LocostUSA.com
http://www.locostusa.com/forums/

A locost EFI
http://www.locostusa.com/forums/viewtopic.php?f=53&t=17208
Page 5 of 10

Author:  BBlue [ April 29, 2015, 7:59 am ]
Post subject:  Re: A locost EFI

Marcus, once again, look to the Megasquirt world. It uses the "D" series connectors. While people instinctly distrust them, if used in a dry environment are capable of handling anything the PCB can and have proven to be quite reliable. Some people find them difficult to wire, but that seems to be the extent of the problems. They are even used to drive "dumb" coils.

I am using coil drivers mounted in a cut down Radio Shack aluminum project box, mounted on top of the computer with Velcro and wired in with "D" connectors. Thirteen thousand miles and all is well.

Bill

Author:  Lonnie-S [ April 29, 2015, 9:34 am ]
Post subject:  Re: A locost EFI

Marcus,

What I know about EFI & electronic ignition would fit in a thimble with room left over. However, when I was looking at donor vehicles, and realizing I would have to deal with both those topics at some level to get a modern donor engine, I found the Ford EDIS system is considered the easiest to adapt by many and that was part of my donor-deciding process.

I no longer have the Internet references, but there are firms in Europe that adapt the EDIS system to non-Ford engines and vehicles and also for racing applications. It seems to be quite popular there. You might want to poke around on search engines with that in mind. The one thing I do still have is this:

Ford EDIS technical overview ==> http://www.dainst.com/info/edis/edis.html

I realize you're still exploring here, but perhaps that info will be useful.

Cheers,

Author:  rx7locost [ April 29, 2015, 10:36 am ]
Post subject:  Re: A locost EFI

To minimize the EMI/noise sensitivity to your CPU, I'd suggest keeping the coil drivers outside of the main box. Fast, high voltage spikes caused by switching inductive loads are not your friend. They can confuse your digital circuits as well as corrupting the analog signals coming into the CPU. If you keep the coil signals as just that, signals, then a user can use a coil with integral driver or a separate dedicated driver board.

Funny thing about noise and grounding, you never know where it's ugly head will pop up. If you have it solved for JD's installation, that doesn't mean that it will be OK with a 2nd or the "n"th installation. Or that it will be forever OK. For example, in my GMC Jimmy, a bad ground wire connection at the radiator support panel ( about 2 ft long ground wire) can make my ABS malfunction.

Author:  horizenjob [ April 29, 2015, 1:15 pm ]
Post subject:  Re: A locost EFI

Thanks all, every bit of this is helpful. One of the things I need is just dimple stuff about people's impressions of what makes sense and what gives people a quality felling etc.

I'll look some more a the D-sub connectors. THey are not usually water tight, thought there are versions with O-rings or seals. They get real expensive though - like more than $50 for just the connector, and you would need 2... SO the really expensive stuff is out. I'm wondering if it doesn't make sense to just have the wiring harness come out of the box with no connector at the box. It could have 6' or wire on the leads and you could put the desired connectors on them after cutting to length.

Another issue is that many of the automotive connectors are pretty bulky. I'd like this ECU to be about the size of a pack of cigarettes, probably a bit bigger - but not much.

There are two issues pushing me for putting the coil transistors in the ECU. On the plus side I think it might save some real money. The transistors would cost about $1 more a piece, but the coils with the internal ignitors seems to cost $10 or more a piece and also don't come on the wasted spark units. Am I right on this?

The second plus for having the drivers in the ECU is that the IC I want to use to drive these transistors offers extra features. It can measure the spark quality and duration and also check for open or short circuits. When there are problems it can reduce the power to protect the coil and the electronics.

I have some ideas about how to get away with this and keep the noise away from the the processor and analog to digital converters. So we can talk about that. At the moment it looks like the processor and it's friends will be on one tiny board and the coil circuits on another board. Since they would be stacked it makes the unit much smaller and I can try different things for the coil drivers. The coil drivers board would have it's own ground wires and they will come out of the opposite end of the case.

Author:  gavin_eakins [ April 29, 2015, 6:35 pm ]
Post subject:  Re: A locost EFI

What driver is that, Marcus, that can detect those faults?

Cheers - Gavin.

Author:  horizenjob [ April 29, 2015, 7:26 pm ]
Post subject:  Re: A locost EFI

Hi Gavin,
I'm using mostly Freescale parts. I'm torn between dumb drivers and smarter parts. I think the first driver board will use the Freescale 33810. It offers 4 fuel injection drivers and 4 Coil pre-drivers on one chip, so there will be 2 or 3 of them on the board.

http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=MC33810

Author:  GonzoRacer [ April 29, 2015, 7:59 pm ]
Post subject:  Re: A locost EFI

Quote:
I'm torn between dumb drivers and smarter parts.

HEYYYY!!!! Is THAT why you wanted ME to test the prototype????

:rofl:

JDK

Author:  russian2 [ April 29, 2015, 11:45 pm ]
Post subject:  Re: A locost EFI

horizenjob wrote:
I want to use to drive these transistors offers extra features.

Is it just me, or is the scope getting wider and wider?

Just in terms of project management: the software for this ECU has not started one single engine yet, is it really time to select the best ECU connector?...

Author:  horizenjob [ April 30, 2015, 1:54 am ]
Post subject:  Re: A locost EFI

Quote:
Is it just me, or is the scope getting wider and wider?


Well I guess it's my nature to try to understand the whole job of building this ECU. I did not understand a good wiring harness would cost more than the ECU hardware. I sort of thought this from looking at advertisements of some commercial units, but I am getting an education in this area.

I could have a running ECU and no idea how to connect it realistically to anything or what kind of a box to put it in etc. Asking people's opinion here is helping me though.

The first prototype boards may not even need a connector, but it might not hurt to put a foot print down for one and then solder the wires to it. I'm back to my schematic and board layout now...

Author:  russian2 [ April 30, 2015, 7:23 am ]
Post subject:  Re: A locost EFI

horizenjob wrote:
I'm back to my schematic and board layout now...


It is my understanding that your highest risk is with the software. If I would be you, I would pause board layout but focus on getting a test mule ready and run some tests with some temporary hardware.

Can you spin your test engine with a starter (from the hardware prospective)? Do you have the hardware to feed the cranking RPM signal into the EFI?

Author:  TooBusy [ April 30, 2015, 8:05 am ]
Post subject:  Re: A locost EFI

russian2 wrote:
horizenjob wrote:
I'm back to my schematic and board layout now...


It is my understanding that your highest risk is with the software. If I would be you, I would pause board layout but focus on getting a test mule ready and run some tests with some temporary hardware.

Can you spin your test engine with a starter (from the hardware prospective)? Do you have the hardware to feed the cranking RPM signal into the EFI?


Would this be a good point to talk about hooking up Marcus' project to something like a Jim Stim? I hate to keep going back to MegaSquirt world, but doesn't the JimStim pretend to be an engine? Or is it easier in the real world to deal with real world signals?

I'm not an electronics guy.

Author:  rx7locost [ April 30, 2015, 11:12 am ]
Post subject:  Re: A locost EFI

Quote:
Would this be a good point to talk about hooking up Marcus' project to something like a Jim Stim?


I am a firm believer in crawl, stand, walk, run development. The Jim-Stim and an adhoc system would be great for proof of concept. Maybe use a real VR input Aka the RX-7 or Miata instead of the JimStim for those circuits. Full functionality of firmware is not necessary at this point, but a full knowledge of what options you intend to have available would be ideal. There will always be feature-creep. Plan for that somehow. Then I would go right to PCB development. That would be the time to determine connectors which may end up as a 2x"x" dual row header that can be wired to the connector to the outside world to start.

I know a lot of other engineers who go from thought to PCB, skipping the breadboard stage. This is costly and adds time to the development IMO.

Thinking for just a few minutes, if I were to develop a new DIY controller I would plan for......

8 analog inputs: 2 VR/HE, dedicated 12V sense input (not the same as the 12V power);Wideband input, coolant temp, oil pressure(for fuel pump shut off control?), fuel pressure, others?.
8 injector outputs (full sequential) switched ground
8 coil outputs (COP) switched ground
8 assignable switched outputs (fan, tach, variable cam, variable intake volume, EGR valve, purge valve etc) switched ground
4 PWM outputs (IAC, Boost, Cam, fuel pump regulation? etc)
4 analog outputs, alternator
Fuel Pump relay output
Many ground wires, ideally 1 for each Input and "n" for outputs)
Communications lines (USB, Bluetooth, Wi-Fi, CAN)

Probably others too. This is the quick list. Anything less would really limit the marketability IMO. Look at the MS3 capability and don't offer any less. At least with a CAN output, you could expand...... Try to offer these features so that no item requires changing components on the PCB's. And that the use of any 1 feature does not eliminate any the use of other feature, at lease in hardware. Analog inputs using various brand OEM sensors without changing board level components might be a challenge.

Author:  horizenjob [ April 30, 2015, 11:53 am ]
Post subject:  Re: A locost EFI

Quote:
It is my understanding that your highest risk is with the software. If I would be you, I would pause board layout but focus on getting a test mule ready and run some tests with some temporary hardware.


What you're saying is sensible, there are just a bunch of little issues that make it more difficult for me to go that way. It's only in the last week that the weather has become good enough to clean out room in my garage to work on the actual motor. Since everything takes twice as long as I imagine, now the board is impacting the schedule of doing something practical and immediate.

For me, I have been programming computers for 40 years. I wish it was less and I was younger, but there you have it! :rofl: So I worry less about the code aspect, I know it will need work. Things like CAN drivers in addition to just getting an engine to idle.

Doing hardware has lead times. Stuff has to be ordered and delivered, it has fabricated and assembled. THere will be time then to work more on mechanical and breadboarding. My fear is that after I get a breadboard system running, I will still have to go thru those other delays including the real risk it doesn't work and needs to be debugged. So I want to debug with something like the product so that it gets as much use as possible. Eat your own dogfood....

Author:  horizenjob [ April 30, 2015, 12:02 pm ]
Post subject:  Re: A locost EFI

Chuck that's a lot of good input. I was worrying about the issue in your last sentence before I went to bed last night at 4AM. I will match your list up with where I am now and see what it looks like. So I'll write you a longer answer later today.

I think some of my issues go away if I sell this as a board or a complete system with box and wires. Then I can be less worried about the issues of connector cost/quality and harness cost/quality. I just can't help but be attracted to the good stuff in that area. I like handling parts that I know work well, with good fit and feel. The electronics have less price difference between really good and not so good, so they will be really good - at least that's the plan.

Author:  horizenjob [ April 30, 2015, 7:34 pm ]
Post subject:  Re: A locost EFI

Here's a very early picture of the work to build a board. This is what's called the "rat's nest". In this picture the board is the black work space and has a 5mm grid on it for scale. The reddish boxes represent the pages of the schematic and their parts. I have more work to do but this is a check that is helping me visulaize things. THe driver parts are going onthe left side and the processor and analog etc. are going on the right. After things are looking good, I will probably break up the project into two boards.

The second picture is part of the Slotus team test hardware. The chip drivers should be able to identify a good coil, a bad coil and a hot dog. Further experimentation will let us know if we can tell when the hot dog is done. I am happy to report that my junior scientist was able to measure a resistance of between 1 and 5 megaOhms, all while performing the sound effects of Curly, everyone favorite Stooge. Honestly, I am so proud of her...

Attachments:
o5e_ECU_PCB_scratch1.jpg
o5e_ECU_PCB_scratch1.jpg [ 395.92 KiB | Viewed 2663 times ]
Slotus_Test_Fixture1.jpg
Slotus_Test_Fixture1.jpg [ 191.02 KiB | Viewed 2663 times ]

Page 5 of 10 All times are UTC - 5 hours [ DST ]
Powered by phpBB® Forum Software © phpBB Group
https://www.phpbb.com/