Tell me why I shouldn't go AEM

[kawgomoo]

to the OP, link will be all you need. it's been proven time and time again to gain big hp. unless you're building an all out race car and are making your living at the track, then stick with link. AEM im sure is good for squeezing out the last .01 hp, and really, why would someone need to squeeze out that last bit unless they are building a dyno queen and want bragging rights. i am by no means a tuner, and i am anxious to get my hands on v3lite (thats all i'll need). it sounds like AEM is way more than most people need.

this is the most ignorant statement ever.

if you can tune link, you can tune aem or squirt or anything else.

if you want to run speed density there is some bullshittery involved with the link system making it more of a hassle to tune.

if you set the AEM up for Boost Comp {google it} then the tuning cant be easier. Literally one hour on the dyno to sweep in the boost map. then a little driving around on the street.

What makes people think because you have a link means the tables that are locked out are somehow perfect and dont need to be tuned?

The problem with link is it costs AS MUCH as a good system, and OFFERS LESS.

Eeking out the last .01hp is not my goal. But having a car i can hop in, beat the shit out of day in and day out and never have to ####ing work on it. Thats what im about.

Enjoy your linked up crapbox, catch ya on the streets.

 

[ramsack]

The DIYPNP come unassembled as far as I can tell. CPU and RAM aren't really a big factor. The bits are how the ADCs and DACs translate between analog and digital. More bits, more resolution. 10-12 bits is plenty. 16 Bits is overkill. So you would need a 16 bit CPU, and pad unused bits with zeros, and a little more RAM as a buffer. I really don't see the need for a crapload of sensors. All one needs is crank position, throttle position, a MAP and IAT (or any other air-metering device), an o2, a knock sensor, and coolant temperature sensor (not really that important.) The only outputs needed would be spark (2 or 4) and fuel injection (2 or 4) The knock I would think takes most of the toll on processing power. Any other sensors are pointless and a separate datalogger would be more useful because it will be made for it. Things are already mapped so there is minimal calculation a computer has to do, it's basically just looking up info, seeing what needs to be done, and doing it. As I said the DSP and ADC/DAC are the biggest things to improve, but only really driveability, not so much peak power.

 

[abaum565]

Enjoy your linked up crapbox, catch ya on the streets.

All of those informative statements and you take a shot like that

We know you don't like link, but have some respect!

 

[kawgomoo]

The DIYPNP come unassembled as far as I can tell. CPU and RAM aren't really a big factor. The bits are how the ADCs and DACs translate between analog and digital. More bits, more resolution. 10-12 bits is plenty. 16 Bits is overkill. So you would need a 16 bit CPU, and pad unused bits with zeros, and a little more RAM as a buffer. I really don't see the need for a crapload of sensors. All one needs is crank position, throttle position, a MAP and IAT (or any other air-metering device), an o2, a knock sensor, and coolant temperature sensor (not really that important.) The only outputs needed would be spark (2 or 4) and fuel injection (2 or 4) The knock I would think takes most of the toll on processing power. Any other sensors are pointless and a separate datalogger would be more useful because it will be made for it. Things are already mapped so there is minimal calculation a computer has to do, it's basically just looking up info, seeing what needs to be done, and doing it. As I said the DSP and ADC/DAC are the biggest things to improve, but only really driveability, not so much peak power.

This is pretty much all true, and admittedly way over my head. The sensors triggering inputs dont really use much processing power.

The process of how your car determines fuel trim however, does require some thought. First you cross refference speed and load {rpm vs psi}

Whatever cell you are closest too lights up. Now the computer must think..okay cell 5500rpm, 15psi is lit.. but the motor is really at 5608rpm and 15.8psi of boost. So now the computer has to refference all the cells near the cell block it is currently "using". A calculation is made based off the slope from cell to cell in the nearby blocks.

This "slope" is applied to the base fuel table. a rough guesstimate if you will. So lets say at 5500/15psi you had 158 as your fuel map number. But being slightly higher in load and faster rpm makes the computer need to guess high, basically holding over. The word here is interpolation.

It then takes this raw number of 158 and compares it to all the other sensors. Is the coolant temp hot or cold? it then trims tihs fuel number accordingly for engine temp. Ok next, how about TPS? was this a quick stab of the accelerator or gentle roll in? This would be accelerator pump enrichment. you need a little shot of fuel to overcome the momentary lean created by rapidly opening the throttle door for example.

and so it goes, on down the line... this is why i feel a computer does need a fair amount of processing power. remember each of these decisions has to be made twice cycle per cylinder. once for fuel, once for spark.

There are subtle variables people just getting into tuning may not realize. So choosing a system for them is that much more difficult. How do you know what you need, if you dont know what you need?

Learning AEM is certainly a nightmare. But tuning really is pretty intuitive. Make a change, if it works better keep it. if it makes it worse, change it back. The Biggest problem with AEM is the TERMINOLOGY used within the software. They speak there own language. And the terms are not necessarily descriptive of there function.

If you want to take an aem, stick it in your car and ONLY fuel tune it. {basically treat it like a link} you could absolutely do so. Most of the tables {start enrich, crank pulse, etc} are pretty close and very linear. If one gives you a fit you just highlight the whole line then move it up or down until the car behaves like you would expect.

One of the big problems is battery offset. When you have a link, many tables just arent there for u to play with. do you think all these tables remain the same even though you have changed every part of the engine?

why would tuning the fuel base map alone be enough? what about dialing in the TRIMS so they actually have the proper authority for YOUR engine setup.

In the end. Tuning has come along ways since afc's and fmu's. People are realizing the KEY to unlocking performance is the tune. You can buy all the high dollar parts you want, but if the tune isnt spot on they will all end up a pile of molten crap in your oil pan.

So to me, it seems just plain stupid to buy a link for any reason. It is a handicapped system. To say "i will never need the hp" well why are you trying to make your car faster if you dont care about extra power?

if its a daily driver shouldnt you want to be able to make it run perfect. absolutely flawless? My big thing is a car that requires no special knowledge to operate. If it runs 15's or 8's i want to be able to toss my grandma the keys and have her take it to the store with zero issues or questions.

To me AEM offers the highest level of tuning authority and automation of your system, while still offer a manageable level of simplicity if you only want to use its most basic functions.

I dont think the Link is a BAD system, i just think it is HORRIBLY over priced and under featured for its price range.

A very nice AEM feature that probably gets underused in the dsm world is the ability to drive your Fuel Pump off a spare PWM output. This lets the aem control your pump speed like a 300 dollar aeromotive controller.

It also helps keep your battery up, because as we all know dsm have marginal charging systems at best. im lucky if i can hit 13volts with my headlights on fuel pump running.

So you will have to excuse me if i take cheap shots at LINK. But as an american, i am free to do so.

I think above and beyond everything else you should SHOP WITHIN YOUR BUDGET. Dont be another DSM partout.

But for me the choice is simple. If {and when} i cannot afford to use an AEM computer i use DIYPNP products. To me they are very similar {except for knock, MS knock sense is a joke aka completely useless} But on that note i dont tune using a knock sensor..so who cares. Knock sensor is a tuning crutch, but useful on a stock turbo car that may very well have careless owners who are even less careful about there choice of fuel.

 

[Ceddy]

One major thing that the Mitsu OEM based EMS (DSMLink, EcuFlash, Jackal, etc) have that Standalone EMS don't is the Knock Control System.

Standalone EMS tend to try to integrate the Mitsu Knock Sensor by simply looking at Knock Voltage. This isn't adequate, you can have zero knock at 5 volts and 30 knock counts at 1 volt.

A Standalone company will never catch up because of the vast resources Mitsu pours into this one system: multiple patents, millions of dollars, worlds top engineers, etc.

There are many things that a universal EMS can't take into account the OEM ecu already has factored in like engine harmonics and resonance, sensor placement, engine bay noise, and a bunch of other stuff you need a PHD to understand.



Another thing I like about Mitsu's system is they use actual AFR in the Fuel Maps, you can look at a cell and see you have 11.0 AFR there. Other systems use Injector Pulse Time, and in my opinion are a pain to tune.



Someone said above about AEM's processor/hardware being superior. While this may true of the 20 year old 1G ecus, the newer ecus in the Evo8-10s will blow AEM's cpu away. But most of the resources in the newer cpus go unused, they are used because Mitsu is keeping up with the MicroController industry. Like said above the 1G ecu is good enough to power 1000 HP cars.


I'm not putting AEM down at all, its a great EMS. Just telling some of the things I like about Mitsu's ecus.

 

[kawgomoo]

If you are tuning afr in the fuel map...you are not tuning the fuel map at all.

you are tuning the O2 target map.

when using o2 correction to keep the car in tune, the changes are always coming several cycles after the event. make sense?

mitsu has huge resources. true.

we are in a hugely down economy. also true.

mitsu is not building "THE BEST" they are building to a price point.

Not saying aem is the best..by any means.

in the scheme of standalone EMS, AEM is bargain basement.

But aem DOES plug n play, taking much of the hell out of traditional real race car shit.

I think the aem does the best job at bridging the gap between real race car gear, and street car usability.

many of the "street" tuning solution are called such because they are just poor coverups, and under extreme conditions the flawed tunes would expose themselves by melting down motors.

I am NOT AT ALL against TUNING THE STOCK ECU!

In many many ways i think it is the best solution.

But if you want to tune the stock ecu, then TUNE THE STOCK ECU! DONT BUY A ####ING LINK!!!!!

Get an evo8 ecu... cheap on ebay. Call DIYAUTOTUNE and get a n52 or n76 nippendenso connector or break out box. build your adapter cable. poof, done! Or maybe someone already sells evo8 to dsm pigtails? i have no idea! you could even steal the socket out of one of ## old dead ecu's and build the pigtail for free!!!!! need the wiring diagram? shit me too....

now go chip tune your stock shit all day long!!! woo hoo!!!! the evo8 ecu has ALOT of nice features. and can be had for like 100 bucks.

So again. i repeat. if you want to tune the stock ecu. then do so. invest in a emulator and chip burner..or just get an emulator. when you want a chip burned email me the file and ill burn it on an eprom and send it to ya. for free!

just promise me you wont spend a thousand freaking dollars on a link!!!!

 

[zatomz]

I dont know how much I'll help here but I am putting an AEM EMS in my car right now, so you know where my .02 are. They have a plug and play for 92-94 T/E/L and that's what I got. Will I plug and play? NO, but I will plug and spend my money to dyno tune my car because I have never had it done before and I am excited to be there and learn something. This is a hobby, for me anyways, why not take the "Scenic route"?

 

[delta448]

kawgomoo,
I can certainly appreciate your self-taught approach and I agree with some of what you've said... however there's more than just one direction in which to approach things and most people simply aren't patient enough to spend the next year or two teaching themselves how to produce a good calibration on a more complex standalone. -Especially if they're less experienced and don't care about becoming a great calibrator, but just want to be going fast right away. We live in a world of immediate gratification, and for that the K-I-S-S motto works. These are the type of people that Link works exceptionally well for. They get a very popular and solid quality product with plenty of web support on an easy to learn interface with which to discover/teach themselves the basics of how to tune. That's the beauty of ECMLink, you don't have to be an expert to get good enough results.

The Evo 8-9 ECU swap is another can of of worms all itself, which I think you might misunderstand a bit. I previously used one in my 97 AWD, I found it to be a more challenging interface due to the extensive use of code -albeit with abilities limited only by the user's understanding. Flash tuning with lots of required datalogging is probably not nearly as quick as on-the-fly ostrich tuning, but the Evo ECU as far as I know is a Flash-only interface right now... someone correct me if I'm mistaken.

 

[ramsack]

I don't see why one couldn't take out the knock circuitry of a stock computer and put it inline between the sensor and a standalone, or even integrate it into a MS. Then you get the signal you need. If I'm not mistaken, is the knock DSP circuit not one of those extra boards that sits atop of the motherboard?