[rusEfi] '03 Miata with VVT #28

[stefanst]

Since my '99 Turboed Miata runs like a champ on the MS3X, and the '03 naturally aspirated is still running the stock ECU, it was decided to make the '03 the new test vehicle for rusefi.

Here are some issues I can think of right off the bat:

- This car will have to pass emissions, so I need to be able to swap the stock ECU in, when not on the racetrack
- NB2 (2001-2005) Miatas have a different ECU connector. Does anybody have a supplier/part number for this?
- Looks like I need to make an adapter loom, or solder a 'flying' connector directly to the board
- VVT will need to be controlled and it looks like VVT is in alpha-status at best

I do have the wiring diagrams for the car somewhere, so this shouldn't pose a major problem.

Any other thoughts./potential problems/ recommendations?

[AndreyB]

- NB2 (2001-2005) Miatas have a different ECU connector. Does anybody have a supplier/part number for this?

Does http://rusefi.com/wiki/index.php?title=Hardware:OEM_connectors#72_pin look right?

https://octopart.com/search?q=1123038 shows good availability, I've ordered a couple from Onlinecomponents.com - these guys have a $50 minimum so I can share one of mines once these show up.

[stefanst]

That's the correct connector indeed.
BUT: In order to be able to easily swap between stock ECU and rusefi I would like to have a few additional I/O, for example for the external MAP sensor, boost control and I'm sure I can think of a few more, if I really try.

So I was thinking about using the 122pin connector that has the same pattern as the three plugs on the 72 pin, but has two more plugs. It's being sold on Digikey: http://www.digikey.com/product-detail/en/8-5353015-1/A123262-ND/2310242?WT.z_cid=sp_8004_20120425_buynow&site=us&csp=WT.z_cid%3Dsp_8004_20120425_buynow%26TE_cid%3D8004-1472158567365

[AndreyB]

Now I want a VVT donor vehicle for myself as well. It was a bit of surprise to me that http://miami.craigslist.org/search/sss?sort=date&max_price=3000&min_auto_year=2001&query=miata and http://dallas.craigslist.org/search/sss?sort=date&max_price=3000&min_auto_year=2001&query=miata get 2001+ Miatas for 3K and less while NYC prices seem to be about twice that.

[stefanst]

In order to make any sensible use of rusefi with the '03, I need a WB O2 sensor.

Fortunately, I still have one sitting around unused.
But its' output is via CAN bus and nothing else, and -unfortunately- it appears that rusefi has no module for accepting sensor inputs from CAN bus yet.

So I can connect a microprocessor, reading the CAN bus data and generate an output that rusefi can understand. But what type of input will rusefi accept? Of course an analog signal will be accepted, but doing the a/d - d/a - a/d conversion seems rather silly. Will rusefi accept any digital communication as O2 sensor input?

[AndreyB]

But its' output is via CAN bus and nothing else, and -unfortunately- it appears that rusefi has no module for accepting sensor inputs from CAN bus yet.

Can you please start a topic about that WB controller? Since we have a CAN chip on the Frankenso board you have this could be doable.

See also http://rusefi.com/wiki/index.php?title=Manual:Hardware:CAN_sniffer

[kb1gtt]

Please see P401 on page 4 of the Frankenso schematic. This offers 5V, GND, CANH and CANL. It also allows an optional termination resistor.

As russian commented, please start a CAN WO2 thread for what ever model number your device is. When you do please include as much information you can about the CAN communications for that device. To date the CAN BUS communications have been minimal. I believe someone was starting an interface based on a certain standard, but I seem to recall that it's capabilities so far have only been to communicate with one very simple device, I think it was a dash board or something similar, which shows we have functional CAN bus, we just need to know the packet structure.

[AndreyB]

Got my order - connector looks pretty small in real life but seems to match eBay pictures for VVT Miatas.

[AndreyB]

What is your plan for MAP sensor? I want to buy same exact sensor maybe even from the same source and mount it in a similar fashion to have both 03 as close as possible.

[stefanst]

It's a GM style 3bar sensor from eBay. Don't remember the seller.
Similar to this:
http://m.ebay.com/itm/New-GM-STYLE-MAP-Sensor-3BAR-3-BAR-For-Electromotive-Motec-Megasquirt-With-Plug-/272262609557?hash=item3f641dba95%3Ag%3A6n8AAOSwuhhXUS8B&_trkparms=pageci%253Ac9e2b1de-b319-11e6-8194-74dbd18031fb%257Cparentrq%253A9bd86e931580a34fcc58f1a8ffecf364%257Ciid%253A2

[AndreyB]

Is there any concern about quality of these under-$20 sensors from China?

[kb1gtt]

It's from china, so yes a concern, but in reality the technology is pretty close to go or no go, so I would expect that if it fails it will likely fail really obviously. These china units are lower cost because they bypass many of the automotive reliability and safety checks. Just bypassing the burn in testing is going to greatly reduce the MFG costs. However it also means you might have a failure as you're doing the burn in procedure when you start to use it.

[AndreyB]

Where are you grabbing +5v from for map?

[stefanst]

I have a secondary connector (molex jr.) coming off the adapter board that is used to deal with signals not included in the original harness. Splitting 5V and signal GND from the adapter plate.

[AndreyB]

I like, I guess would do the same using some weather pack connectors.
Next lazy question - what is the part number for brake booster line T-adapter?

[stefanst]

No idea- Autozone "T" collection.

[AndreyB]

It's from china, so yes a concern, but in reality the technology is pretty close to go or no go, so I would expect that if it fails it will likely fail really obviously.

the plug would not lock on the sensor - the sensor clearly has a retaining tab but it's either too short or not located where it needs to be located. requesting a refund

[stefanst]

Mine had the same issue. It was off by maybe a mm or two. I used an exacto knife to trim the tab on the connector hooking up to it and all is well now.

[stefanst]

I got it starting and running well enough. The alternator control has me baffled. No matter what I do the alternator is going full-tilt. I'm suspecting a wiring screwup on my side at this point.
Also idle control seems to have 0 effect. Another screwup of mine I'm sure.
Question: Are all PWM frequencies freely configurable, or are some shared?
Tomorrow I'll pick up my timing light from my buddy who borrowed it, so I can verify base timing.

[AndreyB]

Idle control - do you have a diode somewhere?

My wiring screw-up was confusing +5v wire and TPS wire, but my harness is actually not done yet.

I'll take your tune especially once timing is confirmed

[stefanst]

No diode. But I was running the IAC at 300Hz and I believe the NB IAC needs 70Hz or so.
Didn't get around to picking up my timing light. Tomorrow....

[AndreyB]

Question: Are all PWM frequencies freely configurable, or are some shared?

All PWMs are frequencies freely configurable, nothing is shared between them.

[stefanst]

SUCCESS!!!!

Engine is running fairly smooth. Idle control works. Even alternator control works. Forgot to wire the WBO2- will do tomorrow.
No VVT control yet, but we'll try that next.
Trigger settings:

Trigger Settings 2003 Miata.png

Trigger offset of 666 is confirmed by timing light. How cool is that I'll have to paint the marks on my balancer though. They're real hard to see. So I wouldn't hard-code the timing yet without final confirmation.

Alternator settings:

alternator settings.png

Now for some weird stuff. This is WITHOUT alternator control. The voltage just randomly drops from 12.2V or so to around 11.5V. I think this is happening when we're sampling voltage at the same time an injector fires. Even though the injectors have a separate 12V supply line, ground is probably getting pulled up by 0.7V or so.

voltage drops.png

I'm also seeing some fluctuations in AFRs. May be caused by deadtime compensation being influenced by the voltage drops.

And here's a log of the alternator control working:

alternator control working.png

Also attached is the MSQ and the whole looong msl from tonight's session.

[AndreyB]

Even alternator control works.

Please elaborate - how did you wire it?

Thank you for playing with it!

[stefanst]

HL5 -> 3M on the connector. That's it. Voltage seems steady while revving.

[stefanst]

I'm having a clear WTF moment here. Trying to fire the '03 up to take a look at the voltage problem and I'm not getting any spark. I even reverted to a previous tune, thinking I did something. No spark. I need some help thinking.....

2016-12-01_19_44_52rpm_No_spark.png

I'm not running into boost cut, rev limiter, anything like that. Fixed or dynamic spark makes no difference.

[AndreyB]

I cannot explain it. Injectors are triggered so the trigger is decoded. What has changed? Have you updated firmware? Have you changed anything in the project? Can you load your last working tune?

Is ignition disabled for some reason?
showconfig in console would show some things

[stefanst]

Stupid user is stupid.
I found a dwell table online that I wanted to use. Dwell at 0rpm seemed totally unnecessary to me, so I left that out. Apparently that's needed since cranking RPM is below 500 (the lowest value in my table). Funny thing is that it's not using the interpolated dwell from the dwell table, but the correct cranking dwell. Makes me wonder what happens if we exceed the max rpm in the table...

[stefanst]

On to some more fun features:
I scoped the supply voltage and it's noisy as all get-out:

12V ripple at board.png

Also scoped the GND on the board vs. chassis:

board_gnd_vs_chassis_GND.png

So most of the noise is actually in the supply line. And there's not much else hooked up to it. This is also without the alternator running.
This means to me that we need to do some more smoothing for the vbatt signal. I'd really like to see the smoothing factors for all analog inputs to be freely configurable of course, since different cars may require different smoothing.

Also, just to be a total jackwagon, imo we should measure the voltage of the injector supply line and not the supply to the board, since the main task of the measured voltage is to adapt the dead-time for the injectors and to control the alternator (of course dwell may be compensated as well, but a little variation here really doesn't make much of a difference).

[AndreyB]

I believe on my board VBatt was improved by some hardware change - was it a larger capacitor or was it a different resistor? It's somewhere on the Neon thread.

Also your tune shows expAverage is disabled:
<constant digits="3" name="slowAdcAlpha" units="coeff">1.0</constant> - that's configurable somewhere on analog inputs dialog.