Meth Failsafe and Patches
 

Hi All,

Rather than filling up my V4 thread with the failsafe stuff we decided to move all that discussion into here.

So what we are doing is working out how we can get an external meth failsafe system to trigger the ECU to run failsafe maps...

Take it away Jamie :)

Cheers
Dave

MAP SWITCHING
Ok, first thing we are determining is how Tephra has used the ICS auto mode switch to switch to alternate FUEL/IGN maps. This is the biggest failsafe you can ever have.

BOOST LEVEL SWITCHING
The next thing is boost level. Most guys use a MBC so even if their alky kits failed and the alternate maps were used, you'd still be pushing some serious levels of boost. Using a MBC, in conjunction with the stock wastegate solenoid (boost solenoid), there might be a way to have the ECU switch on/off the boost solenoid so its CLOSED when the alky kit is fine, and its ON when the failsafe kicks in (I'm now thinking that if it just runs the stock boost map at around 17-19psi, that'd be good). If someone took the time to alter the boost map to run a different boost pressure, that'd work too. I'm assuming most people would want their alternate map to be tuned for the most boost/power they can get without their akly kit activated. That way if the alky system is busted or they can't afford more alky to refill the reservoir for a while, their car isn't a complete DOG on the highway.

The MAP switching is done using the ICS switch. This is what the wires at the ICS switch do:



Green/White - (+)12v for light INSIDE the actual ICS switch (activated when parking lights/headlights are turned on)

Black/Yellow - (-) Ground for light in switch



Blue - Manual trigger wire for ICS relay (connect this wire to (-) ground to activate ICS relay) only requires 3mA of current
*NOTE* The IC sprayer pump will stay on whenever this wire is connected to ground, its intended for momentary activations. Keeping this wire constantly activated for over 30 seconds could kill the sprayer motor.

Blue/White - Auto ICS wire (momentarily give this wire (-) ground once to turn on AUTO mode, must release and connect to (-) ground again to turn off AUTO mode) only requires 3mA of current

Black - (-) ground source for the MANUAL and AUTO triggers


Ideally, you'd want the AUTO ICS trigger switched from momentary ground switching to:

ON = connected to (-) ground
OFF = no connection to (-) ground

Doing so would simplify the use of most alky kits failsafe outputs to trigger the ALTERNATE FUEL/IGN maps.

David, can the map switching be triggered under any conditions instantly? I'd assume you mimic'd the way the stock fuel maps switch between high and low octane maps, I could be wrong though.

I'm about to purchase a meth kit and this is something i'm very interested in.

Will these patches work anyway with people without an ICS switch like the 05+ Evo's

They should, Tephra was working on the ROM's for cars without ICS switches recently (He just got in a huge accident with his Evo so his priorities are a bit different at the moment). If you have the failsafe setup wired directly to the ECU pinout, the actual ICS switch isn't required at all.

If you've ever DL'd one of his patches and didn't donate any money to him, now would be a great time to show your appreciation and throw him whatever you can (even 5 bucks would help I bet).

If you are kind enough to help Tephra out, go to www.Paypal.com and send any donations to his paypal account by typing in his email address which is "donations@globalshare.net".

holy crap...i hope it wasn't his fault lookin at the laptop..."dontations coming +1"

Dude, that totally blows.

yup it sux :( right around christmas too :(

but please don't feel like you need to donate :)

anyways back OT, i'm not sure if using the ICS is the way to go...

maybe if we can find a universally free ecu input we could tap that and control the map selection via that... what do you think?

But every ECU has a wire on that pinout I believe. Its just a matter of the end user going to the ECU harness and tapping that wire. I'm looking for a wire everyone can "live without completely" and since no one needs an Intercooler Sprayer when they have Alky or water injection (most use the stock ICS reservoir anyways) there is no loss of another feature. Seeing as the non USDM IX's came with the ICS still (so I was told) I bet the ICS switch plug is still under the cup holder. Someone would have to check though.


Oh please, everyone should donate! Especially if they've ever used a feature you created,lol Its a community and god knows you do your part....times 10! Consider it a bribe from all of us to keep working on the ROM mod's if it makes you feel any better :D

Also, I'm not sure if you mentioned whether its possible to change the trigger commands to:

connected to ground = Alt Map ON
no connection to ground = Alt Map OFF

Has anyone checked for the plug under the console yet?

Not that I know of, wanna be the first? (hint hint) ;)

I will tomorrow when I get to my shop.

I know people will appreciate what a neat thing this is. I spent a good deal of money several years ago on an ebc and before that on an aftermarket emc. Both in an effort to get a failsafe on my water injection.
Now with relatively little money you can have everything those modifications gave me in a neat stock ecu.

This time around the features are redundant for me but I will be watching for the next upgrade.

merry christmas ecu master tephra

I did not find any unused connector. :( It looks like we will be adding some wires.

Damn. All we really need is there to be a wire on the ECU pinout. As long as thats there...everything is still pretty easy.

Very true but I was so hoping that the connector was there. Then all WE had to do was cut a hole and buy the ICS switch so that it looks factory.

Still could buy an evo VIII cupholder/switch and hardwire it. It's only 5 wires and 2 of them are just grounds (one is for the manual ICS sprayer relay but thats not even needed) :) One needs to get +12v when the parking lights come on and one goes to the AUTO switch wire on the ECU. I could draw a diagram if anyone wanted to do it themselves for that factory look.

I just saw what happened to your car, hope that works out. Looks fixable actuallyIt works for me, but this requires one to modify the DV, find a 3 port solenoid, have a 03/4 evo, play around with the ICS table so they respond only to throttle position, and know pneumatic systems. Not to mention have tephra make it work backwards .

I have logs showing it work, the load, air, and rpm slope drops at exact same point of first breached knock sum. I feel safer with this more than my flow monitor system. It just that it requires alot of exp for learning the system (not worth it unless you got the time). I just happened to integrate it into another working system rather easily, but it took me a year to get the original configuration down.

um front left wheel is fubared as well, snapped the steering control arm...

well if most ppl sacrifice ICS for meth then I guess its perfect.

You guys work out the best solution and I can code it up.. (when i have time :P)

Did we work out if the ICS bottle has a sensor for low water?

:thumbup:

Yeah, I'd say 95% of users will not need the ICS system since the alky cools the air charge far better anyways.

I still need to get to my ICS reservoir and test some stuff. Its 20 degrees and wet here so that puts a damper on my motivation,lol. From pics I've seen, it doesn't appear that there is a low level indicator wired to the reservoir.

Hey Teph, I don't know if you answered my question about changing the auto mode triggger from "momentary on" to activate and deactivate the ALT MAP to:

Not connected to Ground = ALT MAP #1 (aggressive maps with alky kit)
Connected to ground = ALT MAP #2 (conservative failsafe maps)

Found the pin for the only unused ADC channel on the Evo IX.

https://www.evolutionm.net/forums/sh....php?p=5101108

Jamie, um might be a bit hard to do it that way.

But in light of MrFreds 'last ADC' discovery we may be able to wire it up however we want.

The auto switch grounds the point on the ecu. I plan on incorporating the auto button and attatching it to a relay from the failsafe. I will set the time for switching maps to 0 secs. and increase load.

Teph the maps should switch instantaneously right?

jack of spades: would be great if it worked like that. But you would need a device to ground itself when your meth system is functional and you are not in failsafe mode.

Right now, the ECU needs to see a momentary connection to ground and THEN it needs to see the ground go away in order for the map to switch. Simply connecting it to ground doesn't switch maps. It needs to see the ground turn on, then off to do anything. I have already completely gutted the ICS switching setup and tested how it does its job. I can make my control unit mimic this with a momentary relay switch but its far more complicated than having the map switching occur the way I mentioned in the previous post.

There has to be some other wire on the ECU that does one thing when its connected to ground, and another when the ground connection is disconnected. IF we can find the code for that wire, maybe we can manipulate the auto mode input to react the same way.

Evoryder, the idea is to have the ECU use the aggressive MAPS when there is no ground connection, and use the failsafe MAPS when the ground connection is made. My control unit, along with many other alky kits have a grounding output wire which is designed specifically for triggering failsafe mapping. Thats why I am trying to see if Teph can make the stock ECU do what other alky kit users can do with stand-alone ECU's.

Im going to try some testing Jamie....i'll see if I can get this to work.

The easiest input I can think of to mimic is the clutch switch. I believe thats simply a ground/no ground switching setup in the ECU.

Alternating toggle, momentary switch can be made.

But if it gets out of step!!!

Exactly my point :D With it set to be [ NO GROUND= alky map, GROUND= safe map] its cut and dry. With a momentary switching device, it can totally become a backwards effort.


I really wish someone made a tutorial on how to disassemble and track down a feature in the ECU. Like do a step-by-step on how to find the stationary rev limit code or something.

That would be great....

guys we might use that free ADC input that MrFred found, setup would be like this:

ADC0F 5v = altmap
ADC0F 0v = normal map

I think that would provide a more stable solution, can your meth boxes emit 0 or 5V when broken/ok?

There must be unused pin available within the many pins on the ECU?

I can use a 5V voltage regulator that flows thru a transistor to be on or off depending on the triggers from the failsafes. Write up a code on a 94170014 rom with that ADC input and I can try it out tomorrow or tuesday (paid days off :D ).

Thats what Tephra is talking about. MrFred found one unused input that I believe all 8's and 9's have available, yes?

yes

I've got it nailed for the 9, but not yet for the 8. I had to make a 30 wire harness that I used to connect to all the unused pins and test each pin one-by-one. Total effort including working my way through the ROM code and prepping EvoScan was probably close to 40 hrs. My wife is ready to kill me.

Lol, my fiance always is about to kill me when I work on my car projects....then she buys me cams.

WOMEN.

:updown: ...great work so far.

If ADC channel ( if it is a true ADC) can modify fuel and ignition map based on a 0-5V meths flow sensor - we will have a truly ultimate close-loop fail safe.

The more meth you inject, the more aggressive the timing advance and more fuel can be taken out. Need to write a patch for the meths vs fuel/ignition trim %.

oh yeah I know all about the Wife (Girlfriend) factor :) hehehe kinda why I havn't done anything for a week or two :)

Nice work MrFred, is ADC0F mapped to another memory variable or is it just the raw input in the code?

We might as well make this the official input for mapchanging on the nonICS equipped evo's as well, to standardise it. That way people can wire in a switch to flip between maps.

This week I will try and get a ROM out for someone that wants to help test this new mod, once it works then I can roll it out for other ROMID's with instructions on howto wire it up. Maybe Jamie since this is your baby?

would be nice...teph

If we can locate the pins for the VIII's vs IX's jamie or myself would be more than happy to test.

Its definitely an ADC channel. I've already checked the scaling. :-) High/low map interpolation for meth content is now just a matter of some coding.

This is the most effective and reliable configuration.{thumbup}
Our meths failsafe box has a set of "voltage free" change-over contact.

ADC 0F is written to a RAM variable, but its not utilized anywhere in the code. So you have the perfect hook to take the signal and use it to control whatever maps you want.

If there is an equivalent ADC input on 8's then I think this will be a better solution than going with the ICS inputs.

This will virtually make the engine acts like a duel fuel car - with a press of a button. See you guys in tomorrow. It has been an exiting day for me.

and finding out if the 8's have the same 2 available pins or something similar

I think I will leave the interpolation for another day - getting altmaps to everyone is P1 atm :)

after getting my car fixed too :P hehe

ok for 94170015 ADC0F is 0xFFFF8928 (2 byte)

on the 8's pins 74 and 77 are free according to bez...

guys what about using a potentiometer, a clicky one so we can have more than 1 altmap? useful or not?

I'd be happy to test this out. The sooner we get this ball rolling, the better off the whole community will be ;)

Like Richard said, since flow sensors usually increase output voltage from 0-5v, down the road we can find a way to have the mapping interpolate between the 2 map variations. This would be very similar to the ethanol sensor mapping in the newer flex fuel cars. Step one, getting the maps to switch with a 5V trigger and a 0v trigger.

right on!

ok first step is to find the unused pins that go with ADC0F. I suspect they will be the same as 88590015 ROM's, but can't be sure. Do you want to use the info MrFred has provided and do some multimeter taps to see if those pins are being used?

if they are the same then you will need to rig up some 5v input and then we can log that in evoscan/ml. once we have that working then I can work on some code to allow changing maps based on this 5v input.

edit - the code might be tricky as I will be flying blind, I can't test it coz I got no car :( so you guys will have to bear with me until I get it right :)

I plan on making a 5V voltage regulator tomorrow for less than $5 anyways, no issue there. So, these pinouts have no wiring hooked up to them at all? I'll have to make a wire insert of some sort to tap into the harness locations and see if anything is going on with those locations. I'll have to see how MrFred was doing it exactly.

ok you using 94170014 yeah? for ADC0F u need to log 0xFFFF8928 and 0xFFFF8929.

I imagine FFFF8928 isn't used because 255 is just 1 byte - but log it anyways just to be safe :)

K. I might be on later but the woman is calling....might be done for the night,lol.

sure thing - just one last thing... becareful with the ECU and volts (you know this) but for others, I don't want any blown ECU's :)

lets go jack let go...lets go jack lets go

If making a regulator is easy, then sounds like a plan. Another route would be to go to radioshack and buy a 1-AA or 1-AAA battery holder. This will get you ~1.6 V which is good enough to determine which pin is the unused ADC pin corresponding to ADC 0F.

Rather than using just one wire, I wired up every unused pin. I used 22 ga wire. I stripped off about 1/2", folded the stripped part over on itself, and then pushed it into the plugs with the plugs removed from the ECU. Then I secured the wires to the harness using masking tape. This way, I was able to launch EvoScan (key "on", engine "off"), and test each wire one-by-one without having to climb under the ECU or constantly turn the key on/off to move a single wire to a different location. I tested each wire first with a voltmeter to determine if it was generating a voltage. I didn't mess with the ones that did. I probably should have then tested the resistance of the wires I wanted to hook up to the battery, but I didn't. Worked out fine though. No problems with the ECU that I can see. You might want to do it though for piece of mind.

You'll want to use the USDM Evo 8 pinout as your guide. There is a listing on the RRE website. Looks like there are only 10 unused inputs, so it should be much easier for you.

http://www.roadraceengineering.com/e...-ecuwiring.htm

https://www.evolutionm.net/forums/at...hmentid=114700

https://www.evolutionm.net/forums/at...hmentid=114701

Easy enough MrFred, its wiser to use the lowest voltage possible for the preliminary testing to minimize any damage (even 0.5v would be better). Simple voltage drop trick is to add diodes in series with the battery, each diode will drop the voltage by 0.7 volts. Putting 2 diodes in series with youe battery would be 0.1-0.2 volts, enough to make a ripple on the evoscan graph without any damage.

So it seems all i am really doing is recording the particular ADC port in evoscan and sending a signal to each empty pinout until it registers on evoscan. I can do that in the morning.

We'll have to find a way to add a pin to the ecu harness, is there ANY wire that isn't used on that harness that we can steal???? Next task will be finding a distributor that sells the required pin so everyone can legitimately add a wire that we can officially ADD to the stock ECU harness.

So how exactly am I adding these to the evoscan data setting tables so i can log these locations? I'm rusty,lol.

just like 2byte load/rpm.

find an unused pair of MUTS and put 0x8928 in one and 0x8929 in the othre... then just copy the 2byte xml for evoscan!

Ok, to be sure... I could temporarily change my MUT 00 location to FFFF8928 and MUT 01 to FFFF8929 in my MUT table in ecuflash. Then alter the 2byte load xml so it looks like this:

Or would it be best to make a seperate XML datalist item for MUT 00 and another for MUT 01 so each only has a single requestID?

I can't sleep,lol. Did some digging and found that a company called 'AMP' makes the ECU connectors for most cars. They usually have a small wire crimp connector and a large wire crimp connector. The part #'s are Large pins: AMP p/n 173631-1 and small pins: AMP p/n 173716-1. If someone wants to find anyone who sells these in very small quantities, you'll be a hero haha. Tyco and Digikey sell them but in HUGE quantities. You can request a few free samples though.

EDIT: FOUND 'EM! try www.newark.com for the crimp-on pins. $10 cents a piece...pricey,lol. I just ordered 10 connectors so i can just make my own pinout for every blank location in my ECU harness for possible future use. Total cost, $1.35 shipped :D

I can help on most electronic signals generation, just let me know.

I have also found a few pins for the EVO ECU - an old EVO project has since been abandon - looks like a possible restart. Let me know if any want to one or two to try. Just pm me your postal address.

http://www.aquamist.co.uk/forum/gallery/EVO/pins.jpg

Before ordering, check the location lugs on the wire plug, there are quite a few versions. See the red lines for the mis-alignment.

http://www.aquamist.co.uk/forum/gallery/EVO/pins2.jpg

I already ordered so I'll be the guinea pig I guess. They take 2-3 business days to get here. If they are wrong...I lost one dollar,lol.

There isn't a problem a sharp knife cannot solve.

Ahh, now that I see that updated pic in your post there must be different width pins so I think I ordered the wider of the two. We'll see I guess.

Get some sleep... is 11am here.

If you can put a 1K-10K resistor in line to your voltage source. This is to limit the current so if you make a mistake, it will not blow you ECU. ADC channels don't draw much current, so you will still see 80-90% of the voltage after the resistor.

Its not correct for USDM Evos. Emissions related sensors are used on those channels.

The low byte of ADC 0F is logged to MUT 83 for all Evo 8s and 9s that I checked. So you'll only need to hijack a MUT channel for the high byte (FFFF8928).

I'd take you up on the offer, but the pins on my Evo 9 ECU have a square profile.

Why dont ye use a patch harness for ECU testing???
they are not cheap but saves messing around with your OEM wiring loom and are available for the 8 and 9,
all the pins are wired out for testing and makes it easier to add in things like switches and sensors???
just a thought as I said they are not cheap but very handy for any future testing,

here is the link in case ye are interested also if anybody needs wiring diagrams just ask I have all market diagrams for the 8 and JDM for the 9,

http://www.roadraceengineering.com/evo.htm

Here's the catch. Its $100+ for one of those harnesses where we can just add the additional wires we need for less than 5 bucks. AND...those harnesses probably are missing the actual wire locations we need. Don't forget, the pin locations we are testing don't have ANY wires on them in the stock harnesses.

yeah I understand the price is a killer,
but all the pin locations are wired I have one of the evo 8 ones use it for bench flashing of ecu's expensive but handy for testing,
anyway just thought Id share the link for anyone interested,

Thanks for the input Barry. In our situation, its just more cost effective to just do it the good ol' McGyver way,lol. Going to test my ECU right now, I'll post up my findings soon.

Ok, so I got nothing. MrFred, when you logged with evoscan, what were all the variables set to for your data list item settings? Maybe I have it set to different settings than you did. I used the MUT 83 for the 8929 and changed MUT 73 for 8928. I got nothing on either. Only one wire showed voltage too, +5V.

Its good that you saw +5V. At least we know that the wires were making contact with at least one of the pins.

For EvoScan, I had key "on", engine "off" with EvoScan set to show the log on screen. I was watching several of the ADC channels. ADC 0F is logged to MUT 83 on your ROM (and mine). MUT 83 should read 0 until you hit the correct pin with some volts. The value displayed in EvoScan will scale linearly with voltage with the max value being 255 at 5 V. I just went wire by wire until I saw ADC 0F change. Once I found it, then I checked the scaling using 1, 2, and then 3 AA batteries in series.

Its not a guarantee that ADC 0F is mapped to an unused pin with the Evo 8s, but it seems likely. My first suggestion would be to ball up the stripped wiring more so that you're sure its making contact with the pins. Also, can you post the entry in the EvoScan Data.xml file that you used to log MUT 83?

I tried each wire at 0.5 volts thru a 10K ohm resistor, then i tried them directly connected to the AA battery like you originally did and got nothing on evoscan. I logged on evoscan the same way you did. I changed one request ID to "17" which is TPS and it read 34 while at rest and increased when the pedal was pressed so I know the code works, just the request ID's didn't seem to be with the wires I was checking.

I guess I'll have to be more certain the wires are making a connection and try again, unless you see something screwy with my evoscan xml stuff. Post up yours and I'll just copy it to my XML file.

Just tried it again, wires 'seem' to be in just fine. Still reading '0' for both MUT locations.

Here is my EvoScan entry:

<DataListItem DataLog="N" Color="" Display="ADC_0F" LogReference="ADC_0F" RequestID="83" Eval="x" Unit="int" MetricEval="" MetricUnit="" ResponseBytes="1" GaugeMin="0" GaugeMax="255" ChartMin="0" ChartMax="255" ScalingFactor="1" Notes="" />

when you apply the voltage where are your grounding the "other end" to?

I grounded it the sheet metal that the ECU mounts on. JofT did see 5 V on one pin, so it seems like he was grounding ok.

My methanol flow sensor (Labontemotorsports) has a 0 - 5vdc output proprtional to methanol flow (0 - 2000 ml/min).

How easy would it be to patch that into the ecu and log my methanol flow on LogWorks!:)

That would be very cool I think. If not cool at least very useful

Very easy if there is an unused ADC input on Evo 8s.

I grounded mine at the cigarette lighter terminal. Like Fred said, my ground source was fine since I read 5V on one of the wires.

Are you certain the MUT codes are the right ones for my ROM?????

This is annoying. The EVO8's pin profile is rectangular.

http://www.aquamist.co.uk/forum/gallery/EVO/8MR.jpg

You are so close. {pick} Which are the unused pins you are hoping the ADC0F is mapped to?

All of the unused pin locations on this chart:

http://www.roadraceengineering.com/e...-ecuwiring.htm


PIN #52 shows +5V, and either PIN 31 or 32 was actually empty (diagram says it should have a wire for A/C). MrFred, what pin location was the ADC0F input when you tested?

I noticed the previous owner had a S-AFC installed at one time and cut the Mass Airflow wire so I think I might pull that pin out and use it to test all of the empty locations to ENSURE I am getting a good contact. I wish it wasn't 30F degrees out and wet right now, its gonna make it a fun time.:rolleyes:

The Evo 9 plug layout is completely different than the Evo 8. There are three plugs and more pins. On the Evo 9, ADC 0F is mapped to pin 64. There is no pin 64 on the Evo 8.

I just had one thought. Have you tried driving the car while logging ADC 0F? This would be another way to be sure that its not connected to something already.

If it turns out that ADC 0F is not mapped to any unused pins on the Evo 8, another option is to highjack the rear O2 ADC input. It would be extremely easy to create a software rear O2 simulator to prevent the rear O2 CEL, and then that ADC input could be used for whatever.

how is it coming along jamie?

Letting the car warm up right now. Once I confirm that the pin connections are 100% good, then I know for sure that the ADC0F isn't using any of those pins. The 8's have 2 wires we can use, the ICS auto mode input and the ICS output so there just may be 2 configurations, one for the 8's and one for the 9's.

So what exactly will be needed...will you hard wire or will you be using a harness or something? If we use the auto ics pin that means we can still use the button and switch on the fly?

The ICS switches are not ADC's, so there would be no map blending with that option. Just failsafe.

It all depends how Tephra can set it up. I prefer going the 0V/5V triggering concept much better. If thats the case, it'd be best to have that wire hardwired at the ecu harness. Lets just hope we can find a way to use an open port on the VIII ecu.

I do to...just would be nice to have the option to switch maps still using the same button.

PIN 42! Thank god,lol. It registered on FFFF0x8929 only. Whew, that sucked. I literally had to un-install and reinstall the Mass Air-flow pin into each empty location. Those things are a pain to take out,lol.

I got a reading of 86 out of 255 when i gave it 1.70 volts. Sound about the same as you got Mrfred?