Pages: 1 2 [3] 4 5
Author Topic: Why cant motronic controlled cars be fast?  (Read 39441 times)
vwaudiguy
Hero Member
*****

Karma: +53/-37
Offline Offline

Posts: 2024



« Reply #30 on: June 05, 2017, 08:47:27 AM »

Wow you guys are still yapping about a post made 2 years ago, and on top of that this troll thread?

In the past there was a point where the majority of ME cars just weren't fast... that is a fact.  They made numbers, but never translated on the street.

Regardless of what one thinks, there is a difference between making a pull on the dyno in a single gear and racing.

Apparently it's more important to talk shit about posts on other forums and believe there is no point to do something unless you need to use anything else except pump gas.

Having a bad Monday?  Smiley
Logged

"If you have a chinese turbo, that you are worried is going to blow up when you floor it, then LOL."
NOTORIOUS VR
Administrator
Hero Member
*****

Karma: +58/-7
Offline Offline

Posts: 1056


« Reply #31 on: June 05, 2017, 11:51:03 AM »

Having a bad Monday?  Smiley

Well it is Monday, and it's dark/rainy out... so maybe  Cheesy
Logged

SCHNELL ENGINEERING BLOG ·  STANDALONE ECUS · TUNING · DYNO · WIRING · PARTS · VEMS
Google Talk: NOTORIOUS.VR
n00bs start here: http://s4wiki.com/wiki/Tuning
nyet
Administrator
Hero Member
*****

Karma: +604/-168
Offline Offline

Posts: 12236


WWW
« Reply #32 on: June 05, 2017, 12:01:34 PM »

Bottom line, there are some things that are harder to do in ME than standalone.

It doesn't mean it is impossible.

Still, the AZ criterion of "Motronic based cars are slow, so Motronic sucks" is moronic.
Logged

ME7.1 tuning guide (READ FIRST)
ECUx Plot
ME7Sum checksum checker/corrrector for ME7.x

Please do not ask me for tunes. I'm here to help people make their own.

Do not PM me technical questions! Please, ask all questions on the forums! Doing so will ensure the next person with the same issue gets the opportunity to learn from your experience.
Jim_Coupe
Hero Member
*****

Karma: +16/-12
Offline Offline

Posts: 663



« Reply #33 on: June 07, 2017, 02:19:32 AM »

interesting reading anyway Smiley I liked this thread Smiley  (I am a disable everything guy.. for my own cars) There should be a thread on how to make an Me7 really stupid and simple... Just ignition and fuel kinda thing..
Logged

E85oholic
NOTORIOUS VR
Administrator
Hero Member
*****

Karma: +58/-7
Offline Offline

Posts: 1056


« Reply #34 on: June 08, 2017, 10:12:07 AM »

Bottom line, there are some things that are harder to do in ME than standalone.

It doesn't mean it is impossible.

Still, the AZ criterion of "Motronic based cars are slow, so Motronic sucks" is moronic.

Let be totally honest here, everything is harder to do in ME vs. standalone.

The plain fact is engines will run on absurdly bad programming and inputs quite easily unless you're totally off.  Not saying that it's good but I've seen it many times in the past where someone gets their car tuned and it runs, heck it even makes power, but it doesn't drive properly at all.

At the end of the day, ME has a larger hold on the power output than most will ever know or admit.  Of course it's not IMPOSSIBLE, but at the end of the day not many understand exactly how deep the reach is for even the most basic control principals.  90% of "tuners" simply manipulate the ECU just enough to keep it happy for what it expects to see from it's 1000's of parameters programmed by Audi/Bosch for a docile every smooth car with a warranty since that is the easiest route to take.

A car running standalone vs a car with an ME tuned like described above will drive, and feel completely different even though they could make roughly the same power on a dyno.
Logged

SCHNELL ENGINEERING BLOG ·  STANDALONE ECUS · TUNING · DYNO · WIRING · PARTS · VEMS
Google Talk: NOTORIOUS.VR
n00bs start here: http://s4wiki.com/wiki/Tuning
nyet
Administrator
Hero Member
*****

Karma: +604/-168
Offline Offline

Posts: 12236


WWW
« Reply #35 on: June 08, 2017, 11:21:48 AM »

At the end of the day, ME has a larger hold on the power output than most will ever know or admit

For 99.99999% of hardware setups this is plainly not true.

There are literally only 3 things

fuel
timing
boost

Getting ME7 to do any of those the way you want to is not a problem until you run into MAF/load limits.

Quote
A car running standalone vs a car with an ME tuned like described above will drive, and feel completely different even though they could make roughly the same power on a dyno.

I have yet to drive standalone car that didn't have pure shit part throttle.

"But it is a track car" people say. Drag strip, or TRACK car?

Because even a track car has to be controllable part throttle, even in an AWD car, or you are just killing tires.
« Last Edit: June 08, 2017, 11:24:17 AM by nyet » Logged

ME7.1 tuning guide (READ FIRST)
ECUx Plot
ME7Sum checksum checker/corrrector for ME7.x

Please do not ask me for tunes. I'm here to help people make their own.

Do not PM me technical questions! Please, ask all questions on the forums! Doing so will ensure the next person with the same issue gets the opportunity to learn from your experience.
Mechsoldier
Full Member
***

Karma: +5/-11
Offline Offline

Posts: 106


« Reply #36 on: June 10, 2017, 01:16:20 AM »

For 99.99999% of hardware setups this is plainly not true.

There are literally only 3 things

fuel
timing
boost

Other than the processing speed of the ECM. I think a LARGE part of the reason cars are putting more power out now is how fast they can monitor sensors. I think I saw ME5 monitors the O2 10 times a second, ME7 does it 100 times per second, it allows you to push things a little further without having to worry about a condition that can cause engine failure without the ECM catching it.

I have yet to drive standalone car that didn't have pure shit part throttle.

"But it is a track car" people say. Drag strip, or TRACK car?

Because even a track car has to be controllable part throttle, even in an AWD car, or you are just killing tires.
And that's what my very long, very rambling post was about. You CANNOT get the throttle control from standalone you can get from a the ME7 without tons of background logic they just don't build into them because track guys don't need them. You'll NEVER build even a 450 hp car on stand alone you'd feel comfortable putting your grandma behind the wheel of and sending her to the store, period.

NOW, I did find something interesting. Anybody ever hear of these guys?

http://pi-innovo.com/openecu-module-comparison/

They make ECMs that are designed with rapid implementation in mind. They use a very easy to use interface. Their controllers can be used to control nearly anything from DSG trans, or even using solenoids to make a standard mt an AMT. 

They've been around 20 years, I ran across them looking for hybrid motor controllers, and found they had done a very interesting project in conjunction with some UK companies to take and replace the factory Bosch management because they didn't want to hack it, and they put their ECM in the car, changed the hybrid strategy, and reduced CO2 emissions 40% over the emissions drive cycle, and fuel economy went from 42 mpg, to 64. I always fealt that the VW hybrids got garbage fueling...not sure what changes they made, but PI Innovo's software meets all federal and world wide environmental standards.

They support can and linbus, they have built in OBD2 and diagnostic ability.

I wouldn't even call them stand alone, they're open source OEM. They have worked with Landrover, Range Rover, and currently make the Mk2 City Golf ECM for VW in Africa. Cool stuff.

https://connect.innovateuk.org/documents/3030546/3708203/H4V_OVERVIEW_FV_DPCHARTERS.pdf/e8018002-9c94-4ecd-b187-f4ac41135c8d

The ME5 series had ONE 16 mhz processor, so could do 16 million computations per second, and sampled the O2 10 times per second ( I can no longer confirm but believe I was taught in VW tech training)

Our Me7 ECM is a 32 bit system from 2 16 bit chips running at variable clock speed from 33 to 43 but the official speed is 25 hz, and looks at the O2 sensor 100 times per second, doing total of about 25 million computations per second up to 43 mhz.

The C167CR processor is listed at the lower end of processors on Infineon sites, it was specified for like ABS modules actually, but MOST of the better chips don't fit this socket. It looks like there WOULD be an upgrade but only to like 40 HZ.

Pi-Innovo's M670 runs at 256 mhz, it has a sample resolution of up to .001 Seconds. You can literally check the O2 and knock sensors 1000 times a second. Makes me wonder if you could even use standard narrowbands since the resolution is enough to check the frequency so much. Regardless the Pi Innovo stuff seems pretty great, easy to use, and it probably costs less than Maestro does  Grin

They support direct injection, automatic manuals like the DSG, and much more. It's basically big enough you could run an ABS system off of it too.

This WOULD be a serious upgrade, something that can monitor the knock sensors, maf, O2, AND uses factory strategies like ME7 or something? Plus it passes readiness and stuff, would be basically undetectable even in Cali.

For example, and I've gotten a lot of shit for this but I don't care, it's tested. I have an engine builder who makes like, streetrod engines for 30 grand, does Viper engines, lots of race shit, and he told me to lean it out to 13 on my sisters mk4 GLI 1.8t which ALREADY had the APR stage 2 tune and it RAPES it. I don't have figures YET I'm doing the breather system then Im going to dyno it with the stock, APR 1 and 2, GIAC, and other tunes I've backed up to check mine against them.

I would never tune a customer car that lean, it's pushing the limits and has very little safety cushion. But if I have to drive through death valley and it pings I'd flash it, so it doesn't scare me. I ALSO bump the timing so that it's pretty much always pulling 5 -7 degrees of timing without detonation, I've shifted the octane requirements so that now I could run 96 or 99 octane it would adjust automatically, 91 is now like running 87 from the factory.

The resolution isn't enough in the ME7 ECM though and I've HEARD detonation while the ecm sees none. With the PI Innovo even you could even change the trigger signal from the knock sensors that initiates timing retard and all that good stuff.

Getting mounts, or other vibration on a heavily modded car could be easily tuned around. The limitation I think we face is the processor speed and thus the proper or BETTER resolution from the sensors. The resolution you could get from 25 BILLION computations per second would allow the diagnostic programming to pick up on misfires better and faster because a better waveform to analyze, likewise a maf that isn't reading right, or when you get a detonation or hickup the better signal res could even help you sense detonations that way.

This shit looks pretty awesome, I'm stoked that it does FSI and stuff as well, with fuel boost circuits, quad VVT control, and much others. I BELIEVE it says you get unlimited phone support too.

I looked at some stand alones. AEM EMS-4 is like a 32 bit 50 Hz processor, and their BEST kits are 200 or so. Megasquirt's NEWEST kit is only 16 bit. Haltech doesn't easily display the exact speeds but from looking at their quoted speed for sensor resolution it looks to be about 50 hz for the cheapest ones, up to 200 hz.

So it appears processor speed IS really important, and the factory setup is at least partially limited from that. I mean, think about it, the goal is to burn as lean as possible, since lean is fast, while also keeping the cylinder and EGT temps in check. And ideally if our computers are good enough, then you stay at the EXACT border of that threshold. IF you can't check the sensors enough you can't do that, then you risk crossing that threshold faster than the ecm knows and corrects. I mean, ideally if you could run it so that the temps get a little higher than you'd like for a nanosecond then pull it back to cool it slightly, you'd be golden.

Only the highest end stand alones for 2k or 3k do that while also having torque management profiles, obd, etc. I think Pi- Innovo is going to be way cheaper and more intelligent, it clearly works if they're doing things for busses, oems, etc. Their ecms are priced at OEM price level, I can't imagine it being any more than a grand for the highest level. As the site shows it requires some knowledge. I'm interested.


Logged
Mechsoldier
Full Member
***

Karma: +5/-11
Offline Offline

Posts: 106


« Reply #37 on: June 10, 2017, 01:30:41 AM »

interesting reading anyway Smiley I liked this thread Smiley  (I am a disable everything guy.. for my own cars) There should be a thread on how to make an Me7 really stupid and simple... Just ignition and fuel kinda thing..

It's a really BAD idea, read my couple posts. The Honda guys do that and blow shit all the time. A quick sum up, lean is power, ideally lean and max ignition timing right up to just before detonation, and EGT/valve temps acceptible.

Removing the safeguards essentially means you can't push the tune to the limit, because you could blow up before the ECM notices knock. IF you have different motor mounts the knock sensor signal may even be so different the ECM doesn't recognize it as knock. IF your breather system and your MAF are correct, you can push the limits much further. For example a 1.8t won't pull much or any timing on 91 octane, but if you read the pull in value block 21 with VCDS it will pull 3 - 5 degrees timing to correct. My tunes pull 3 - 5 on 91 or 92 octane. But when the ambient temp and humidity are right, or I add octane, it automatically increases power.

If I were to remove the knock sensing, I'd never be able to do that because it would detonate all the time. I have about 20k miles on my tune like that, no issues, at a MAX of 13:1 AFR under full boost 100% load or whatever, and I'm running 13.5:1 or 14:1 at 65%, 75% load from 2500 to redline.

If I turned anything off, I'd be fucked. And it makes a LOT more power. IF your MAF doesn't read right, and your car isn't on point it won't run it. These are the reasons why companies like APR are known for weak tunes. I have the Alientech Setup, the legit. And I know guys who are the tuners for C2, APR, JHM, and Addict Motorsport here in Washington, and the remote tuners ALL have told me they can't push it further because the tunes wouldn't run on some cars. Addict related that he also fixes EVERYTHING wrong first.

I can't even put my tune on 5% of the cars I try it on, if the breather or maf are fucked,there's oil up under the intake and stuff, I put it on and it's the exact same. Then they fix that crap and it hauls. The tuners don't want that bad PR," Now they say I need $600 in breather stuff and a maf, fuck them!" etc.

The whole tuning stages thing is kinda bunk and imho to sell parts. I've taken a backed up APR stage 2+ from my car and put it on a STOCK 1.8t, and it runs FINE, and honestly isn't as much less hp as the car with the downpipe and stuff too...

If you have to turn shit off your car isn't right. Maf reading should be 3.5 grams per second at idle and when sitting still at 2000 rpms should be about 10.5 or 11%. If you're at 2.5% and 7% the maf is bad (or the breather has internal vacuum leaks sucking from the intake boot, or the valvecover, or the block....
Logged
IamwhoIam
Hero Member
*****

Karma: +48/-105
Online Online

Posts: 1053


« Reply #38 on: June 10, 2017, 03:38:16 AM »

really funny last few posts here... ME5 does not exist and has never existed. Knowing exactly what you're talking about is really important, before you start dumping a complete logorrhoea that hundreds of n00bz will be reading and taking for absolute truth.
Logged

I have no logs because I have a boost gauge (makes things easier)
_nameless
Hero Member
*****

Karma: +327/-456
Offline Offline

Posts: 2734



« Reply #39 on: June 10, 2017, 06:39:00 AM »

 Tongue m3.83, m5.92
M= Motronic
E= Electric pedal control
...
 Tongue
Logged

If you are in the market for a tune and would like the ease of downloading and flashing a dyno tested tune for a fair price check out https://instatune.sellfy.store/
Mechsoldier
Full Member
***

Karma: +5/-11
Offline Offline

Posts: 106


« Reply #40 on: June 10, 2017, 04:39:54 PM »

really funny last few posts here... ME5 does not exist and has never existed. Knowing exactly what you're talking about is really important, before you start dumping a complete logorrhoea that hundreds of n00bz will be reading and taking for absolute truth.

I didn't bother putting in the exact ECM number big whoop. Point was that we've gone from 8bit in digifant, to 16bit in the mk3 ECM, and the Audi throttle cable ECM
And now we have 32 in the mk4.

Point is processor speed DOES matter. The rest of the info is valid.
Logged
eliotroyano
Hero Member
*****

Karma: +47/-8
Offline Offline

Posts: 807


« Reply #41 on: June 10, 2017, 06:35:52 PM »

Friends I was reading and I am impress of the amount of knowledge many people have. From my point of view we are controlling an analog system (engine, gearbox, etc....) with a digital system (ECU, TCU, etc....), so processing speed is important , but how the analog system reacts to the digital one is what really matters. Then a "sufficient capable ECU" can control an engine, gearbox, etc... in terms of how many variables it can act to obtain the desired "control". So the main difference between OEM vs Standalone ECUs is just how much knowledge we have about it. This discussion remembers me CISC vs RISC computational systems in the past. As a simple novice is just my point of view.
Logged
Mechsoldier
Full Member
***

Karma: +5/-11
Offline Offline

Posts: 106


« Reply #42 on: June 11, 2017, 10:13:00 AM »

Friends I was reading and I am impress of the amount of knowledge many people have. From my point of view we are controlling an analog system (engine, gearbox, etc....) with a digital system (ECU, TCU, etc....), so processing speed is important , but how the analog system reacts to the digital one is what really matters. Then a "sufficient capable ECU" can control an engine, gearbox, etc... in terms of how many variables it can act to obtain the desired "control". So the main difference between OEM vs Standalone ECUs is just how much knowledge we have about it. This discussion remembers me CISC vs RISC computational systems in the past. As a simple novice is just my point of view.
Kind of. Like I wrote you basically gain no advantage unless you spend the big money for the higher processor speed.

And the ECM gets analog signals from sensors too.

But as I was saying, the more you can push the file ignition timing and learning the fuel out to the threshold of cylinder temp or egt being out if wack and no detonation, the better.

Because of the slower processor speed you can't do that because it doesn't sense detonation fast enough. As we increase the processor speed to 256khz you're now doing 25,600,000,000 computations per second instead of 25,000,000. It means you can push the shit out of it. The 2.0t runs like 16:1 at hwy cruise.

The systems with processor speeds fast enough to justify the upgrade are thousands of dollars.

Me7 does good enough for me.

Sent from my SM-G955U using Tapatalk
Logged
prj
Hero Member
*****

Karma: +1035/-438
Offline Offline

Posts: 5884


« Reply #43 on: June 11, 2017, 01:31:33 PM »

Because of the slower processor speed you can't do that because it doesn't sense detonation fast enough.
You probably shouldn't talk about that when you don't know difference between mhz and khz...

Also, processor speed has not mattered for knock detection for about 20 years now, due to how the algorithm works and how the integration is done over the signal in a separate DSP.
It baffles me how someone can type up so much bullshit without having any actual idea how things work.
« Last Edit: June 11, 2017, 01:33:25 PM by prj » Logged

PM's will not be answered, so don't even try.
Log your car properly.
Mechsoldier
Full Member
***

Karma: +5/-11
Offline Offline

Posts: 106


« Reply #44 on: June 12, 2017, 12:24:08 PM »

You probably shouldn't talk about that when you don't know difference between mhz and khz...

Also, processor speed has not mattered for knock detection for about 20 years now, due to how the algorithm works and how the integration is done over the signal in a separate DSP.
It baffles me how someone can type up so much bullshit without having any actual idea how things work.

Coming from the guy who essentially claims that motronic can sense knock as well as a modern Tricor ECM?

I gotta go, I'm trying to get in contact with the engineers at TI and tell them not to bother with this since it's not a problem and there's no gains to be had trying to reduce the load on the IC.

Logged
Pages: 1 2 [3] 4 5
  Print  
 
Jump to:  

Powered by SMF 1.1.21 | SMF © 2015, Simple Machines Page created in 0.034 seconds with 17 queries. (Pretty URLs adds 0s, 0q)