1994 MX5/Miata/Eunos 1.8 NA (Japanese Import)

[deviator]

Is there a way to operate the fuel pump test with less run time? The reason I ask is I'm thinking of sticking a bottle in the engine bay and running the test. Obviously 3secs may well fill a bottle unless I pipe it out to a large Gerry can.

I was hoping there was a de-pressurise valve on the injector rail as with some other cars, but no such luck.

[AndreyB]

Is there a way to operate the fuel pump test with less run time?

Not right now.

But I think it's not worth taking the fuel hose off, no reason to torture yourself. I am really in favor of shortening the wire to GND in order to hear the relay if not the pump, and then doing the same with the software. Unless it's really noisy there both sounds are pretty obvious.

[deviator]

Updated. Now on Console 20141114; firmware=20141115@5394 MX594d@COM4

[deviator]

OK what do I need to short? I want to be 100% certain on this (for obvious reasons)

[AndreyB]

You want to ground fuel pump relay wire for a second.

If you have an unplugged harness plug, it's light green wire in position 2T - it's the fourth wire from the end of the larger plug - see http://www.rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_1994

If you have Frankenso board, it's probably easier to touch the ground side of E4 transistor - all the pins on the left side on this picture:

E4 is the leftmost transistor in the cluster of eight transistors.

As for the source of ground, I usually use the pin on the inside of the door hinge. There is also a cluster of GND on Frankenso - that spot where a naked wire connects three or five pads together, between the unused smallest plug and the largest plug.

[deviator]

Progress! The pump and relay fired straight into life. Looks like one issue is about to be solved.

[AndreyB]

Progress! The pump and relay fired straight into life. Looks like one issue is about to be solved.

So we know it's the right wire. Now need to figure out if or why not Frankenso is/is not controlling this.

[deviator]

Is it worth me metering out 2T to GND to see what it does while issuing the pump test? Or do you have a better idea?

[deviator]

Bad news I'm afraid. ECU connected, ignition on. Ran fuel pump test 3 times, no noise, nothing on meter. Then there was an ascending high pitch whine/whistle followed by a blank display on the Frankenso. This time only micro usb connected so laptop stayed up. Pulled plugs, disconnected battery.

Reinstalled factory ECU, car still works fine, well it starts and runs.

Brought everything inside. Disconnected Discovery and Frankenso. Connected to laptop. Only 2 lights. Computer doesn't recognise USB device. Then that familiar smell, pull USBs from laptop.

The STM32F103CAT6 is warm. RIP Discovery #2.

[AndreyB]

I am getting a batch of new boards next week and at some point in December I hope to get a bunch of these assembled by the skilled Chinese hands, I will mail you a replacement Frankenso hopefully within 6 weeks from now.

In the meantime we can try to figure this out on a bench if you have time to participate. Do you have any 12V power supply? I need to go offline for a couple of hours, let me talk to Jared and we will come up with a research plan if you - please - please - would agree to help.

I'm sorry

[deviator]

The story so far.... (The purpose of this is to pull together everything that happened in 7 pages into one single post)
1994 Japanese Eunos 1.8 imported into the UK.

Stock ECU Pictures
http://i754.photobucket.com/albums/xx182/the_deviator/MX5/IMG_20140827_180953_zpsvvik5cmd.jpg
http://i754.photobucket.com/albums/xx182/the_deviator/MX5/IMG_20140827_181039_zpsflklcnb1.jpg
http://i754.photobucket.com/albums/xx182/the_deviator/MX5/IMG_20140827_181110_zpszdjsed4y.jpg

Frankenso/Discovery board pre-Shipping pictures (from russian)
http://rusefi.com/images/Frankenso/frankenso_set_for_miata_1994.jpg
http://rusefi.com/images/Frankenso/frankenso_set_for_miata_1994_back.jpg


AEM Wideband kit ordered (Bosch Lambda/O2) and installed.

5v/12v link wire moved P602/P604 to P604 only.
http://i.imgur.com/zM493tR.png

Once arrived, flashed to firmware 4591. Connected to car. Try to fire. No luck, nothing from fuel pump. Display on Frankenso shows....
http://i754.photobucket.com/albums/xx182/the_deviator/MX5/IMG_20141103_192510_zps42e32861.jpg
http://i754.photobucket.com/albums/xx182/the_deviator/MX5/IMG_20141103_192516_zps817041ce.jpg

No issues other than failure to report.

As russian said “Because V0.0 says Vbattery = 0 (!?) and "M9.3" says MAF 9.3 volts and both is totally unrealistic”

A new firmware 5040 and now the Frankenso reports "unrealistic CLT -285".

Next a Skype session with russian. We had just updated the firmware and he was Team Viewing my PC checking things. When a burning smell appeared, laptop powered off. Car & Laptop were fine. Both USB leads were connected. Since then russian has advised against the mini USB lead being connected whilst the ECU is in the car.

The board had been powered up for around 3-5 minutes. Reset after the firmware update and up maybe 1-3 minutes before the failure.

This fault was the U1 rectifier....
http://i.imgur.com/tQJTe3G.png

kb1gtt asked me to test the duff board. Results were D1 is measuring 248 ohms in one direction and no connection in the other. D3 is measuring 10 ohms is both directions.

Two more Discovery boards ordered. This time the VDD pins were bend and broken to remove the chance of a short to the Frankenso.
http://i754.photobucket.com/albums/xx182/the_deviator/RusEFI/IMG_20141113_151914_zpsbafzren0.jpg (Please note, all 3 VDD pins were trims, NOT just two as in this picture)

The Frankenso board alone was connected to the car, ignition on. Each pin on the Discovery header was checked and nothing above 5v was found anywhere.

Everything ready for another round of tests.
Fuel pump test – nothing
Injector test – Fine, can hear/feel these opening
Fan Test – Nothing
Spark Test – No spark found.
Later this spark issue is believed to be solved (but I've yet to test) by russian
“Yesterday another 1994 was started once we've realized the coil wires were flipped in rusEfi configuration - see http://rusefi.com/forum/viewtopic.php?f=3&t=729&start=10#p12354 “

Report a whine from the idle solanoid.
set_idle_pin none
writeconfig

Suggested by russian solved this noise.

I compare my wiring harness to http://rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_1994 I map my colours out (looking into the back of the harness plug)
http://i754.photobucket.com/albums/xx182/the_deviator/RusEFI/BPF3Pinout_zpsd26a6437.jpg

This leads to me finding 3 different coloured wires to what I was expecting (bold in the above picture).

These are for...
2F MAP Signal Ground
1O Stop light switch
1E Check engine light

So probably not much of an issue at the moment.

Firmware updated to include the above mentioned swap coil pack wires over. Now 5394.

With both boards connected to car and ignition on, a bridge wire was connected between 2T on the harness plug (Fuel Pump Relay) and GND. This brought the relay and pump into audibly working.

The board was probably only powered for 1-2 minutes.

Next a connect a diode test multimeter between the harness plug 2T and earth. I connect the laptop to ECU. I issue 3 fuel pump tests with no noise what so ever. Then there is an ascending high pitch whistle/whine. The display on the Frankenso goes out. The laptop is still on, I pull the USB, (Only micro connected) turn the ignition off and isolate the battery.

The car has been returned to the stock ECU and appears to work, so does the laptop.

The Discovery board is disconnected from the Frankenso and connected to the laptop. I get 2 normal red lights but no flashing blue. I get 'USB Device Not Recognised' from Windows and a now familiar smell appears. I pull the USB plugs. The underside of the STM32F103CAT6 chip is warm. Another Discovery is toast.


Several things I've noticed which may or may not be relevant...
1. The Frankenso display has never displayed anything other than 0.0v
2. The Frankenso display has displayed out of range MAF voltage, 8-9v is common. I've had 9.2 whilst sat on the bench.
3. When both boards are connected to the car for more than 2-4 minutes. The Discovery board fails. It's like a component is failing, or something unexpected is going hot in the car? Although as noted below, it must have powered for 10 minutes whilst I was spark testing the other night.
4. The fuel pump output is not working.
5. The spark has yet to be confirmed.

Some of these may well be red herrings.

On a side note, the other night I was working outside with the ECU connected for more than 10 minutes whilst I was running Spark Tests I had no issue with the board. Thinking out loud, could multiple fuel pump tests generate a short on the Frankenso?

[puff]

first of all my condolences.
just fuel pump test, without actually cranking?
could it be the fuel pump relay switches on something else, which drives some unexpected voltage to the board?
has it ever shown any realistic adc input values?
are you totally sure on where your efi connector wires go to/come from?

[kb1gtt]

deviator I sent you a PM with skype info. I believe it has gone un-read.