Mirosquirt Install on Subaru EJ20K

[stew3223]

Hello all,

Hoping to seek your collective advice.

I recently purchased and installed a Microsquirt module in a '98 Impreza swapped with a JDM EJ20K version 4 STi engine. This is a wasted spark engine from the factory and uses the common Subaru 6/7 crank trigger. I've done a number of these swaps before but this is my first experience with Megasquirt. A few issues...

1.) I can't seem to find how to enter the lambda sensor scaling. I''m using an AEM wide-band O2 gauge with 0-5v analog output. The only options I see are under: Vehicle Project/Project Properties/settings and the only choice is "wideband (default)". I also see that under the Tools tab there is a selection for "calibrate AFR table" hpwever its grayed out and I can't select it.

2.) My factory map sensor is a 2 bar sensor. The scaling is such that the inputs need to be: 0v = -160kPa and 5v = 200kPa however tuner studio won't except a value lower than -100kPa. I understand this doesn't seem physically plausible however I verified the scaling with a hand vacuum pump and a DVOM. Any options here for entering a custom scaling? The sensor is a Denso 2201AA090.

3.) Again, I'm missing something with entering the scaling for the rest of the sensors. I can't select "calibrate thermistor tables" under the Tools menu. So although my coolant and IAT sensors are giving reasonable values, I can't verify the scaling. This seems so simple I feel stupid asking.

4.) Despite the above issues I attempted to start the engine to verify that my Cam and Crank triggers are working correctly. While cranking the engine seems to be having an intermittent sync issue. I noticed the "Not RPM Synced" and "Ready" indicators going green/red over and over. I am registering an engine speed, but it too cuts in and out. This engine uses VR sensors. I have Cam signal wired to VR2In+, Crank signal to VR1In+, Cam and Crank ground as well as shield ground wired to VRIn-. I can also see in the log the engine speed goes from ~300 to 0 and back repeatedly, leaving me to believe theres and issue with the triggers. I've attached a log but I'm not sure if there's a way to look at the sensor signals them selves in megalog viewer. Any help is appreciated.

I also attached my .msq for you to review. I was able to get the car to kind of run, sorta. But only for a few seconds, so I feel likes close and I'm just missing something. Obviously I need to get the above issues sorted first.

Thanks in advance!

[stew3223]

Okay, so I figured some of this out last night.

I figured out that I had to click "Unlock Calibrations" under the tools menu to be able to scale the AFR and Thermistors. I'm not sure why these wouldn't always be freely configurable but anyways...Lambda, CLT, and IAT are taken care of.

I still can't scale my stock MAP sensor properly, however I ordered a GM 3 bar MAP which will be here today.

I was able to get the car to start and idle. There are still sync errors on cranking, but once it's started they go away. So it seems my triggers and wiring is OK.

Hopefully I'll have another update tonight.

[hybrid]

The ini file often has min/max limits you can set. Have a look through it.

[stew3223]

So today seemed like a step backwards.

I installed a GM 3 bar map sensor. Re-scaled IAT and CLT sensors. Fixed an issue with IAC. After this, The engine starts and idles but it's constantly getting sync errors now, even while idling.

Is it normal for the fuel pump relay to switch off when the engine loses sync? I can hear the relay clicking every time the sync lost indicator flashes.

I checked the mechanical timing of the engine and ensured it was correct. Same issue persists.

I took a composite log. Any insight would be appreciated. Not really sure where to go from here. I'm going to do a compression test tomorrow just to make I'm not battling any other issue.

[stew3223]

So here's a question I just thought about...

Since the firing order on the Subaru's is a bit different than most 4 cylinder engines, do I need to pair the injector wiring different than how the manual states to wire it?

I currently have Injectors 1/3 pair together and Injectors 2/4 paired together. However, because the firing order is 1-3-2-4 does this have any effect on how the injectors should be paired?

Do I need to change the sequencing?

Thanks,

[hybrid]

Since you are not running sequential injection then it doesn't really matter.
You could try switching to 4 squirts per engine cycle, which will give smaller squirts more often.

Sync errors could be a sign of noise. Are you running resistor plugs and supression leads?

[stew3223]

Since you are not running sequential injection then it doesn't really matter.
You could try switching to 4 squirts per engine cycle, which will give smaller squirts more often.

Sync errors could be a sign of noise. Are you running resistor plugs and supression leads?

I'm using NGK BKR7E spark plugs, very standard for this engine. I'm not familiar with suppression leads? I'm using the stock plug wires.

I'm also using the OEM shielded circuitry. I spliced the connection in at the ECU, so there's maybe a small gap in the shielding.

Do I need to be using something else? The composite logger definitely doesn't look right...I'm certainly losing some of the cam signals, and showing Lost sync reason 21. Is there not a calibratable threshold for the VR signal voltage?

I attached a screen grab of the trigger log. Also attached another log from this morning.

[hybrid]

There are no pots to control the sensor input like there is on the MS2, but there are noise filtering settings within TS you can try to mess with.
Ignition Settings -> Noise Filtering

[stew3223]

There are no pots to control the sensor input like there is on the MS2, but there are noise filtering settings within TS you can try to mess with.
Ignition Settings -> Noise Filtering

Okay, the noise filters we're disabled. Is it pretty standard that some filtering is required?

Is there anything else I could consider trying?

[hybrid]

I haven't used the noise filtering on either Micro or MS2, but I'm using hall sensor input.
Having said that, I switched from Micro to MS2 because of some weird issues with it, and I'm now a firm believer that unless you need the small footprint of the Micro (for a bike etc), then use MS2 instead.

Anyhoo, there's plenty of people using noise filtering, so it's fine to use.
I can't help as far as your trigger log goes as I have never fully understood the composite log - I've only ever used crank triggers, without cam.
I think graph 1 is your crank trigger, graph 2 is your cam trigger, but I can't remember what the red indicates.
I also don't know that engine, so don't know how many teeth (or missing teeth) are on each of your triggers.

If you just do a tooth log, you should be able to see a similar length bar for each tooth, then a longer bar where the missing tooth is.

[stew3223]

I haven't used the noise filtering on either Micro or MS2, but I'm using hall sensor input.
Having said that, I switched from Micro to MS2 because of some weird issues with it, and I'm now a firm believer that unless you need the small footprint of the Micro (for a bike etc), then use MS2 instead.

Anyhoo, there's plenty of people using noise filtering, so it's fine to use.
I can't help as far as your trigger log goes as I have never fully understood the composite log - I've only ever used crank triggers, without cam.
I think graph 1 is your crank trigger, graph 2 is your cam trigger, but I can't remember what the red indicates.
I also don't know that engine, so don't know how many teeth (or missing teeth) are on each of your triggers.

If you just do a tooth log, you should be able to see a similar length bar for each tooth, then a longer bar where the missing tooth is.

Why should there be any difference between Microsquirt and MS2 except for IO content? As far as I know, isn't Micro just a pre-assembled version of MS2? I don't need the extra IO...so what exactly is the advantage of MS2 over Micro?

Is it possible I should be adding resisters to the VR circuits? Is that necessary with or possible with Micro? A lot of my reading points to people requiring inline or shunt resistors with the 6/7 trigger...but they're all on MS2 not Micro.

[hybrid]

You can try inline resistors. Can't hurt, especially if others have had success. Micro is essentially the same except for its small size. Lots of components jammed in very tight together, more susceptible to noise - in my opinion (and experience).

Sent from my A0001 using Tapatalk

[stew3223]

You can try inline resistors. Can't hurt, especially if others have had success. Micro is essentially the same except for its small size. Lots of components jammed in very tight together, more susceptible to noise - in my opinion (and experience).

Sent from my A0001 using Tapatalk

Okay, thank you very much for your input.

Re: resistors... Here's one thread I'm referencing... http://www.msextra.com/forums/viewtopic ... 01&t=39453

I want to clarify... he's run the signal straight through a 12k resistor and a 2.2k resistor accross the VR + and - ...Does that sound right?

[rickb794]

I would take the time to look at your plug wires, the use of low resistance (spiral wound) plug wires is a grey area. They do not offer the suppression that the higher resistance wires (carbon core) do.
Measure the resistance if you are not sure. Not sure where to draw the line but a 1k ohms minimum does not seem unreasonable.

Also check the impedance on your injectors. Section 4.10.2 read it all

The third area to look at is grounding. Section 3.2 critical!

http://www.msextra.com/doc/pdf/Microsqu ... re-3.4.pdf

Some it seems have to use it but I feel noise filtering is a bandaid, therefore a last resort.
My reasoning is if something is that noisy there is probably something wrong, and it is likely not stable.

And speaking of noise USB rarely works well in the noisy automotive environment.
A very short, or no cable USB to serial convertor is best. Bluetooth seems to be a good alternative if your computer has it.

[stew3223]

I would take the time to look at your plug wires, the use of low resistance (spiral wound) plug wires is a grey area. They do not offer the suppression that the higher resistance wires (carbon core) do.
Measure the resistance if you are not sure. Not sure where to draw the line but a 1k ohms minimum does not seem unreasonable.

Also check the impedance on your injectors. Section 4.10.2 read it all

The third area to look at is grounding. Section 3.2 critical!

http://www.msextra.com/doc/pdf/Microsqu ... re-3.4.pdf

Some it seems have to use it but I feel noise filtering is a bandaid, therefore a last resort.
My reasoning is if something is that noisy there is probably something wrong, and it is likely not stable.

And speaking of noise USB rarely works well in the noisy automotive environment.
A very short, or no cable USB to serial convertor is best. Bluetooth seems to be a good alternative if your computer has it.

Thank you very much for the reply.

I'll measure the plug wire resistance tonight.

All Subaru EJ engines, so far as I know, and specifically the EJ20K use high impedance injectors. I will verify their value tonight, however I don't expect this should be an issue.

My grounding is setup as follows...and I'm now realizing I may need to revise this...

Microsquirt Pin 22 is wired to B84 - Pin 19 (see attached image)
Microsquirt Pin23 is wires to B84 - Pin 46 (see attached image)

I mistakenly thought these we're different ground locations. I'll move Pin 23 to B84 - Pin 94 ... This should be a different engine ground location.

I also realized I may have wired the VR grounds and shields wrong - This was confusing in the Microsquirt reference material. Currently, I have VRIN1+ wired to B84 - Pin 8 Crank position sensor signal (see attached image), VRIN2+ wired to B84 - Pin 7 Camshaft position sensor signal (see attached image). I have VRIN- wired to B84 Pins 29 and 28 as well as 54 (shield). I should probably have the shield on a separate ground, correct? Sensor ground?

[stew3223]

Good morning,

Plug wires measure between 9.5K ohm and 10k ohm. NGK BK7RE's are categorized as "resistor" type spark plugs.

Injector resistance are all around 11.5 ohms - High Impedance.

I replaced the one of the ground locations as described above - no change.

I re-wired the shield ground from VRIN- to ground - no change

I changed polarity of the VR sensor inputs because, frankly, the reference material for this product is confusing... No change.

VR Sensor Confusion.jpg

Finally I threw the hail Marry and installed a 36-2-2-2 crank sprocket. From what I've read, people have had greater success using this pattern with MS AND it doesn't require the use of a Cam sensor when doing non-sequential injection and wasted spark. I changed the appropriate settings in the .msq and fired it up. I'm still getting constant sync loss errors, both at cranking and while running. The engine idles very rough, and since the tach is reading engine speed and then not (0RPM) the controller is cycling in and out of cranking/not cranking, running/not-running and this is cycling the fuel pump relay on and off continuously at about ~4Hz. I wasn't sure if this was a symptom or a cause so I tried to jumper the fuel pump relay, no change.

Finally, desperate and extremely frustrated, I decided to check the cranking compression, just to make extra sure this engine is healthy. I measured 150 to 152 PSI cranking compression on all 4 cylinders ... probably some of the best numbers I've ever seen for one of these engines, especially for a used engine.

latest .msq attached as well as a tooth log and a data log. Really, I'm very unsure of where to go from here...

[rickb794]

Not sure I would trust the OEM harness for your main ground(s), they should be tied directly to the motor.

Those are not typo's, VR sensor polarity is "adjustable". I have not yet learned how to determine which way is correct.
I believe that yours is correct as the tooth pattern has peaks instead of plateaus.
Or perhaps the tooth logger is not the correct way of viewing the pattern.

Hopefully someone experienced will chime in.

[stew3223]

Not sure I would trust the OEM harness for your main ground(s), they should be tied directly to the motor.

Those are not typo's, VR sensor polarity is "adjustable". I have not yet learned how to determine which way is correct.
I believe that yours is correct as the tooth pattern has peaks instead of plateaus.
Or perhaps the tooth logger is not the correct way of viewing the pattern.

Hopefully someone experienced will chime in.

I have a hard time believing this is a grounding issue. The OEM grounds I'm using are engine ground points. HOWEVER I will try running two new engine grounds to see if there's any change.

[rickb794]

I have successfully used the OEM harness grounds that did not even go to the motor (they went to the body).

I depends on many factors and grounding may not be a factor in your instance.
Take a resistance measurement and check for any ground loop voltage before going too far down the grounding rabbit hole.

Does your sync status change with the motor running?
In other words does it sync while cranking then loose it when it fires?
Without sync I'm surprised it runs at all.....
Your log will show the reason for sync loss with a number.

[stew3223]

I have successfully used the OEM harness grounds that did not even go to the motor (they went to the body).

I depends on many factors and grounding may not be a factor in your instance.
Take a resistance measurement and check for any ground loop voltage before going too far down the grounding rabbit hole.

Does your sync status change with the motor running?
In other words does it sync while cranking then loose it when it fires?
Without sync I'm surprised it runs at all.....
Your log will show the reason for sync loss with a number.

I'll measure the resistance to ground when I get home. I agree, I think these are good grounds. I've used them before on other swaps with different ECU's.

It repeatedly cycles between "synced" and "not engine synced" both while cranking and idling. I should qualify "idling" ... It's barely running, with constant and overwhelming misfire. When the speed signal drops out, the controller thinks its below the "cranking" threshold so it goes back into ASE and a number of other cranking/starting correction are applied as well, it appears. It's also cycling the fuel pump relay on and off. So yeah, it runs, if you can call it that.

The log is giving me "lost sync reason 30." The only information I can find on that is " 30 = 36-2-2-2 " ... So not super helpful. Is there someplace I can find more reference material on the lost sync reasons? or any diagnosis?