T-leg Piggyback ECU R&D

[93forestpearl]

wheeeew!!! Just read the whole thread. I am definately interested as this is sounding better than name brand piggybacks, besides being much cheaper.


For the most part, I've been able to follow what you guys are doing and talking about. I used to be an electrical engr. major. Took a microcontroller class using the 68hc11. EE wasn't for me so I switched to mechanical.

The "LegaCU" or whatever sounds exactly like what I'll be looking for come fall.

Dan

[free5ty1e]

V - are we just disagreeing on what units to be adjusting the trims in? I envisioned the software allowing the user to select which units to tune with, giving a choice of:

ADC value (which is the index for pulling from lookup tables)
voltage
converted MAF reading (g/s)

Then we'd have to make runs anyway to see what results were yielded from the last batch of settings. With a wideband O2 sensor, the software could even identify which RPM/MAP data points were overly rich or lean, and the next tune can begin from there.

I'm not sure where our miscommunication comes in, but "pushing it through the transfer function" to me means pull the value from the lookup table in g/s (or CFM) and use that to tune. Right? If that's the case its just a matter of selecting those units for tuning, whichever units are selected it will still be the same range and step size for allowable trims. It'll just be pulling readings from a different table in ROM for the display; the index value would be the same.

Dan - the 68HC11 was a pretty pointless class imho, I remember having to take that. Nothing like making sure the new engineers learn how to utilize obsolete technology, huh?

Thawa - good find, I'm sure that info will come in handy at some point. Never hurts to keep all the info in one place that might even be remotely related.

Nico - the laptop interface will be an option for later, for easier tuning and run analyzation. Shit, I might even take some of what that street dyno program does and implement that - screw paying for dyno time, that can be calculated and graphed with the inputs we'll already have access to in this product. But anyway you will be able to navigate menus and tables using the buttons and LCD, no laptop will be needed to tune. It'll just make it easier.

[vrg3]

Yeah, I guess one way to look at it is that we were disagreeing on the "units." I just feel that the only way to go is the trim the converted MAF reading.

Yes, "pushing it through the transfer function" means reading/interpolating out of the lookup table. And then pushing it back through basically means a binary search of the lookup table and one interpolation.

But you couldn't use a PGA to do this trim, right, since you can't convert between volts and g/s in hardware?

Yes, with a wideband sensor you can very effectively provide feedback for adjusting the tables. Even the narrowband sensor can provide some not-useless info. With a wideband you could even try to make it auto-tune the fuel trims based on a table of air/fuel ratios, but that's probably not really worth it since it'll affect ignition weirdly, most tuners will want to control things anyway, and it'd be more responsibility / liability.

[93forestpearl]

free5ty1e-

Yeah, it was a pretty lame class. That seems to be the theme here, classes can be way outdated, besides the fact that many of the professors have trouble speaking english. We worked on pretty dumb assignments like de-bouncing a key pad and crap like that. It was cool though, to understand that these microcontrollers make engines look really slow.

I'm liking mechanical engineering alot, even though I switched in my fourth year We are trying to get a Formula SAE program rolling here on campus to give our ME program a focus and to get the community involved. Its hard to get sponsors when you don't have previous iterations of the car to show.

Keep up the great work.

-Dan

[free5ty1e]

I feel your pain, Dan - my years at UCF were not really up to par with what I feel I paid for. Professors rarely taught their own classes, sending TA's who could barely speak the language instead. And outdated only begins to describe the technology we were taught there

...but I digress. Vikash, I gotcha, you're right about the trims really only being useful when dealt with in actual units of g/s or cfm. I just wanted to provide the option to actually look at what voltages or ADC values those trims turn out to be, mostly for my debugging capabilities but also for the meticulous enthusiast. I figure one of the 6 buttons (thats how many we have using only the I/O shared with the LCD, i.e. we have 7 pins to control an LCD and read 6 buttons) can be a "change units" button for whatever currently-selected value is on the display.

So yes, I see now that using the PGA method wouldn't be very fruitful as it would be a different type of trim than we want here. There's plenty of time to work with the ADC -> DAC method anyway, no need to implement something super high-speed.

[vrg3]

Cool, okay. It can't hurt to implement other things too... I just wanted to make sure you were clear on what it would mean.

[NICO]

ok now that we got both you electronic geniesses to agree free5ty1e + vrg3 together are we going to test and build the ej22turbo-tuner?

an about the test pilots you need five pepole where do i sign and send the cash to cuz for sure i want the first one. my legacy turbo is so over qalified that it puts pepole trying to race me in shock how i just left them looking at the green lights

one more thing do i need to get a lm1 reader or am i ok

[free5ty1e]

Cool... now that we agree, I do believe we'll be tuning soon...

Nico - if you're handy with electronic circuits, one of the beta chips I send to Vikash will be reserved for you, it'd be great to get some data from a heavily-modified Legacy such as I think you've got there. If you need a full prototype sent, there is a possibility that there will be time to construct a second. If everything's ready to test before I've got to leave for Nevada, I'll build you one for just reimbursement of the parts cost. I'll know more on that once I've gotten a good, inexpensive backlit display selected for production units.

And no, you won't need anything else - just hook your car back up to the stock ECU (if you're not still connected to it) and you should be able to control everything from the unit's interface. The laptop interface will be added later once we've got a fully functional and tested LegaCU Piggyback m'heah, and will be a completely unnecessary option - just adds convenience for tuning and reading run data.

[NICO]

all right free5ty1e i am ready for this, vrg3 are you up for it. both you guys toghther are makeing history on this board and makeing a crazy tuning tool for the famous SUBARU LEGACY TURBO EJ22.

[azn2nr]

jake. can you try to put your 440's back in and see if the ecu will calibrate after a few days. nico was saying that his pinks started to work after a little bit and today i talked to noah down at cobb and he said his 440's had no problem after a few days and that the ecu calibrated to them (he is using a 94 usdm turbo leg ecu). he also said that trey cobb hooked the ecu up to something similar to the scan tool and that they could see how the ecu was changing things to compensate.

[jake15]

i left them in for a weekend.... and the ecu just didnt like them for some reason i guess. i'm going to get them flow-tested and stuff before they go back in. but i want to wait until i have something to control the fuel, then i wouldnt mind just swapping the injectors and leaving it stock for a few days to see if they will compensat, but i'm not going to put them in just to take them out again... i'm lazy

[impostor]

Trey Cobb was looking into how the ECU works?! C'mon guys, Acessport for the SS!! I know, Im too optimistic.

azn2nr-I see what youre saying. I have a 95/96 WRX ECU, the connectors are the same. I searched around and found the pinouts for a 96 WRX and sketched the connectors of both ECUs side-by-side for ease of comparison. They are mostly the same, except I assume the WRX has individual coil packs for each plug. Each cylinder's ignition is controlled seperatley on the WRX, whereas our legacy it is 1/2 and 3/4. I havent had time (or patience) to figure out what it would take, but I think for someone who better understands electronics than I do, could come up with a way to make a JDM standalone work fairly easily.

Jim

[azn2nr]

never say accessport and ss in the same sentence.

even though trey is a good guy i would never want access port for my car. and it wasnt accessport he was using. since the ecu isnt as complex as the ecu for the wrx noah was saying that trey may be able to tune the maps that are already on the ecu without a piggy back.

[-K-]

All I can say is WOW. I wish I had been able to get on the Inet more in the last few months because I just had a chance to read all of this.
I have a few things to add I just wish I had been able to do it sooner.

First off, why are we going to trick the ECU into doing what we want when we seem pretty close to just "Fixing" it to our liking? What I'm talking about is our ECU has a port for the EEPROM chip. What that does is make the ECU use new lookup tables in the chip. That gives access to all the tables we need, boost, ignition, fuel, load. Now we know about the VOS Chip for the Liberty RS, he tuned the chip for a few basic mods and that's great but....
What I'm thinking is make this emmulate an EEPROM chip. But it would let the user tune it, adjust the maps. I mean instead of making tables to adjust the values the ECU gets it would let us adjust the actual tables the ECU used.
This would also open a huge World market because it would be very easy to make it work on the early WRX. There are not many options for them and none would be close to the same price range.

Talk to some good tuners about the functions and values on the maps so it's easy to tune. What you have is a great product with great features, get some subject matter experts to help make it user friendly.

I like the idea of a bigger MAF to go along with this. I guess you could just adjust for it but I don't like the idea of maxing it out and guessing at the amount of air after that. I will be looking at making a bigger MAF housing when I get home that will have 10-20% more airflow. That should make 440cc injectors a good enough match for the stock ECU, timing will still have to be looked at but I'm not sure it'll be a problem.

[free5ty1e]

Hmm... I'd like to see one of these EEPROM sockets in a turbo ECU, you say we already have an open place to put altered control maps?

Hmmmmmmmmmmmm...............

[vrg3]

Yes, there's a socketish thingie in there for replacement ROMs.

But our stock fuel control isn't really "mapped." And you certainly couldn't hack it to fuel reasonably beyond the limits of the MAF sensor in a simple way.

[-K-]

Well if you hook up the mini computer I'm sure you could make regular maps then convert them to the lookup tables to feed to the ECU. Make the maps user adjustable and you could scale in bigger injectors, bigger MAF, play with the boost control, timing, everything I think. That would be the ultimate.

Also I have heard that you can datalog the MAF and MAP and then convert the MAP signal to what the MAF would read and take the MAF out. Some people want a MAP based ECU and I have heard of this working.

There are chips like the VOS that are just a different tune, they work great but are for the EJ20G. They are also not adjustable.

Just making a chip for the EJ22T would also be interesting but would take a lot to tune and you have to tune it for a certain set of mods. Say 16G, exhaust, intercooler, 440cc injectors, and a bigger MAF. That would also be a cool way to go but it needs to be made by a tuner and you would need the right mods. I like the idea of an adjustable ECU that you just plug in.

[free5ty1e]

Hi Everybody! (...Hi Dr. Nick!)

so, I finally got a chance to get back to work on the LegaCU system... had some family and work issues recently, which sure does seem to get in the way of development.

So anywho, I've got the firmware written to do all the conversions and display on a larger 4x20 backlit LCD. I'm working on the trim table interface for adjustments, and for now just have the DAC mirroring what it reads from the MAF sensor. I'll be running both the ADC's sampling rate and the DAC's update rate at 1kHz, which should be plenty fast enough for anything our engines would need - and the processor will be able to maintain this with zero effort while taking care of all the other functions this thing will perform.

Should I still be expecting an ECU and harness to wire up the prototype with...? I just need the connector male and female counterparts to solder together to create the first LegaCU Plug and Pray connector. PM me for my shipping address if you can help me out with this.

Vikash - any info or specs you can send me on the scan tool functionality would be helpful too, as I have the beginnings of that feature in my latest version of the code.

[Kelly]

Im ready to do some R&D too.
Although My car is not heavilly modded at the time, it is modded, and I also have a set of 440's waiting to be used.

I also have access to a wideband. (and theres a dyno near by)
I'm very interested in this still, and could possably help you sell some in the future.

[THAWA]

How much longer are you going to be in florida? I don't want to send the ecu's and have them get there after you moved.

[free5ty1e]

Kelly - I'd definately like to have you test your 440's with the piggyback. Where are you located? At the very least, I bet I can have a beta chip burned for you before I move.

Thawa - I'll be here in FL until May 9th, so there's still time. The forwarding address will be in effect after that date as well so I shouldn't miss any mail. I'll PM you my address, let me know how much it costs to package and ship and I'll reimburse you with Paypal. Thanks!

[vrg3]

You want to incorporate the use of the Select Monitor port? Sure, I can give you some info on it. Give me some time to compile it.

[free5ty1e]

Vikash - Indeed, that will be an excellent way to get inexpensive scan tools to Legacy owners, if nothing else. Just PM or email me with the info once you gather it together, shouldn't be a problem to implement.

[J-MoNeY]

TTT.

[IronMonkeyL255]

This could be quite interesting if it pans out......