prescaled down and rescaled up to do cylcount squirts per cycle and updating it's own PW as often as possible inside that.

Why don't you send them an email with your plans and ask if they think it infringes on their rights at all? That could be a good way to gauge if you will get sued or not. Currently they have no standalone product that behaves as you are talking about and one integrated product that kinda behaves like that, but kinda doesn't.

You will still gain the superior atomisation properties doing what you are, and if your goal is large injectors on a small engine, or better fuel economy, you will probably achieve that. Just not the throttle response aspect.

Best of luck with it anyway :-)

Fred.

[Tara]

Thanks for the honest critique. I knew it wasn't ideal but I'm looking for something to help smooth my idle. I'd also eventually like to go to stagged injection with an ITB/FI setup I'm working on. That will drop me down to only one injection driver and that would kill my idle. This is a way around that.

Primarily, though, this was a way for me to add CPLD/FPGA programming to my resume.

Thanks

[Fred]

Thanks for the honest critique. I knew it wasn't ideal but I'm looking for something to help smooth my idle. I'd also eventually like to go to stagged injection with an ITB/FI setup I'm working on. That will drop me down to only one injection driver and that would kill my idle. This is a way around that.

It should do as you wish if you make it work right :-)

Primarily, though, this was a way for me to add CPLD/FPGA programming to my resume.

That's secondary for me, and a good excuse when the women in charge gives me grief for not getting a job ;-)

Fred.

[ababkin]

Thanks for the honest critique. I knew it wasn't ideal but I'm looking for something to help smooth my idle. I'd also eventually like to go to stagged injection with an ITB/FI setup I'm working on. That will drop me down to only one injection driver and that would kill my idle. This is a way around that.

Primarily, though, this was a way for me to add CPLD/FPGA programming to my resume.

Thanks

Hello and welcome to the forum Tara
I am very happy we have another person who is into programmable logic and desire to apply it for controlling an engine (the first person is myself but i've only recently started with that, so still very much a noob and still learning VHDL)

As Fred mentioned, this kind of operation will result in worse response to transients than a true sequential system, but i think you will still get better idle just because of the fact that you'll be opening injectors only once per engine cycle (as opposed to several times as in banked/batched systems)

If you really worry about B&G frowning at you legally, just re-implement your power supply according to some free/open design (i am sure there are better ones around). I wouldn't worry about using same mosfet as MS does - this is in no way infringing. Anyone can use any parts. The IP is 'HOW' you use them.

Alex

[Tara]

I guess if the power supply and MOSFETs are the issue then it's best to just leave them out.

Here is the design for the first chip:


Here is the design for the second:


I couldn't keep everything on one chip even with 12bit counters so I'm stuck soldering the 16 I/Os from one chip to the other.

The Xilinx USB programmer is about $250. My husband read that the $40 USB cable that Digilant sells could be used but you have to use their software. Has anyone used either one? I know the parallel cable is less but the laptop we're using doesn't have a parallel port.

http://www.digilentinc.com/Products/Det ... av2=Cables

[Fred]

BTW, the suing/patent/copyright stuff is only applicable if you decide to produce, market and sell these things. If not, you can pretty much do as you please :-)

Congratulations also on being our first female member! (that I know of) Most impressive :-)

Fred.

[ababkin]

Congratulations also on being our first female member! (that I know of) Most impressive

INDEED!

programmable logic + car stuff + female = your husband is very very very lucky guy!



hit up some cpld/fpga forums, people love the digilent stuff - chances are their cable will fit the application.

[Fred]

Tara, I ran across this today and thought it may interest you. I believe you should be pretty safe as they only have a patent on the basic MS infrastructure so far and it's laughable anyway.

http://www.diyefi.org/forum/viewtopic.php?f=9&t=304

Fred.