[rusEfi] Miata NA 1996, 1.8L stock (#7)

[tomiata]

I ordered a SLC B Free 4.9 with the text display. I'll give it a try.

[kb1gtt]

About IAC, take note that your fuel delivery will change significantly as you change the IAC. It changes the pressure in the manifold, which changes how much fuel flows across the injector. Granted the fuel pressure regulator should keep that reasonably consistent, but it probably doesn't do it as well as you might hope. This could make for some short term issues with dynamic reactions.

[kb1gtt]

Tracing what is perhaps 1992 ECU I find this

Code: Select all

12V --> 12ohm IAC --> pin 2W or 3Q 8.2 ohm --> diode with cathode --> GND
                                           |-> TO220 to .47 ohm --> GND. 

The TO220 could be analog controlled. However it has minimal heat sinking. So due to a lack of thermal heat sinking, I think it's driven saturated, AKA switch for PWM. So I think the IAC sol runs at a constant 12V/(12+8)= .6A, and when the PWM drives it low, it increases the current to about 1A. This does not seem to perfectly match the above http://rusefi.com/forum/download/file.php?id=888 but it's probably getting close. That capture shows it going from 12V to about 8V. I would expect that with this ECU I traced, I would see 8V going to 0V instead of 12V to 8V.

So I would like to suggest we move the diode. It currently uses the diode with the white band connected to 12V. Lets try this. Same diode, but remove from 12V. Add 8 ohm resistor to diode where 12V was connected, then connect the other side of 8 ohm to GND. Remove the 34 ohm resistor. Then change the PWM duty to 250Hz or 500Hz. I think this will make it play nice. If I'm lucky, I'll get another scope capture and it will look the same as the above noted picture.

As a recap for another thread that mentions 500Hz duty instead of 160Hz, In the above graphic, it notes the OEM ECU used a duty of a bout 4mS, so I would expect the duty to be about 250Hz.

[AndreyB]

Add 8 ohm resistor

Average through-hole resistors are 1/2W, what kind of resistor do we need in this case? 2W? 4W?

For a test I want to try four 1/2W 33 ohm in parallel.

[kb1gtt]

The OEM ECU used what appears to be a 1 watt. This seems odd to me as I would expect 12.1 ohms + 8.2 ohms = 20.3 ohms. Then 12V/20.3ohms= .59amps. Then 8.2*(.59^2)= 2.79 watts. My best guess is that they expect the ON time to be low. Basically when the transistor is active, this 8.1 ohm is not active, so a 30% duty cycle would produce under 1watt in that resistor.

[kb1gtt]

Final wiring should look like this.

Code: Select all

12V --> IAC sol --> pin 4Q / W55 --> low side drive chip --> GND
                                 |-> + Diode - (white marked) --> 8 ohm --> GND

I believe the + side of the diode and low side driver chip is already installed like this. You just need to move the - (white marked) end of the diode, adding an 8 ohm resistor and connecting it to GND instead of 12V. When you do please take a scope picture to see if it is generating 40V spikes.

This is assuming your IAC sol is the same as the one used in a 92 ECU I have traced. I guess it's possible that this 92 ECU had a different sol. So this might not work perfectly. I have 2 ECU's here that I looked at and they both seem to have the same circuit, so I think this will play nice.

[tomiata]

I made a of bundle resistors, 5 47 ohm and a 100 ohm, 1/4 watt each, it measured about 8.3 ohms, should handle about 1.5 watts.
Here is a picture of the wave form,

iac-11-20-14.jpg

The buzz is back. And the resistors got pretty hot, only tested for about 10 seconds at a time. Didn't try to start the car with this in place.

[kb1gtt]

Hmmmm, that's matching what I was expecting. I see the signal going from 0V to 5V, and spikes that are tolerable. The best I can guess is that your OEM scope traces from above, do not match the circuit in this ECU. I really feel the need to make a circuit using one of these 2amp op-amps.

[tomiata]

Update on status, I had some cold solder joints on the frankenstein board that caused some erratic behavior. Now I can start reliably on rusefi when warm, and I'm refining timing and fuel settings. I'm stumbling along with that, any references on tuning would be appreciated.

And I got a 14point7.com WB O2 sensor kit and installed it. I'm not sure of the data that it's showing yet. It has a display showing temperature and lambda values. It takes a minute for it to stabilize and produce output. After it gets going the temperature value bounces around wildly. Not sure what's the deal there. The lambda value may be correct, but I don't know to expect it to be. It varies some, but mostly stays above 1.2 up to 1.36 (max value).

I'll try to get a video uploaded to show how it sounds. After starting, there is a hum sound coming from the MAF area, I'll try to capture on a video. It might be related to the IAC pwm freq, but the sound is not coming from the IAC side of the engine. Before starting, there is no noticable sound from the IAC.

[AndreyB]

[video][/video]

[spags]

Hi tomiata,

I've solved the ICV noise on my 94 miata, it turns out the hum is not the ICV it's in the airbox. Touch the airbox while running and you'll feel it. I think its caused by a harmonic with the icv frequency and the intake chamber.

Here's the link to the fix, works for me and hasn't broken yet.
http://rusefi.com/forum/viewtopic.php?f=3&t=729&start=30#p12431

As for the WBO, maybe double check your grounds, I hear those things tend to be sensitive to inconsistent grounds.

EDIT: Just had to say this after seeing the video of your car running, kickass yet another miata!!

[kb1gtt]

Fancy tool for finding and tracking noise http://m.harborfreight.com/mechanics-stethoscope-69913.html?utm_referrer=direct%2Fnot%20provided or the lower cost approach is to put your eat on a scree driver handle while you put the tip on the engine parts that you want to hear.

[tomiata]

Thanks, doing the "set_idle_solenoid_freq 500" seems to have eliminated the hum while running. But, now I can hear a 500Hz faint buzz sound coming from the ICV when the engine is off. I think I had the freq set to 250 before and didn't notice it when the engine was not running.

I'm trying to figure out the WBO. It's not showing good data yet.

I can start fairly well when warm and idles ok. If I adjust the fuel tables manually with set_whole_fuel_map, it changes the idle rpm. Setting to 4 makes it idle at about 1400rpm, setting to 6 makes it idle at about 1100 rpm. I don't understand this.

[AndreyB]

First of all, I really need to turn idle solenoid off while engine is not spinning - that's https://sourceforge.net/p/rusefi/tickets/117/

I've actually spend some time trying to tune idle today. I did experience the same thing when I was LOWERING fuel and the RPMs were going UP. I cannot explain it

I've started with 12ms giving me 1400 rpm (image how rich it was when you read the final value), I then went down to ~6ms and RPMs went to 1700. Step by step I was lowering fuel until some AFR values appeared on the display (my controller shows 0.0 if outside of the 9-19AFR range), I was also lowering idle_pwm to lower the idle rpms

Long story short, my results on the neon: ~900rpm idle with 2.8 ms fuel and idle_pwm 350

I'd suggest go lower and lower, until it dies. Once it does, you know you were too lean. Too rich does not stop the engine while too lean does stop it pretty fast.

[puff]

shouldn't EFI consider how lean/rich the mixture is when calculating ignition timing?

Long story short, my results on the neon: ~900rpm idle with 2.8 ms fuel and idle_pwm 350

what was the afr? what was the AFR after it changed from zero?



another theory: you were running too rich - explains bad o2 readings mentioned by tomiata, some fuel was going away unburnt to exhaust. by reducing injection time you sorta offset timing a bit. (shifting it away rom valve overlap?).

[tomiata]

Yes, I'm sure I'm running too rich. I smell gas, and my family tell me I smell like gasoline after I've been out working on the car.

[tomiata]

I did roughly the same procedure as @russian:

Long story short, my results on the neon: ~900rpm idle with 2.8 ms fuel and idle_pwm 350

And got down to 1.1 ms fuel and idle pwm 320, going lower killed the engine. My WBO showed a lambda down to 1.25, but doesn't change much through the process. I don't trust it yet. (I wonder if it got coated with unburned fuel or moisture and not giving good data. I had put it in my other car to test out, and it showed what looked like reasonable data.)

But, it won't start again after tuning down to these leaner values. I had to bump back up to 5ms of fuel and pwm to 400 or so to start up again. And then could tune down to lean again. Did that a couple of times.

Wrestling with usb serial connections for TS and console is a pain in linux. Spags, are you dealing with that in your setup? I saw your video with you using a linux laptop.

[AndreyB]

I did experience the TS issue today - https://sourceforge.net/p/rusefi/tickets/118/

[tomiata]

This thread has some example timing maps for the Miata 1.8L engine:

http://forum.miata.net/vb/showthread.php?t=237921

[AndreyB]

I think TS issue is here (that's in the rusefi.ini file):

#if CELSIUS
intake = scalar, F32, 8, "°C", 1, 0.0
#else
intake = scalar, F32, 8, "°F", { 0.01 * 9/5 }, 3200.00
#endif

so I am providing value in CELSIUS, but looks like the conversion is not right (the { 0.01 * 9/5 }, 3200.00 ) part

Any chance you would want to figure it out?

[meXanicus]

My version

Code: Select all

#if CELSIUS
      intake = scalar,  F32,8,"°C",1,0.0
#else
      intake = scalar,  F32,8,"°F",{1 * 9/5},32.0
#endif

TS calculates so. For example, 30 degrees Celsius, then 30 + 32 = 62 * 9/5 = 111.6. I do not know how to say that the TS must first multiply and then add 32
P.S. I do not understand English very well, so excuse me for google.translate.

[tomiata]

I'm back after a long break. From updating to latest firmware I'm running again. And I have a working WBO now.

I was able to start from a cold engine just now, and it runs fairly smoothly. A bit of a hiccup ever 5 or 10 seconds. AFR shows 9 to 10 range. Idle RPM sits steady around 1900. Changing idle_pwm has no effect now which is strange. Any known problems relating to that?

[AndreyB]

Idle RPM sits steady around 1900. Changing idle_pwm has no effect now which is strange. Any known problems relating to that?

Nope, as far as I know idle_pwm should work

[spags]

Hi tomiata, good to have another miata running the board!

The code for set_idle_pwm was changed to make more sense at some point. Now it takes a value between 0-100 and it will ignore out of bounds values.
I found that 40 makes mi car idle at around 1200, 35 is more like stock.

[tomiata]

Thanks a lot, that was it. set_idle_pwm at 28 puts me to idle at 1000.

[spags]

Awesome,

Mine requires about 10% more. I wonder if it's because of the temperature difference.

Are you running the solenoid frequency at 500? Wonder if that would have to do with variation.

[tomiata]

Hello after a long absence. Remember the clutch squeal? I just replaced the clutch and think I see the root cause. I think the throw out bearing was seizing up occasionally and grinding against the pressure plate. I put in a new clutch, fly wheel and throw out bearing, and other parts and got it running again.

See the groove cut into the fingers of the pressure plate.

IMG_20160107_215622~2.jpg

Hopefully I can try out rusefi again soon.

After a few rounds of trial and error and
speculation, I got a good tip from here:
http://forum.miata.net/vb/showthread.php?p=7158314#post7158314
And isolated the problem to the clutch, disabled the clutch safety interlock and I can start without scaring the neighborhood.

Now I can get a decent start on rusefi with no "horrible" sounds. For added flair I put the frankenstein in a box....

IMG_20141002_204826.jpg

... and got lots of help from russian and kb1gtt. Thanks!

[kb1gtt]

It your no longer scaring the neighbors, how are you going to advertise your doing cool stuff

Good to hear the progress. What is the next step?

[tomiata]

I need get it legal again with inspection and license.

[AndreyB]

One of the last tabs on the current rusEfi console now allows you to test injectors and fuel pump - these you would hear from your driver seat. Spark is a bit more complicated but could be tested as well.

On the pictures from your fresh logs note how top two lines - the trigger signal - always get wider after the first cycle with spark and fuel? Do you have any theories for what it is? Could it be a backfire causing the engine to suddenly slow down once you try igniting stuff? Could it be the the spark output wires are flipped?

That's all my ideas for now I see the perfect trigger input. I see 7ms fuel pulse which is OK. I see that after the first cycle something changes. What does not make much sense is why is it after full cycle and not after a first spark, if it would be a wrong coil it would probably be right away not just after four sparks?