Simple evap system control with megasquirt
 

I have an NB2 with ms3 pro Evo. The car has an engine swap, but that's irrelevant to this topic, and it's boosted. I want to do something with my evap system. It's a street and track car.

Removing the entire system means building pressure in the gas tank, and drilling a hole in the gas cap means sloshing gas onto the track.

Moving to a simple breather under the car means being a bad person and killing penguins.

I'd like to do some kind of simple evap system that could keep the tank as a closed system and then vent into the intake under the right circumstances, like stock.

Has anyone done this? I presume I could just run a simple valve that opens when the car is at mid-throttle and mid vacuum, and stays open when the car is off.

Pos-cat for the question; this is highly interesting to me as well (for the stock system, and I admit I have not searched yet). I assume it would be rather simple, since you already have the inputs (MAP & TPS), so you could wire an output to trigger the OEM "actuator"/valve?

I personally eliminated the solenoid valve and simply installed a check valve. The flow at idle is low enough that I have no issues idling as low as 600 RPM. You could use a spare digital output if you wanted to control a solenoid with RPM/map qualifiers.

IMO evap is required for applications similar to mine. Return-style system running e85. The fuel temp increase coupled with the sealed tank can cause a significant amount of pressure. I relocated the evap canister to the rear driver fender-well (where the fuel filter is located on passenger side NA). I have been through inspection and passed.

Edit: and parked in a garage, that I prefer doesn't smell like a refinery.

That's an interesting idea. I was thinking I'd want to avoid drawing a heavy vacuum on the tank at WOT, but the boost would take care of that.

What size injectors are you running? Any extra fuel could make it harder to idle big injectors.

I have a 400hp e85 capable fuel system with ID850 injectors and a DW200 pump. I can maintain a stable static idle with this setup at 550 RPM. I bet I could go under 500 if I close the idle adjust screw.

Good to know, thanks!

Ted, I'm running it with just a check valve, and so far so good. Thanks for the suggestion.

Glad I could help

Sorry if this is a dumb question but - would replacing the solenoid valve with a simple check valve also work on an NA with standalone / ms3?
I don't want to remove the canister and would like to have it actually work. :)

If I were doing a fresh install today, I would hook up the purge valve and drive it with the ECU. A little homework is probably required, but a good side project for a beginner where the risk is low, but one can experience satisfaction and provide something new in the community.

But yes, a check valve works on a NA chassis.

Anyone have an idea how much vapors a typical canister holds (days at temp) before they are released into the atmosphere anyway? Another way to ask... Ted, how many days in the garage before you begin to smell alcohol?

EDIT:
Also sparked by this thread... I'm idling at 800, but have been wondering if I could go lower. I might play with lower, now that I'm looking into improved fuel economy. When A/C is on, I like to run about 1200RPM, though, to make the system effective.

If I don't have evap in the summer the garage will smell after 1 night.

Well just replacing the solenoid with a valve is much less :effort: than either routing another wire into the engine bay or hacking into the Mazda loom but yeah I guess I could use one of the programmable outputs via conditions in tunerstudio.

But what conditions ? :hsugh:

And how long WITH evap? Not that I plan to re-install, but a question I’ve had for some time.

DNM

Well I am drawing from memory, but I believe the solenoid is already wired to terminal "X" on an early NA, so one would simply need to route that to a spare open drain output (I would probably pick an unused injector channel and wire inside the ECU case).

Conditions are pretty simple, websearch should confirm, but I would think RPM greater than say 1500 and load between 40 and 60kpa would do the trick.

No smell with check valve EVAP, ever. Gas cap doesn't let a bunch of air out when filling either.

I did discover that with hot fuel and a low tank, the purged vapors at idle would cause a rich condition (especially with gasoline IIRC). A solenoid would presumably fix this.

Yeah that's what i meant - solenoid is wired to pin 4T / Yellow+Red on my 95 EU 1.8 loom. I got unused low side outputs on my ms3 basic DB37 connector.

Ok thanks I'll look into that!

To follow up on this:

According to miata diagnostics wiki evap control on the 1.6 works like this.


And the '92 workshop manual states:
[quote]
Operation FUEL VAPOR VALVE
The solenoid valve (purge control) is controlled by duty signals from the engine control unit to perform purging
of the charcoal canister. Purging is done when these conditions are met:
(1) After warm-up.
(2) Driving in gear.
(3) Accelerator pedal depressed (idle switch OFF).
(4) Oxygen sensor functioning normally.
[quote]


Now, regarding the 1.8 system I couldn't find any specifics on the web. Closest thing to working conditions I could dig out of the german 93 1.8 workshop manuals was this diagram:
Attachment 228195

So tl;dr during partial load with warmed up engine, acceleration and full load conditions the purge solenoid is venting. In all other conditions it is disabled.


So rpm > 1000 + map(40-60kpa) would also cover partial load on a cold engine where the solenoid should stay disabled. How can I also cover coolant temp with only 2 possible conditions for my programmable output ? :dunno:

MS3 with loops can stack in the CLT along with the other two parameters.

DNM

Activating the EVAP without calculating the consumed fuel volume and just letting it apply a strong vacuum inside the fuel tank is a bad idea and will likely cause the fuel tank to crack in the long run.

Right I guess this is where my noobish attempts digging at this thing end.
So check valve instead of solenoid even worse idea?

Do you have an idea how to get some use out of the canister? ​​​

At the minimum, since you can't write custom logic/firmware, you would need a pressure sensor on the tank, and only activate the solenoid when the pressure inside the tank is above barometric and the intake manifold pressure is between 80kPa and barometric pressure. This will allow you vent pressure from the tank to the manifold at a "controllable" rate. You will also need a failsafe valve that will allow fresh air into the tank should the vacuum inside the tank fallow below approximately 90kPa and also set a CEL if this persists. This is somewhat doable on an MS3 if you have enougj inputs, outputs and loops left.

Link pnp ECUs all come with functioning purge capabilities. I don't personally use the function, but here's a copy of the help section on the topic:

Where is the in-tank pressure sensor on a 1991 miata?

There isn’t one

https://cimg3.ibsrv.net/gimg/www.mia...3a1273d4f0.jpg
​​​​​​'94 1.8 fuel and vacuum system





https://cimg6.ibsrv.net/gimg/www.mia...6e6f077f51.jpg
'92 1.6

exactly

Thought I might chime in with my recent experiences:

I just got my car running with the new built engine and turbo setup; car is still technically a MSM, but is built like a run of the mill turbo NB. BP4W with forged rods and pistons, flattop intake, LS coils, FMII 2560R kit, with AC/PS

In the process of removing and re-installing the engine I decided to delete the EVAP canister and solenoid in the engine bay to make room for a catch can. Everything in the rear of the car remains the same, except with a DW200 fuel pump.
I originally did as a couple of users suggested and ran the fuel tank vent line straight to the intake manifold with one of the crappy little plastic check valves inline to prevent pressurizing the intake. I found that after cruising around in high vacuum for a long time, the car would start to crackle and pop on decel as if my decel fuel cut weren't working properly. Additionally, the car would begin to mysteriously and intermittently idle on 3 cylinders when hot and would have a dead miss on cylinder 1 under very light throttle. When this occurred my AFR's would go pig rich and the car would idle like garbage as my O2 control tried to pull the fuel back. As soon as I dipped into boost, cylinder 1 would wake up and everything would be fine.

My investigation proceeded as follows:
Disconnected the coil for cylinder #1 while idling on 3 cylinders and found no change in idle; pulled the plug and it was black, fouled and reeked of fuel. Coil appeared to be firing with a healthy spark jumping to the mounting bracket
Fresh plugs, new day... no change. After driving for about an hour the mysterious misfire and bad idle had returned. Pulled plug #1, fouled again.
Checked compression suspecting bad rings or valve problem; all cylinders were an even 150 psi.
Swapped coil 1 and 2 suspecting intermittent coil issues with heat soak... No change.
Changed fuel injectors from FF 900's to FF 640's, suspecting a leaky injector... No change.
Removed PCV and vented both sides of the valve cover to breather tank, suspecting oil flow from PCV to be fouling out the plug... No change.
Disconnected fuel tank vent line and capped the intake port... PROBLEM SOLVED

This past weekend I drove 200 miles and ran an autocross event and experienced no issues while the vent line was disconnected and venting to atmosphere. So far the only plausible explanation I have is that the engine was sucking extra fuel from the tank under high vacuum and at idle, fouling the plug and causing the obnoxious burble tune noises. As soon as I got into boost, the check valve would do its job and the issue would disappear.

Right now my plan is to either run the vent line back down into the frame to keep the fumes out of the engine bay or to implement a solenoid that is only open for a few seconds at a time during steady state cruise.

Check valve goes to evap canister, not the tank.

I removed the small canister and solenoid valve in the engine bay. I connected the smaller diameter hardline in the engine bay to a port on the intake manifold with a check valve inline.

To clarify, I deleted what is labeled as the "catch tank" and the "purge solenoid valve" in the engine bay. The large charcoal canister at the back of the car remains. For some reason deleting the solenoid valve and catch tank seems to causes issues on an MSM, or at least on my particular car.
https://cimg7.ibsrv.net/gimg/www.mia...3592eaeb42.png

Would that be because you're now pulling 20hg of vacuum at idle on the fuel tank with a direct connection between the hard pipe and intake manifold? You've bypassed everything that controls vacuum flow rate if I'm not mistaken?

Yes, I believe that's why I'm having problems. What I don't understand is how people can do that on an NA and not have similar issues.

On an NA it looks like the charcoal canister has a vent line going directly to atmosphere, which would keep the pressure slightly below ambient but probably wouldn't suck hard enough to get a substantial amount of fuel. It is unclear to me how the check valves before the charcoal canister operate.

On an MSM the vent line is present, but I wonder if the leak detection pump is normally closed and is allowing the line to get down to substantial vacuum.

Depin the fuel tank pressure sender and wire it to an analog input; you’ll be able to monitor tank pressure.

if you’re stock, it’ll be in the OBD2 data.

Does an NB2 have a fuel tank pressure sensor? There's one listed on the 99-2001 evap system diagram but there doesn't appear to be one listed in my '05 evap system manual or on NB2 the ECU pinout. The EVAP leak detection pump doesn't appear to have a pressure output.

​​​​​​Also the FSM shows that the leak detection pump should be normally open if there's no power. Further investigation is required.

Either way, it appears that NA6, NA8, NB1 and NB2's all have EVAP systems that are different enough to require different strategies or deletion methods.

Why not get a GM FTP sensor and tap the top of the fuel sender assembly like GM does? Should be a linear 5V signal that you could use.

@jawatkins95
The NB2 does have a pressure sensor for the tank. It is mounted above the gas tank. Currently I am working a return system on my 2001. MS3PnP Pro and Trackspeed turbo'd for years now. Would have trouble putting gas in the car if on a road trip. Ended up doing a switch on the dash so could hit it as coming up on gas station. Worked well for that. Engaged and small throttle conditions it will definitely add fuel. if at idle could kill the motor. I don't have a solution just yet to mine. Limits of Megasquirt are a problem. One thought is a low pressure switch( idiot light type oil switch?) and Use ECU to control power to it during desired engine parameters.