[BMWcurious]

Originally Posted by f32_027

From reading the information it seems to work fine as I thought the boost parameters and when meth kicks in would be controlled by the flash tune. But I guess I'm missing something.

No you are not missing something This is exactly what I am getting at. Plus ignition and fuel adjusted too.

It can be switched or variable (ie smooth or progressive) depending on what the methanol kit can output.

[proTUNING Freaks]

Originally Posted by BMWcurious

Once you add an analog to digital converter along with any other required interface hardware, and have the firmware sorted, trying to handle latency/determinism and delayed boot due to running an OS, you'd be making another JB4? The concept of an I/O expander is something I've been considering, but the Pi, as suited as it is to what it is being used for, would not be my choice.

Concept of flex fuel on a platform that has capacity to run 20% ethanol given hardware fuel system limitations is also something to keep in mind as for what it can do today. As for the Pi you're not necessarily stuck with linux either and could use an RTOS no big deal tons of ways to skin the cat, not that we're into hurting cats!

[BMWcurious]

Originally Posted by proTUNING Freaks

Concept of flex fuel on a platform that has capacity to run 20% ethanol given hardware fuel system limitations is also something to keep in mind as for what it can do today. As for the Pi you're not necessarily stuck with linux either and could use an RTOS no big deal tons of ways to skin the cat, not that we're into hurting cats!

I do love a good cat

My thoughts about the Pi were particularly in response to flex fuel and port injection where I think it is totally unsuitable in a way that those not familiar with automotive control tasks may not appreciate when they may look at the GPIO count, processor speed, memory and price only, without considering the specific needs of latency, determinism and interface hardware needed for automotive tasks.

If it wouldn't be too much work to redo your existing services to work with an RTOS, I still wouldn't trust soft RTOS tasks anything like automotive grade microcontroller hardware with phased lock loops, peripheral control processors and automotive/industrial geared timer arrays to decode toothed wheels and schedule injection pulses or even communicate with deterministic latency to the DME. By the time you've lost presumably desirable Linux features and added automotive peripherals over SPI, you may as well have an automotive grade microcontroller providing the peripherals where it can autonomously control them over an internal bus without even bothering its own CPU with interrupts.

I'd trust a Pi, especially if also running Wifi and ethernet, far less than a dedicated PIC in these circumstances, never mind something more appropriate like an S32K or Tricore.

I'm sure there is a lot of interesting stuff that could be done with a Pi that is not time or safety critical, but that isn't port injection or flex fuel. I wish there was more convergence between microcontroller and applications processors, as the amount of RAM and flash is still poor on the microcontroller side. However, when I see Linux, ethernet, an SD card interface and HDMI, garbage collection and a heap, I run a mile from wanting that controlling my engine and see latency everywhere. It all screams of user interface and is great at that.

[harkes]

Originally Posted by BMWcurious

I do love a good cat

My thoughts about the Pi were particularly in response to flex fuel and port injection where I think it is totally unsuitable in a way that those not familiar with automotive control tasks may not appreciate when they may look at the GPIO count, processor speed, memory and price only, without considering the specific needs of latency, determinism and interface hardware needed for automotive tasks.

If it wouldn't be too much work to redo your existing services to work with an RTOS, I still wouldn't trust soft RTOS tasks anything like automotive grade microcontroller hardware with phased lock loops, peripheral control processors and automotive/industrial geared timer arrays to decode toothed wheels and schedule injection pulses or even communicate with deterministic latency to the DME. By the time you've lost presumably desirable Linux features and added automotive peripherals over SPI, you may as well have an automotive grade microcontroller providing the peripherals where it can autonomously control them over an internal bus without even bothering its own CPU with interrupts.

I'd trust a Pi, especially if also running Wifi and ethernet, far less than a dedicated PIC in these circumstances, never mind something more appropriate like an S32K or Tricore.

I'm sure there is a lot of interesting stuff that could be done with a Pi that is not time or safety critical, but that isn't port injection or flex fuel. I wish there was more convergence between microcontroller and applications processors, as the amount of RAM and flash is still poor on the microcontroller side. However, when I see Linux, ethernet, an SD card interface and HDMI, garbage collection and a heap, I run a mile from wanting that controlling my engine and see latency everywhere. It all screams of user interface and is great at that.

What planet are you from?

[BMWcurious]

Originally Posted by harkes

What planet are you from?

England

I am just too excited by what is now possible with these ECUs and love to discuss new control possibilities with people planning stuff and think about how to make it work

[harkes]

Originally Posted by BMWcurious

England

I am just too excited by what is now possible with these ECUs and love to discuss new control possibilities with people planning stuff and think about how to make it work

Just very impressed with your level of knowledge on this subject. What is your bagground? Automotive engineering?

[BMWcurious]

No because I graduated in medicine not engineering.

Yes because I have been doing electronics and programming for about 35 years, tuning for 17 years including professionally, reverse engineering ECUs and adding new code to them for 12 years. Best known for my previous open source work for Mitsubishi Evo, and commercial work for Cobb on Nissan GT-R adding new features for tuners.

I sold a medical practice just over 2 years ago to retire at 40, but after 3 months got bored and decided to get back into ECU code development. Few small contracts, but mostly R&D over the last year on my own initiative and funding.

[zinner]

Originally Posted by BMWcurious

I'm sure there is a lot of interesting stuff that could be done with a Pi that is not time or safety critical, but that isn't port injection or flex fuel. I wish there was more convergence between microcontroller and applications processors, as the amount of RAM and flash is still poor on the microcontroller side. However, when I see Linux, ethernet, an SD card interface and HDMI, garbage collection and a heap, I run a mile from wanting that controlling my engine and see latency everywhere. It all screams of user interface and is great at that.

Ha I just left a position which was heavy in java tuning. It's amazing that java just stops when it performs GC and how much performance can be gained or lost by proper sizing and generation ratios.

[harkes]

Originally Posted by BMWcurious

No because I graduated in medicine not engineering.

Yes because I have been doing electronics and programming for about 35 years, tuning for 17 years including professionally, reverse engineering ECUs and adding new code to them for 12 years. Best known for my previous open source work for Mitsubishi Evo, and commercial work for Cobb on Nissan GT-R adding new features for tuners.

I sold a medical practice just over 2 years ago to retire at 40, but after 3 months got bored and decided to get back into ECU code development. Few small contracts, but mostly R&D over the last year on my own initiative and funding.

All make sense now. Awesome

[BMWcurious]

Originally Posted by harkes

That would then mean that a simple switch could be installed in the cabin grounding the pump? I can then decide if I want to go WOT with meth @ 25psi or just run pump gas @ 16psi - the DME will auto adjust. Because that would be freaking amazing!

Replying here instead of other thread to keep on topic.

I think this is a bad idea because every time you open the pump switch and then go WOT at least the first time, the DME will have to react to going lean and knocking. Why not prevent it happening in the first place and use the lambda and knock control lightly and as a secondary line of defence?

[Twisted Tuning]

Originally Posted by f32_027

What's the difference between a controller such as this:
https://torqbyte.com/collections/pro...orqbyte-cm5-lt
or this one...https://twistedtuning.com/collection...c-1-controller (which is the same).

and a JB4?

I will be using a flash tune, so if I don't need the tuning capabilities of a JB4, am I better off to use something like the above controllers?

I do not want to have any gauges or equipment in my cabin.

Simply put, WHAT did you use as your controller and WHY??

(I would like to hear from people who have experience with flash tunes and methanol injection. I'm not looking for any opinions other than on the questions I've asked. Also, I am not partial to one method of controller vs. the other - I am a blank slate and just want to hear what's best and why)


THANKS !

The Torqbye and my controller are the same for the most part (We're partners). Shortly i'll be offering mine with Bluetooth built in though to be used with their mobile app. So that will be another good thing. right now bluetooth needs an added extra module

Originally Posted by harkes

I would also like to go full flash but so far nothing else can offer what I want.

A function that is increasing boost based on meth flow is unique for the JB4, and I also want to be able to quickly switch on/off meth, as meth for everyday driving is totally overkill for me.

My problem is also that I have a rether small (1gallon) tank and dont really wish any bigger. When that run out I will have zero boost with something like the Torqbyte - with JB4, well soldier on!

Thanks for remembering me f32_027

You can Set different maps in the torqbyte/mc-1 if you wanted. thats the great thing of a flash based Meth controller. You can also turn it off if you added an Arm switch at the time of installation which is simple.

The JB4 actually doesn't add boost by itself. You actually tell the JB4 how much boost to add by adjusting the Boost additive. Then if the car isnt pulling timing, after a few pulls, it will add the boost. If its pulling timing and avg ignition goes down. it will reduce boost.

Funny thing is, the JB4 has no idea what the DME timing target actually is. Which is why the JB4 does not increase or reduce boost dynamically. it does so over time.

Originally Posted by Mike@N54Tuning.com

The JB4 is the only solution that actively monitors meth flow and uses that to adjust how much boost is added. The only solution that listens via CANbus to your throttle, pedal input, gear, DSC request from the DME, etc, and cuts off meth flow when required to avoid backfires or engine problems. To list a few of the advantages.

It's a pretty simply setup too. They even offer free back end flash maps you load via MHD and free tuning services. So you save on the map pack, custom tuning fee, and logging pack within MHD. You just need a JB4, the $249 MHD flasher license to load the BEF, and the FSB based meth kit.

Mike

the JB4 in its basic form, monitor pump current, NOT flow. and doesnt do it very well. Which is why even BMS tells you any reading over 60 in the Meth column is considered 100% because it does not really know.

The TB/MC-1 controller, the current limits for failsafes and etc can be adjusted completely for any pump you use. And the TB/MC-1 can monitor ACTUAL flow by adding a flowmeter. And failsafes can be tied into that as well.

The TC/MC-1 also can use PWM for the injection solenoids, regular injectors or PWM the actual pumps. And also can use, Throttle, Pedal, and etc. Pretty much any 0-5v source can be used as a trigger or control medium for the meth or failsafe.

BMS has to offer FREE support and Maps to run with their JB4, because IT NEEDS IT to work fully. Kinda hard to make someone pay for something that is needed for your product to work 100%

Originally Posted by harkes

Of course the Torqbyte monitors flow (well pump current actually) otherwise it would not have any safety features.
The CM5 cannot raise/lower boost only shut it off. No meth = no WOT action.
The CM5 can actually be used as boost controller but that would not work well with the DME as it would take counter measures detecting boost above target. This is where JB4 does its magic making the DME think all is well.

At the end of the day read the information from Torqbyte and see if it fit your application.

there are many ways to setup the TB/MC-1 failsafes. The main failsafe is dropping boost, because quite frankly, if there is a meth flow problem? what should you do? drop boost.

Boost control works. I have used numerous External EBC's on these cars without issue. JB4 again, Not needed.



Also, below is a couple of qoutes from another thread i was in and meth questions came up. It will expain a bit of how the DME can be tuned to protect itself with meth.

Originally Posted by Twisted Tuning

The JB4 map 3 really isn't anything special. You set the boost and the additive, and if it doesnt pull timing it increases boost. and if its pulling timing, it decreases boost.

Maybe people arent aware, but the Factory DME does this. Slightly different but same exact result. The DME can alter load targets based on a number of things. IATs, knock correction, coolant temps, and etc.

For this particular point i'll highlight the knock correction logic.

In reference to using meth injection flash only. What i normally do is tune the car with the meth system at full tilt. Which is giving the highest octane you would get from the system. Once that is done, i then tailor the DME to reduce boost (Load) as a function of knock correction. The DME does this in real time. JB4 does not. JB4 averages and needs a few pulls to reduce boost target, and also needs a few pulls to increase boost as a function of the ignition timing.

Fact is, the JB4 does not know what the timing target is. Which is why its needs some pulls to average it out.

DME knows, and does everything real time.

So, flash only lets say we tuned the car cleanly at 25psi with meth full tilt and minimal to no knock correction. Now, i program the DME to reduce load to whatever i want it to in the instance of a knock event. and how severe. So lets say a -4 or -5 knock correction happens. I can have the DME drop load to 20psi or even lower. And if you really wanted drastic, stock load targets.

Thats just some of the functionality of the DME. Hope that answers your question.


Upgraded turbos will generally need a custom tune.

Originally Posted by Twisted Tuning

sorry, I completely forgot to mention the Flex Fuel stuff. Everything I just said was Native DME stuff.


But yes, flex fuel will allow Load, timing and fuel adjustments automatically (once tune is setup properly) via an Ethanol sensor or if one is using meth they can use a flow sensor or something of sorts to be the determining factor.

Yes, you could simply put a switch for the low side of the pump to turn it off. Albeit, I wouldn't let the meth sit in the pump for too long without use.

[Mike@N54Tuning.com]

Originally Posted by Twisted Tuning

the JB4 in its basic form, monitor pump current, NOT flow. and doesnt do it very well. Which is why even BMS tells you any reading over 60 in the Meth column is considered 100% because it does not really know.

The TB/MC-1 controller, the current limits for failsafes and etc can be adjusted completely for any pump you use. And the TB/MC-1 can monitor ACTUAL flow by adding a flowmeter. And failsafes can be tied into that as well.

The TC/MC-1 also can use PWM for the injection solenoids, regular injectors or PWM the actual pumps. And also can use, Throttle, Pedal, and etc. Pretty much any 0-5v source can be used as a trigger or control medium for the meth or failsafe.

BMS has to offer FREE support and Maps to run with their JB4, because IT NEEDS IT to work fully. Kinda hard to make someone pay for something that is needed for your product to work 100%

The FSB is setup to monitor current and the 60% setting is the appropriate setting for the Aquatec pump BMS includes with their WMI kits. If you are using a different pump, you may need to adjust the value up or down. The rest of the mapping is internal so customers don't have to waste their time or money to set it properly. The JB4 also fully supports physical flow sensors, a combination of both, or nothing at all (aka the spray and pray approach you encourage), depending on the situational needs.

The TC/MC1 is a dumb controller not aware of any engine data so you'll never be able to tie in a safety for engine data like running lean, knock, fuel trims maxing out, DSC intervention, gear, oil temperature, etc, etc.

BMS offers free back end flash maps that are fully unlocked because they make their money selling the JB4 and everything else is just to provide their customers with the best and easiest possible tuning options. The JB4 is a powerful tuning device, one that doesn't cost that much more than the TC/MC1 controller, but certain things are best done in the flash map directly. Like VANOS adjustment, speed limiter delete, exhaust burble, base fuel scalar ranging, etc. Having both gives you the best of both worlds without dealing with locked flash maps. Since the fee structure is setup where the MHD flash license, map package, and logging software are all priced ala-cart the JB4 owner only needs the MHD flash license. This saves on the map package, logging package, and custom mapping expenses. Winds up costing the same or less for a much more flexible and powerful tuning solution, with better resale value, and a wider range of support.

Mike

[Twisted Tuning]

Originally Posted by Mike@N54Tuning.com

The FSB is setup to monitor current and the 60% setting is the appropriate setting for the Aquatec pump BMS includes with their WMI kits. If you are using a different pump, you may need to adjust the value up or down. The rest of the mapping is internal so customers don't have to waste their time or money to set it properly. The JB4 also fully supports physical flow sensors, a combination of both, or nothing at all (aka the spray and pray approach you encourage), depending on the situational needs.

The TC/MC1 is a dumb controller not aware of any engine data so you'll never be able to tie in a safety for engine data like running lean, knock, fuel trims maxing out, DSC intervention, gear, oil temperature, etc, etc.

BMS offers free back end flash maps that are fully unlocked because they make their money selling the JB4 and everything else is just to provide their customers with the best and easiest possible tuning options. The JB4 is a powerful tuning device, one that doesn't cost that much more than the TC/MC1 controller, but certain things are best done in the flash map directly. Like VANOS adjustment, speed limiter delete, exhaust burble, base fuel scalar ranging, etc. Having both gives you the best of both worlds without dealing with locked flash maps. Since the fee structure is setup where the MHD flash license, map package, and logging software are all priced
ala-cart the JB4 owner only needs the MHD flash license. This saves on the map package, logging package, and custom mapping expenses. Winds up costing the same or less for a much more flexible and powerful tuning solution, with better resale value, and a wider range of support.

Mike

I honestly don't recall "Encouraging spray and pray" method that you speak of. FYI, the DME has trim safeties embedded in itself. And is setup on timers and ranges that can be adjusted to Limp out when trims are maxed. That's native DME on the N54 and N55. As stated, DME natively has numerous Knock and temp safeties. Hence using the DME for it, in REALTIME. Seamless integration.

Considering how my post distinctly stated how to tune meth with the DME. I'm still unsure where you got that I'm "Encouraging Spray and Pray".

End result:

JB4 drops into a lower boost map as a failsafe (IE- drops boost)

TB/MC-1 drops boost as a failsafe

DME... DYNAMICALLLY lowers boost (load) and increases (in realtime, instantly, not over several pulls like the JB4) as a function of all sensor inputs as a safety. and is seamless integration and nothing can run the engine better.

So long story short, which is what was asked. JB4 is not needed to run meth. And you aren't Missing anything when a good controller and DME are used, unless you like the steering wheel controls.

BMS offers FREE BEF's because they have to. If they didn't, BEF's wouldn't exist. I calibrate people on all platforms using all methods including the JB4. so I know what it can do and can't do and what nuisances and good things it brings with it. And I never hesitate to tell people what those things are. But don't confuse my statements as hating on a JB4 (as most people do).

[BMWcurious]

Is JB4 logging knock on individual cylinders on MEVD17 DMEs? If so at what frequency? Does it log DME fuel trims or a difference between target and actual lambda to look for AFR error?

[gx5_027]

Originally Posted by Twisted Tuning

The Torqbye and my controller are the same for the most part (We're partners). Shortly i'll be offering mine with Bluetooth built in though to be used with their mobile app. So that will be another good thing. right now bluetooth needs an added extra module

That's good to hear - when will you be releasing it with the bluetooth addition?

[gx5_027]

Again - thanks to everyone for the fruitful debate!

It's refreshing to see people directly addressing what other people are saying, respectfully - and I'm thankful people are speaking from experience.

[hardparker]

I'm about to run an HPC progressive variable voltage controller off of the tmap as input to it with a harness patch to tap into. Plug and play simple and inexpensive.

I'll be using a tee off of the washer pump output. The WMI pump easily pulls from the washer tank through the pump. The inline pump that feeds the windshield washer nozzles barely impedes supply flow, and certainly not enough to restrict anything. It'll pour the whole tank out under gravity if you disconnect the tube from the pump output.

Concerns regarding "it'll catch fire and blow up" due to the washer tank need not fret. It's all external to the engine bay and will be running nothing but OTS WWF.

Horsepowerz here I come!!! DME, let your smarts keep me safe when I f**k up and make me power when I treat thine N55 well.

[Mike@N54Tuning.com]

Originally Posted by Twisted Tuning

Also, below is a couple of qoutes from another thread i was in and meth questions came up. It will expain a bit of how the DME can be tuned to protect itself with meth.

So the plan would be to let the DME fall in to a state of super-knock that's pulling 6+ degrees of timing advance, and then use that as the indicator to adjust load? I would think the idea is to lower load BEFORE the super knock events occur. After is too late to avoid piston damage. That is why the JB4 progressive system holds boost low UNTIL meth is fully flowing and then only raises it proportionally. So if meth never starts, drops, or stops, boost is lowered before super knock can even be registered within the DME!

Mike

[Twisted Tuning]

Originally Posted by BMWcurious

Is JB4 logging knock on individual cylinders on MEVD17 DMEs? If so at what frequency? Does it log DME fuel trims or a difference between target and actual lambda to look for AFR error?

AFAIK, the JB4 for the Fseries does not monitor timing corrections, but raw timing. How it works to increase boost is avg the ignition advance to where it "THINKS" timing is good. I say thinks, because the JB4 has no idea what the timing target is.

Im not sure what the refresh rate is.

Originally Posted by f32_027

That's good to hear - when will you be releasing it with the bluetooth addition?

once i get off my butt and get some time to finish it, lol. Hopfully within the next couple months.

Originally Posted by hardparker

I'm about to run an HPC progressive variable voltage controller off of the tmap as input to it with a harness patch to tap into. Plug and play simple and inexpensive.

I'll be using a tee off of the washer pump output. The WMI pump easily pulls from the washer tank through the pump. The inline pump that feeds the windshield washer nozzles barely impedes supply flow, and certainly not enough to restrict anything. It'll pour the whole tank out under gravity if you disconnect the tube from the pump output.

Concerns regarding "it'll catch fire and blow up" due to the washer tank need not fret. It's all external to the engine bay and will be running nothing but OTS WWF.

Horsepowerz here I come!!! DME, let your smarts keep me safe when I f**k up and make me power when I treat thine N55 well.

Ive run 100% meth before with no issues. But for customers i usually tell them to remain around an 80/20 mix of meth to water just for flammability.

I would make a separate outlet for the meth pump if youre going to use the washer tank just to be safer.

Originally Posted by Mike@N54Tuning.com

So the plan would be to let the DME fall in to a state of super-knock that's pulling 6+ degrees of timing advance, and then use that as the indicator to adjust load? I would think the idea is to lower load BEFORE the super knock events occur. After is too late to avoid piston damage. That is why the JB4 progressive system holds boost low UNTIL meth is fully flowing and then only raises it proportionally. So if meth never starts, drops, or stops, boost is lowered before super knock can even be registered within the DME!

Mike

I do believe in my quoted posts i stated tune the car with meth at full tilt so you know the limits. Then if by chance the DME see knock while meth is active it will reduce boost.

If you're not running meth for whatever reason run a none meth flash, done. Not worried about running out of meth and etc. because why? the meth controller will dump boost if meth runs low or there is a pump or flow issue. DME does the rest. ALL IN REAL TIME, not over several pulls so the JB4 can figure out if the timing is good or not.

[gx5_027]

I don't think it makes a difference, but I have a coolingmist pump and the bms 3 gallon stealth tank. I'll be using one cm-7 nozzle.

[kevinjt]

subd

[Mike@N54Tuning.com]

Originally Posted by Twisted Tuning

AFAIK, the JB4 for the Fseries does not monitor timing corrections, but raw timing. How it works to increase boost is avg the ignition advance to where it "THINKS" timing is good. I say thinks, because the JB4 has no idea what the timing target is.

Im not sure what the refresh rate is.

The JB4 monitors ignition advance in cylinders 1-6 at around 10 times per second, and determines timing stability by comparing the average actual timing against a factory timing profile which is programmed in to the JB4, and by evaluating timing drops in individual cylinders as they occur whose timing is slowly reintroduced by the DME.

That timing data is used in a variety of ways by the JB4 for auto tuning purposes, as part of the integrated water/meth safety (just because you're injecting fluid doesn't mean you're injecting enough to prevent knock), and any other internal algorithm that needs to take in to account engine knock and timing advance.

Regarding methanol safety it's a complex issue. The DME runs long term timing and fuel adaptions and even if using meth on a "pump gas" map timing can be advanced by the DME, and when that octane and fueling is suddenly removed, quick knock occurs the DME is not always able to properly deal with, and pistons can be broken. This was a common issue with the N54 in 2008 using the JB3 system and non-integrated meth kits before BMS introduced CANbus and their WMI safety integration systems. Running WMI without a properly integrated safety system is just playing Russian roulette with your engine. Running less aggressive tuning can reduce the risks but who installs a WMI kit to run less aggressive tuning?

Programming the flash tune to react to knock with load changes is too slow of a response to be effective in this regard. If you can run a flow sensor input in to the DME then that would be the best way to do it "flash only". I use quotes because you're still using a piggyback for the WMI and flow sensor control. Just not a JB4 piggyback. But like most features implemented "flash only" it becomes a compromise solution. Kinda does the same thing, but not really. Devil is always in the details.

Mike