rusefi.com

Thanks to Andrey for all the support so far, I'm here and I've signed up.


Year 1994
Engine 1839
Transmission Manual
Colour 50 Shades of Red.

My first plan is to get it up and running on a relatively standard spec MX5 (Well Eunos Roadster). My car has a K&N panel filter, aftermarket exhaust and the timing mod done. Once this is running, I am hoping to port the ECU and my newly found knowledge across to the MK2 (UK spec) that I have here.

Welcome and thank you trying this out

I think I have a stock US 94 1.8 ECU in my bin of stock ECUs, I'll confirm later tonight.

Cheers. I can't wait until it arrives and I can start to play!

Can you take some pictures the harness plugs where they go into your stock ECU when you have a moment? Looks like 1990 is different from 1994 which is different from 1996. No rush with that.

I am pretty sure that I will gather all the info online, just want to be uber-sure which one we are dealing with.

Sure, I'm in the other car today, but I'll try and do it tonight when I get in.

EDIT

Is there any specific pins? Colours you want me to look out for?

deviator wrote:Is there any specific pins? Colours you want me to look out for?

Just an overview pic would do it - I mean I definitely want to see some colors, but that's mostly just to confirm that I am looking at more or less the right diagram.

As requested....




Hope that helps. I've ordered the wideband+controller today and I'm getting excited.

Perfect, that's exactly the pics I was looking for.

deviator wrote:I've ordered the wideband+controller today and I'm getting excited.

What WO2 did you get? I can mail a pin wire end to add into one of the unused positions of your connector.

Your case has arrived, in terms of components I now have everything but one diode. I am planning to ship your board in about a week.

Nice one. My wideband is here, so not long now.

One of the pending issues is how does the MAF sensor provide the signal to the ECU.

On my Ford Aspire with a MAF which is my MAF experience so far, the sensor was providing current (not voltage) and it had to be pulled down with a 1K to be measured, one would measure 1K resistance between ECU pins 2B and 3A - see for pinout.

On the 1994 Mazda ECU I have, I measure 2.25K resistance between 2O and 3A

I've uploaded some pictures of the 1994 Miata ECU here.
MAF input is pin 2O, I can trace it to a via below T822 silkscreen, on the back this goes to a 10K R432 in the middle. Also there is a short trace to a 1.5K R431 which is connected via that wide track to the middle pin of the green regulator?! WAT? Anyway, multimeter shows 2.15K resistance between 2O and logic GND, probably because of D005 which has 615 Ohm resistance.Anyway, the place of EEs is on a bonfire alongside the witches.


2O <- > R471 200K pull-down

2O <-> R472 2K <-> D471 (trace on the back) & D472 (trace on the front) & P842 & R473 18.3K
R473 <->C472 & main chip corner pin

D471 to logic GND
D472 to +5v

See also http://rusefi.com/images/1994_Miata_ecu/IMAG0263_front.jpg http://rusefi.com/images/1994_Miata_ecu/IMAG0266_back.jpg http://rusefi.com/images/1994_Miata_ecu/

See also http://www.ehow.com/how_7816537_test-mass-air-flow-sensor.html

A mechanical co-worker once told me "if he can't see it, it doesn't exist". I think he'd offer some gas to get the bonfire going for the EE's and witches.

Thanks for the pictures. I'll see what I can learn. I'm opening them in GIMP now.

Some info on MAF from the same guy who has helped us with the tach signal requirment for fuel control: http://forum.miata.net/vb/showpost.php?p=7103798&postcount=8

Goes into the mail tomorrow. Best case scenario it would run your engine right away, but it definitely needs tuning and very conservative testings. This is a prototype, please treat it as such.

russian wrote:

LOVE IT! I can't wait.

I've realized there is a mistake on the board.



In the bottom-left corner there is a piece of wire connecting one of the P602 pins with one of the P604 pins. This wire should be cut or unsoldered.
Once the wire is removed, the better option is to solder a piece of wire to connect the two P604 pads together. An easier option which might work would be to only cut the existing wire and apply some configuration changes. Please let me know when you will have time to play with all this and I will help you on the phone.

I was a bit confused by this, but now I see it. What russian is saying, is that the board has a wire installed between P602 and P604, this wire sets the hi/lo driver to use a 5V high side reference. However it needs to be 12V. So you will need to relocate the jumper to be on the P604. Once the jumper is across the P604 via's, the high low will use the proper 12V signal and your ignition should work.

{edit} corrected type{/edit}

F1001 is nice. don't you have proper smd devices? %-)

The proper place for the fuse is upstream from the PCB Putting a fuse there may or may not be a good idea. Fuses can be very difficult to properly size, especially when integrated into a PCB.

F1001 has gotten mixed reviews. Some have pushed that it's good because you can guarantee there is a fuse and you can protect the PCB, others have pointed out that people might see this and think it's good for the circuit, AKA no other fuse. As well it violates KISS, and could be a source of failure. Technically you should have the fuse as close to the battery as reasonably possible. I used to put in a thin trace such that you knew were it would burn if you had a burn. That thinner trace would be located between some fuse holder pads, such that if you burnt the trace it would be easy to replace with a fuse. However that takes a bunch of space, so I tend to leave it up to the the assembler to decide if the fuse will be up stream, or if they want extra protection on the PCB. It seems to be about 50/50 on what people want.

I tend to avoid the SMT fuses, as the heat from other components or heat sinking copper can warp the tripping points from the expected. Picture running a race in the desert with elevated ambient temperatures. It would suck to pop a fuse just because it got warm. Remember fuses are a small resistor that opens because it got to hot. See http://datasheet.octopart.com/0429007.WRML-Littelfuse-datasheet-7554058.pdf for reference. I design for an ambient of up to 50C, however that often allows the PCB components to get to 100C+ such that the chips stay below 125C to 150C. If you mount the fuse to the PCB the fuse temp is closer to 100C. If you put the fuse on some short wires keeping it slightly spaced away from the PCB, you will be much closer to 50C. Per that datasheet, they stop the de-rating at 80C. So the difference of SMT vs wire mounted means with-in components specs instead of exceeding the spec-ed limits.

Sorry it's kind of long winded just to say you may or many not want a fuse. In this case russian guessed there would be a proper up-steam fuse.

Right after a slack period, I am back and have time to play. Wideband is in the car and working. ECU is now jumper to P604 (12v).

I'm going to start downloading the software now.

Welcome back

Realistically you should be fine to warm up the engine, hope for the best, and try to start the engine. It should start and it should not ruin your car. At least I really hope so

If you want to play it more conservatively, you can turn the block on without cranking the car, check the gauges, make sure that you hear the fuel pump, try the 'bench' commands to click the injectors & try spark, and then try to start it.

Feel free to PM me or @ your phone number if you want to chat on the phone.

Unfortunately I have to re-iterate that no warranty is implied or anything

Right time for an update......

ECU swapped. Powers up and looks well. Feeling brave I turn the key








Enough drama, it's cranking but not firing. On the plus side, I suspect I know the issue. I've not heard the fuel pump priming and I have this 'thing' in my mind about the US and JDM fuel pump wire being different? I am not sure on this and I've dug no deeper as it's cold and dark out there!

Pictures of the display after cranking, incase anything jumps out....



On another positive note, reconnected standard ECU and the car took me to the shops fine.

deviator wrote:

Is that with ignition key ON? Can you take a picture with ignition key on?

Because V0.0 says Vbattery = 0 (!?) and "M9.3" says MAF 9.3 volts and both is totally unrealistic. I wonder if you wiring is completely different from http://www.rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_1994 by any chance?

Another explanation would be if the config got wiped out. Can you connect the board to a 'dev console' java application? You would need to take it off the case for that. I think you would need dev console anyway so you would need access to the micro USB connector so you would need to take it off the case for now.

Yes, that is with the ignition on. Without the ignition on, the ECU isn't powered. If I get chance I'll connect it up tonight.

Board removed from casing, I don't have the jumper lead with me as I'm not at home. However I've setup the tools on this laptop as well and it appears the Dev Console connects and reports a firmware. The title bar states.....
"Console 20141102;firmware=20140907@4951 Miata 1994@COM5"

So it looks like it might be a wiring issue.

The Maff value is still reporting 9.2 or 9.3 even when sat on the desk.

Realistically what we need is to allocate an hour or two where both you and me would be online and you would have a phone and a laptop. Oh, and the car and a multimeter

On the other hand the board does power up so it means some wires are right. I think once we get on the phone we will figure it out. Let me PM you my cell. Can you please read the manual about updating the firmware and set_engine_type 20 http://rusefi.com/wiki/index.php?title=Manual:Engine_Type

And the firmware appears to be updated....

New title bar "firmware=20141101@5040 Miata 1994d@COM3"

The display now reads "unrealistic CLT -285". I presume this is coolant temp?

deviator wrote:The display now reads "unrealistic CLT -285". I presume this is coolant temp?

Yes, that's an unhappy http://rusefi.com/wiki/index.php?title=Engine:acronym-phrases#CLT

Can you please post what tempinfo and analoginfo commands output into the Message tab of the console?