Flex Fuel with MegaSquirt Controller
 

The write up in the MegaManual says
http://www.megamanual.com/flexfuel.htm

Does this work with MS-1 as well? I am contemplating switching to MS but need to decide between MS-1 and MS-II.

The main driver for me is to run more timing with the improved octane of e-85, but it's not always available ( and not always 85% ethanol either ).

The writeup addresses fueling adjustment, I assume I will have to add a MS Knocksense to back the timing down if the percentage of e-85 in the tank falls to low and I start to get some knock.

Anyway to use map switching to change the spark map if the concentration of e-85 drops below a certain point?

Is anyone using the flex fuel sensor? How is it working for you? What are you doing about timing?

This is not a "Green" or fuel efficiency exercise, it's all about power.

Thanks in advance,
--Chris

id rather have greener car then have 40whp more. interested!

MS-I does not support flex fuel sensors. MS-II can make both fuel and timing changes in response to a flex fuel sensor with the right mods, but it adds a certain amount of spark advance for every % change.

Unfortunately the amount of spark change required is not proportional with the fueling.

Would map switching based on a value coming out of the flex fuel sensor work?
If not, does the KnockSense provide enough protection?

Unfortunately the software currently doesn't support table switching based on flex fuel input. And I don't recommend on relying on a knock sensor to keep your spark settings in tune.

We have successfully used megasquirt to convert a yamaha snowmobile to flex-fuel. I won't pretend I know all the details as my task on the team was a completely different aspect of the project, but I know we used a flex fuel sensor from a GM vehicle and it automatically adjusts something important depending on the percentage of ethanol in the tank.

I think they (the engine control part of the team) had to generate a gasoline table, and an E85 table, and the megasquirt interpolates between the two tables. That is the part I'm a little fuzzy on though.

:subscribed: In for the power as well.

Matt, do you have any more info on this or any other links to successful installs for street vehicles?

Chris

There's a few discussions on the msefi.com forum - I don't have any particular links to mention but you'll find some accounts of people there running it.

My MS-II is hopefully on Joe's workbench right now having an input circuit and whatever else added to accomodate the flex fuel sensor. In reality, I won't get to the flex fuel experimentation stage until after September, but I will keep you guys posted when I start the process. The #51 SM2 SystemStrategies.com/DIYAutoTune.com Mazda Miata just has too many events scheduled to start playing with variable fuel mixtures until after the SCCA Solo National Championships.


(* Hey Matt, did I have the cap on straight in the interview?:))

It's on the bench, I'm just having some trouble communicating with it. I know the MS itself is operating as the FP output goes true for about two seconds upon powerup and then goes off (priming pulse), I just can't get my laptop to talk to it. I'm not seeing any activity on the Three LEDs, but I'm guessing that's normal for a baseline-configured MS2.

Currently I'm using the new serial adapter that came with your package, I think tonight I'll try using it to talk to my MS, and/or use my RS-232 adapter to talk to yours.

Make sure it installed the drivers. If you need them, they are on our site.

I used the tiny little driver CD that came in the package with the serial cable, and the adapter subsequently populated itself into my Device Manager tree. I changed it from COM8 to COM1 (which is unused) and configured the port for 115200, leaving all the other defaults such as 8N1, the buffers, etc.

I may simply test it with a null modem and a terminal program to one of my other known-good machines to validate the operation of the port. I expect that perhaps I'm missing something in the MegaTune config, which I've only just started tinkering with for the MS2.

BTW, Sam- can we sneak an addition into the car name somewhere? I'm thinking of The #51 SM2 SystemStrategies.com/DIYAutoTune.com Mazda Miata presented by UNC Law. I'll try to get you a hat. :D

Two things?
Who is Sam?
and
To get all that on a hat it will have to be a sombrero. Great idea, even more room for sponsor stickers:)
http://raeannemusic.com/Gallery/albu...0489.sized.jpg

Uhm, yeah... Chris... You'd think I could remember your name, given your choice of username and whatnot... :bang:

As to custom embroidered sombreros, here you are: http://www.asipinatas.com/Sombreros/index.htm

No hard feelings Bob, better get cracking on that Sombrero order. We don't run at nationals until Sept 18th and 19th, but will probably be there from the 12th on. That's roughly 8 weeks from now. They said 4-6 week lead time on the site. :sobrerosmiley:

I was looking into this some time back for the same reasons. Could be a fun project. The FlexFuel sensors were a little pricey but I got the impression that there were used ones to be had (though one model tended to fail?). My main snag was wrapping my head around moving from MS1 to MS2 at the time (getting all of the miata mods converted). I was busy so I gave up - would still love to do it. A lot of e85 to be had in Indiana.

Certainly watching this thread.

Success.

Started by validating the serial adapter itself. Used a null modem to connect directly to another PC and was able to have a terminal session between them at both 9600 and 112500.

Next, used the 'scope to verify that serial data was in fact getting all the way to pin 38 of the CPU (the actual 9S12C64) when I tried communicating, but was getting nothing back on pin 39. No activity in MT, wouldn't pass the "test" in the comm setup screen.

Yet all this time I know the CPU is alive, because I see activity on the FP line for ~2 seconds at startup...

Finally gave up and re-flashed the MS with the new MS2-Extra code, and now I can finally communicate with it. Man... Two days down and I haven't touched the hardware yet.

Anyway, Chris, I think we're finally in business. I'm going to try to get the cam input triggers done tomorrow. If I can get to the point where I see an RPM indication, I'll be happy. Gotta figure out how to configure an NA CAS in MS2. The wheel decoder in this software is way different from what I'm used to.

Good news, Chris- your MS2 is officially Squirting and Sparking.

I built the CAS input and ignitor output stages this evening and just got done testing them. The circuit is a bit of a hybrid- for the inputs I built the 4G63 circuits exactly as spec'd in the MS2Extra docs, including the second phototransistor circuit: http://www.msextra.com/ms2extra/MS2-...4g63.htm#2hall

For the spark outputs, I deviated from the MS2E docs and built them to MS1 spec, using Q6 and Q8 with 5v pullups through 100Ω on the collector side, as I'm more familiar with this circuit and I know it works on both the stock igniter and the COPs.

So I've got good activity on both spark outs, and both INJ outs (I haven't modified those, and don't plan to.)

Now it's decision time. I've got to find a place to mount the two big FETs for IAC and WI control. I've already removed and tossed Q16, the high-current ignition driver, as it's worthless in this application. So I need to make one more space on the heatsink. I'll give you the choice here- I can either remove R37 & R38 which will defeat the injector overcurrent protection circuit, or I can remove Q9 & Q12, which will defeat the ability to run Lo-Z injectors. I'm leaning towards pulling the resistors- their job is to prevent you from doing something stupid like putting four Lo-z injectors on one channel and blowing up the FET, removing them won't affect the functionality of the system.

While preventing me from doing something stupid is generally a good thing, I am quite happy with my 750cc High impedance injectors from RC (SH4-750D).
I think Q9 and Q12 can go.

Joe,
You could also mount them to the case, or get an extra heat sink and mount them under the board, or "stack" the extra heat sink on the top level(this is how we do it here).

Hmmmmm....

I decided against case-mounting on mine simply because I knew that the box was going to be opened and closed a million times as I experimented with it. I guess with Chris' unit this won't be an issue so I can just stick it on one of the endplates. And with the one spot we freed up by removing the 912 (which is truly useless in this application) I think that covers us for heatsink requirement.

BTW- I like the new insulators you guys are shipping with the TO-220 modkits, with the nylon shoulder washers and such. Much better than the old stuff.

Yea, I figured there was a reason you didn't mount them to the case.

Today's episode of "What the hell was I thinking?" is brought to you by the fact that MS2 doesn't even support WI control to begin with. So that's kind of a non-issue now. Chris, your heatsink components and endplates are both safe. Buy a Devil's Own WI pump controller. :cool:

Got the launch control / flatshift circuit and a pair of relay driver circuits built last night. Kept running into trouble as I was missing a few trivial parts (stuff I forgot to buy at Fry's on Wednesday) so today I've got what will hopefully be the final "needs to be purchased" list and we'll be able to finish up with everything else.

I was kinda surprised to see in the MS2E documentation that in order to do table-switching I've got to solder a wire onto one pin of the tiny little QFP48 microprocessor. Not a problem here in the lab (got a Metcal iron and a Leica stereomicroscope) but I can't imagine the average user doing this with a Radio Shack iron and a magnifying glass...

I have a working water injection setup now, it is just that I have to set the "On" point via a screw type adjustment on the pressure sensor under the hood. It's not progressive. I was hoping to just control it via an output map of somekind from the MS-II but if I can't it's not the end of the world.
Great, I hope you are keeping track of what I owe you. ( Picking up the old chub tomorrow. Not sure whether to box it or find a throwaway suitcase for it.
See, yet another opportunity for me to have screwed it up. Thanks again. I am getting excited, this is going to be fun.

Nah, I haven't really been keeping track, it's probably under $20- don't sweat it.

Got all the hardware mods finished tonight. I haven't exhaustively tested all the ancillary I/O yet, but the basics (fuel, spark, WBO2, etc) all seem to be working. A tip of the hat to the folks at DIYAutotune.com who did the principal assembly of the basic 3.0 board- a very clean job of it.

One thing I gotta say- a lot of folks bitch that the MS1 software seems confused and uncoordinated, well, the MS2 software is downright sociopathic. I'm still trying to figure out why the left-hand column in the VE and Spark tables is labeled in % rather than KPa, and scaled only from 0 to 100... Anybody have a clue on this? 'Cuz I'm stumped. I've set Speed Density in every space I can find.

Anyway, pictures. It's not my most beautiful wiring job ever, but it seems to be solid. I did the mods on the MS2 daughterboard to run down unused pins of the DIP-40 and then get picked up again on the other side, so you don't have to worry about wires dangling from the CPU to the mainboard.

http://img295.imageshack.us/img295/4...isfrontys7.jpg

http://img295.imageshack.us/img295/3...risbackoj5.jpg


I've built the following functions onto your board:

1- Relay driver for fan (closure to ground).
2- Relay driver for misc (closure to ground).
3- Input for table switching (active-low).
4- Input for launch / flatshift (active-low).
5- Constant barometric correction.
6- Input for knock (for KnockSenseMS).
7- Input for Flex Fuel.

Full documentation on the mods (pin assignment, etc) will follow. Still haven't started on the harness, though that should be easier now that all the I/O is, if you'll pardon the pun, pinned down.

Pictures like that certainly keep me humble. Thanks for the hard work.

Now, do you want the good news first or the bad???


Good news, the Old Chub has arrived safely in San Diego...






















The bad news... There were only two six packs available for purchase in Colorado Springs.

More good news.

All I/O have now been tested and are operating. The only one I haven't checked out is the FlexFuel input, as I have no way of simulating that signal at the moment. We've got a fancy function generator at work that I'm going to borrow, and hopefully I can produce a signal with that which reasonably approximates what the MS is looking for. I've never actually seen a FlexFuel sensor in real life, so it'll be an educated guess.

Now, onto the boring part: building the damn harness... You friggin' Plug-n-Play people... :fawk:

Great news.
In the words of Harry Callahan (Magnum Force 1973) "A man has got to know his limitations..."

Man, I hate harness-building... Easy, but tedious...

http://img37.picoodle.com/img/img37/...sm_43b8c51.jpg


When you get this, you'll gain an appreciation for my perspective regarding grounds. When building an MS, it's important to have enough ground wires. Too many is enough.

Three to the main connector, and two more are those two black wires. I suggest you pre-run a pair of heavy wires through the firewall to a new ring connector on the head for these to connect to.

Also- what's the status on the connector for your IAC and TPS? I've wired this to use your stock TPS wires, but you're going to need to cut off the stock 3-wire plug and install an NB 4-wire plug. Start hitting the junkyard now, I'll provide instructions to follow.

The harness looks great. The TPS plug is already taken care of, although I ran a wire outside the stock harness to carry the variable signal. I will start looking for the IAC plug, but if I can't find one, I can probably get the one from the "donor" car. I haven't cut the harness on it, and am trying to avoid doing so.

I can suffer with a poor idle until warm with no ill effects until I secure one right?

(How was the beer?)

Mmmm. Beer good.


For TPS, you'll want to wire the new TPS plug into the factory harness as follows, assuming '99 colors on the TPS pigtail:
'99 light green / red to '92 red. (TPS Vref from pin 1N / pin 26)
'99 green / black to '92 light green / white (TPS sig, pin 2L / pin 22)
'99 black / red to '92 black / light green (GND)

Specifically, here's what your pinout looks like altogether:

http://img32.picoodle.com/img/img32/...nm_be1cd74.gif

you guys frighten me you know too much ive been considering doing e85 conversion over here in uk on my 1999/10AE only you seem to go through alot of trouble other than paying someone else to do it is there a simpler way? alot of bio sites just want to sell me/you/anyone plugin piggy back ecu's plus wires and i know on my older car i have to change alot of seals pumps etc.. but you seem to have alot of trouble or am i missing something? sorry new to almost all modding and havnt even supercharged mine yet.:eek5:

It is pretty simple to convert to E85, the difficulty comes with Flex-Fuel I think.

Seals and stuff do need to be replaced, but mostly you just need to re-tune your VE table for E85.

Indeed. To do a straight E85 conversion, I believe you'd simply need to install larger injectors. Flex Fuel is where the complexity comes in- the ability for the system to automatically detect what type of fuel you are running and adjust accordingly, without user intervention.

Is there a way to do flex fuel AND coil on plug?

hi yes just to say i would be going for flex fuel as limited supplies i mean going over to the ring (nurinburg) on one tank aint goin a do it ! i didnt relize people were so helpfull thanks

One has nothing at all to do with the other. You can run COP with pretty much any ECU, on any fuel. So yes, after having converted the fuel delivery and engine control systems for flex fuel operation, doing a COP conversion will work just as well as if you were still running normal petrol.

If you ask good questions, around here you'll get good answers. :bigtu:

Back from the dead.

Joe,
Can you tell from the pictures if I can run low impedence injectors? I can't tell for sure from reading the thread whether or not we deleted that capability.

Also, is sequential injection possible? What all would YOU have to build on my setup to make it happen?

--Chris

I found a post where Joe already answered the question, I can run low-z

I can answer that, Low impedance injectors yes.

Sequential requires an additional 2 injector drivers, either two copies of your current circuits or 4 new ones via JBPerfs injectors&ignition driver card.

To keep low impedance injectors with sequential you'll also need his peak and hold driver card if you use his injectors&ignition driver card

Okay, I think I have mentioned on this forum before that my electrical knowledge begins and ends with "Don't let the magic white smoke out".

I went to the jbperf site, saw the boards and some directions with squiggly lines and colors and came back here with questions.

Question one: Is there a way to utilize the sequential ignition and injection driver card without owning a soldering iron?

Question two: Do I need to be reading the Crank trigger wheel as well as the CAS to know when to inject? ( I have a '99 four tooth wheel on the car, but the sensor is in a parts box somewhere)

Question three: If I go full sequential on the injection, will my current injectors be big enough? I currently squirt twice per cycle right? So full sequential gives me greater control but less capacity?

Yes, make someone else do it for you
Yes
Shouldn't change as far as I know. Injection time just changes, not the amount. Someone else more knowledgeable please confirm/deny.

1) nope
2) just the CAS is fine assuming you're using both signals from it
3) Sequential won't effect your Duty Cycles i.e. you won't have to go larger because of it

EDIT: TriggerZ, Chris doesn't need an additional crank trigger when there's already a crank (and cam) trigger coming from the CAS.

Still confused on #3. Why won't my duty cycles double when I only inject half as often?

doesn't work like that.

you injecting once per cylinder per engine cycle, whereas with batch it's twice per cylinder per engine cycle.

So injecting less often, but more fuel, it balances out.

Half as often and twice as long is still the same ratio. Well, it may change a bit when you factor in injector dead time, but not that much.

Also note that 2 squirts alternating, which is what many Miatas run, fires each injector only once in an engine cycle anyway.

Sorry this isn't sinking in for me.

If I currently inject twice per cycle for 10 ms, won't I need to inject for 20 ms if I do it once per cycle? Assuming ( I know its a big assumption) that is correct, if I was at 50% Duty cycle won't I now be at 100%?

Matt,
How do I tell if I am on 2 squirts alternating? (found the setting that says Injector staging, but the selection above it in tuner studio says squirts per engine cycle = 2)

What I ultimately want to accomplish is individual cylinder fuel trim. Would that be possible with only two injector drivers? (something tells me that is wishful thinking, but you gotta ask to learn)

exactly.

twice per cycle is 2x 10ms

versus once per cycle 1 x 20ms

so the DC is identical (or near as damnit due as Matt says to injector dead time)

In Engine Constants, there's two selections, one is for number of squirts and the other is for simultaneous/alternating injection

Only possible with full sequential. If you think about it if you tried to trim for one cylinder with batch injection you'd actually be trimming from two...

When you are squirting each injector twice per engine cycle, or once per crankshaft revolution, the displayed duty cycle is based upon the time of one crankshaft revolution.

So, let's say you're turning at 6,000 RPM, or 100 revs per second, your injector cycle time is 10 milliseconds (1/100 sec.) The injector squirts for 5 ms each time the crank comes around, so your duty cycle is 50%.

Now we'll say you've gone full sequential. Each injector is squirting once every two revs of the crankshaft. The injector cycle time is now 20 milliseconds (two crank revs) and, to deliver the same amount of fuel, the injector squirts for 10 ms. Still 50%.



Nope. Individual trim requires one driver per injector.

As already answered, the NA CAS is capable of sending both the crank position and cam position signals. However, a trigger wheel at the crank will provide more accurate information to the PCM than a wheel mounted at the cam shaft. This is do to timing belt slop and flex in the cams. Many people call this solution good enough, but it does leave a bit on the table.

However, if you go with the NB super weird 4 tooth trigger wheel for crank position, you will also need to use the weird NB cam position as well, as it is also an unusual pattern, unless you want to write some custom code.

Lucky for me I left that sensor in place to plug the hole in the front of the valve cover. That is the one right?:noob:

Yeah. Did you leave the original cam gear in place too? The cam gear has the weird pattern.

Sure did.

Then you can use the NB triggers (with the Jason/Abe Schmitt circuit) which will be more precise than the NA CAS (in fact you can remove it), and you will be able to run fuel and/or spark sequentially.