Realtime Chip programming?

[Will_in_China]

What are people doing with RealTime Chip programming?

Anything?

I know of somebody that has a handheld that will do something like that.

Others?

Software?

Would be nice to "twiddle the bits" while driving down the road...

-Will

 

[leroy]

I've used the LDTI Pro Series emulator for a few years with good success. http://www.ldti.com/epm_emu.htm It works great for setting idle and cruise while running.

It works well with my old 386 laptop. Jeff Scott was having some problems getting his to link with his Pentium.

The advantage of being able to make lots of changes in a short amount of time can also be a disadvantage. Some things should be changed and then evaluated for a few days under different driving conditions, and notes taken before making another change.

Jim

 

[smeagol]

emulators are out there, but it's not as simple as you'd think.

Basically you have 2 ways to operate them at this point...

1- tweak the individual bytes in a file. When you do this, you have no graphical reference as to what you are changing, you are just changing 39 to 3a (hex). You had better know what you are changing, as far as the location, in the chip file, as well as what units/scale that value has. If you do it way too often, you will memorize some locations, and it can be beneficial for tuning VE maps etc.

2- Modify your chip file with promgrammer/tunercat, upload it as a completely new file to your emulator. You may or may not need to shut the engine off and start back up; you might get away with leaving the truck running while you do it.

If someone could make some software that could make the hex editing along the lines of Speedpro/DFI etc, or group hex data and label them for tables.. it'd make life easier. Anyone with that skill willing to devote that time probably won't give that away. So now let's see the money you have invested... probably have chip burning hardware, chips, an emulator, then pay for their software, maybe a DIY wideband 02 setup, then a scan tool like Datamaster or Diacom. The only reason not to get Speedpro/FAST or DFI etc at this point then is because you like the challenge of dealing with the stock computer, because your money invested is either approaching the cost of an aftermarket computer, or has surpassed it.

 

[leroy]

Good points Brian.

Since I've never used Speedpro, it doesn't seem that bad. I carry a prom dump that I can use for reference. I don't worry about conversion factors that often, since I just want to increment or decrement a value in a table usually. I have the conversion factors if needed though.

Damn, come to think of it, you're right.....I do this way too often.

Jim

 

[Will_in_China]

Thank men,

Just as I figured.

So, would it be valuable if we had something with a graphical interface for the modifiable stock ECM values?

Or is it mostly cost prohibitive and at what cost?

I think I might be able to make something happen here for the community but I don't want to waste my time...

I am this close to getting an eval unit shipped that is yet untested on our ecm.

Basically it is a REPLACEMENT OEM ecm with serial data hooks to tie into an API on your laptop.
When all is said and done, you are left with a good chip that can be copied as needed to any stock ECM.

-Will

 

[canadian]

i fiddled around with it.

i fiddled around with it.

i have some stuff i wrote in vb that allows you to modify various tables in a .bin file, but i haven't touched it in about a year. i was borrowing keith's emulator that lets you modify bits on the fly so something like this could work, but the serial interface you're talking about sounds a little easier.

i can send you what i have if you know vb or want to fiddle around with it.

-- warren

 

[Christian Dionne]

Once you go F.A.S.T. you'll never go back.

I know my next truck's first mod will be Speedpro.

I envy you chip people, I just don't have the patience.

 

[ParTyBoy]

Re: i fiddled around with it.

Re: i fiddled around with it.

i have some stuff i wrote in vb that allows you to modify various tables in a .bin file, but i haven't touched it in about a year. i was borrowing keith's emulator that lets you modify bits on the fly so something like this could work, but the serial interface you're talking about sounds a little easier.

i can send you what i have if you know vb or want to fiddle around with it.

-- warren

Bowley's back?? whoa...

 

[canadian]

yeah, i'm here, but the truck is still in pieces. need to get my ass moving on that one. then maybe i can start working on that software again. i still have to figure out some of the conversions for RPM and things from the .bin file... think that's where i left off.

then i have to replace the tranny to transfer case adapter on the truck meaning the crossmember comes out. if that's the case, might as well pull the bed, clean up the rear end, brakes all around, new shocks, mount the air/air, maybe work on the tt setup.

who knows...

just gotta find time and motivation.

 

[smeagol]

Once you go F.A.S.T. you'll never go back.

I've used FAST, and I'm going back.

FAST is easy tuning, but at the sacrifice of other things. I like knowing what's inside my ECM, and we've made some strides with the stock computer in recent times, with wideband 02 datalogging, bigger injector drivers and so forth.

quick comparison:

TCC lockup only in stocker
better scan tool / datalogging in stocker for the most part
better diagnostics in the stocker
AC shutoff by stocker
boost control by stocker
knock retard better with stocker
fueling & timing CAN be better tuned with stocker (this is an it depends question)
emissions controls with stocker (if you require them)
switching between programs easier with Ultimate setup
no P/N switch input for Speedpro makes idle tuning difficult on some trucks

The biggest factor for me is the fact that I know how the stocker works, I know the code, I know the control algorithms. I ask such questions of Speedpro/FAST and I get... It depends... It shouldn't. With the stocker, I get the ability to change that code to my liking... I have ideas for some cool stuff I'd never be able to do in a FAST setup. Now that I have a milder Sy that I will actually drive, I may test my lean cruise fueling.. I have a way to make it work pretty easily actually. 3-5mpg better on the highway (I'd guess) would be pretty cool.

 

[turbodig]

I've used the LDTI Pro Series emulator for a few years with good success. http://www.ldti.com/epm_emu.htm It works great for setting idle and cruise while running.

It works well with my old 386 laptop. Jeff Scott was having some problems getting his to link with his Pentium.

The advantage of being able to make lots of changes in a short amount of time can also be a disadvantage. Some things should be changed and then evaluated for a few days under different driving conditions, and notes taken before making another change.

Jim

Very true... you really lose the logging/change tracking that you have with
burning. But, on the other hand, for really basic tune work, you really
don't need to know, other than saving the LTDI buffer for each set of
changes that you try.


I've been using mine for about a year now, and it works great, with a few
caveats:

1)It's not "real time"; the ECM burps when you make changes.
2)The interface is severely clunky, as others have mentioned.

That being said, the combination of it and the Wide-Band makes tuning
*much* faster. We tuned about .4 out of Mark's truck in about 8-9 passes,
whereas it would have taken a couple days to do the same the old way.

As long as you have a driver and a tuner, you can make quick work of
things. I usually change stuff while the other guy is driving around to
keep air moving and get the engine and IC cooled down. It chugs and
glugs when you change things, but it doesn't appear to hurt anything.

The thing that really made it all work well, is having DataMaster, the LDTI
software, and TunerCat and Promgrammer all on the same laptop. You]
can make a pass, look at the data, pull open the file in Promgrammer,
do your editing, and write the file back to the emulator. Most times, I skip
Promgrammer and twiddle bits in the LDTIs hex editor, mostly because I
have no life, and have managed to memorize table locations.

This isn't a setup for the average guy, as BG mentioned... you'd probably
be money ahead to get a GEN7 DFI or a FAST. But, if you're setting up to
be a tuner in your area, and plan on doing more than one truck, it's a
hell of a setup to have.

Tech-tools makes a dual-port ram based emulator called the UNI-ROM...
supposedly is true-real-time. This would make for the ultimate setup, but
the cost is over double the LDTI.

The LDTI's parallel protocol could be figured out, using printmon from
www.sysinternals.com. There is also some freeware parallel "sniffer" tools
that could do this as well.

Getting Programmer (or something similar, like the S/W Warren was
working on) to send to the LDTI would be very sexy.

Later,

Dig

 

[Will_in_China]

So it seems pretty apparent that any dashboard-ish, laptop based, stock ecm tool would be of marginal value?

Did you get my PM regarding this greenie?

It seems that those making chips, have learned the hard lessons and the rest are waiting to -buy a chip-?

-Will

 

[smeagol]

Sorry mailbox was full....

We've been making more and more resources available for those who want to do their own chips. Unwritten rule is to not hand it out, but share with those people who have proven themselves worthy. We can't hold anyone's hand every step of the way, it's too time consuming, and the path is half the fun

 

[BillC]

Getting Programmer (or something similar, like the S/W Warren was
working on) to send to the LDTI would be very sexy.

Hey Dig,
Any luck getting the LDTI protocol from the manufacturer? Like we discussed at the Nats, if you can figure out the protocol, I'll put it in Promgrammer so y'all can do direct updates from the editor to the emulator. Also, you'd be able to keep track of the various changes by simply doing a "save as" after each update.

 

[turbodig]

Getting Programmer (or something similar, like the S/W Warren was
working on) to send to the LDTI would be very sexy.

Hey Dig,
Any luck getting the LDTI protocol from the manufacturer? Like we discussed at the Nats, if you can figure out the protocol, I'll put it in Promgrammer so y'all can do direct updates from the editor to the emulator. Also, you'd be able to keep track of the various changes by simply doing a "save as" after each update.

I haven't heard back from there yet, but I'll pitch something at 'em again to see what I can work up... it'd only improve their sales of hardware, so
I'm not sure why they wouldn't.

All else fails, we can hack. I'll download Printmon again and see if there's
an obvious pattern to their transmission.

Later,

Dig

 

[turbodig]

So it seems pretty apparent that any dashboard-ish, laptop based, stock ecm tool would be of marginal value?

Did you get my PM regarding this greenie?

It seems that those making chips, have learned the hard lessons and the rest are waiting to -buy a chip-?

-Will

I don't think so... I think there's a lot of room for evolution in the current
toolset, such that ECM tuning could be more palatable for the average Joe.

Having the GEN7s or FASTs front end to the emulator would be sweet,
would save me a lot of time in tuning things.

Particularly the FAST's histogram display of a given run over the fuel table
you're editing, so you can tell which cells are being used without a lot of
lookup.

This would imply some sort of a Hybred of Programmer and DM, whereby
you could load a uni or GDF to overlay over the current file.


Side note... Is anybody else having trouble getting F77 to match up to DMs
MAP reading? It doesn't seem to be consistent from truck to truck.

Later,

Dig

 

[rarebit]

i was thinking about this project some years ago. what i was thinking of doing was designing a little board that simulated an eeprom. it would use a dual port static ram, a microcontroller like an 8051 or hc11, and an rs232 interface for connection to a laptop. a small program would run on the laptop that looks in a directory for bin file updates generated by promgrammer. this effectiveley would give you the flexibilty that speedpro offers(even wide band correction...but thats another story.)

simply save the bin your tuning and it would automatically update the little eprom simulator. with dual port ram the ecm would see no glitches and therefore would not have to be reset every time you make a change. a more advanced system would be updated directly by promgrammer via a serial port using a simple protocol to adjust only the memory cell altered with promgrammer. this would require the author of promgrammer to write a simple serial protocol to communicate with the eprom simulator.

yet a more advanced system would utilize two flash memories working in tandem so that one could be updated while the other was being used by the ecm. a system such as this would retain its memory even while the whole system was powered down. it would also remain glitch free.

i can think of two approaches to a wide band solution- the first requires scrapping some portions of the existing ecm code(like egr and existing o2 code) to make room for wide band code in the ecm. the wide band code would consist of a table containing the desired a/f ratio verse map and rpm, an update rate value, and correction % value. another table or possibly an equation would be needed to scale the wide band voltage o2 signal into the proper o2 ratio.

the second approach to wide band correction method could be a promgrammer like program that performs the wide band algorithm on the laptop and then updates the ve table directly.

if the second approach could be designed, possibly a self tuning system would result with a little more tweaking. all the user does is modify their truck, and over a period of hours, new ve values are integrated into the ve table. essentially a self tuning truck.

the beauty of this is that it doesnt require the self tuning code to exist in the stock ecm. this auto tuning system would require a promgrammer like program to record the rpm, map, and o2 correction. promgrammer would take samples(say a few hundred) of the previously recorded correction o2 data for a particular ve cell, average all the cell data together, and come up with a new ve cell value. out of range data could be filtered out to prevent anything screwy(fouled plug, faulty map, etc) from ruining the ve table. maybe even allow the user to evaluate what cells the promgrammer wants to update before the actual update takes place.

the actual wide band algorithm to update the fuel base pulse width is rather trivial once all the tables and values are in place- the desired a/f ratio is determined via the desired a/f look up table, that value is then compared with the actual a/f ratio. the difference of these values is then scaled by the correction % value, and finally fed into the final base pulse width at a periodic rate determined by the update rate value...a simple feedback loop.

a similar system could be designed to automatically adjust the timing table depedning on knock counts detected. essentially the computer slowly advances the timing of a user selected cell until knock is detected or a user defined limit is hit. if knock is detected, the computer then backs off the new timing value by a user determined amount and plugs this value into the timing cell. at this point the user could be prompted for a new timing cell to tune. such a system would only be used during the tuning process, and once the maximum timing values are established it would no longer be used.

maybe if a few people are seriously interested in either of these projects and would invest time and money we could create something actually better then speedpro.

 

[JSM]

There is interest there, here is my take on what is holding it up.

1. Not much money to be made on it, $$ rules the world.
2. Those that would like to do it, either
A. Don't have time
B. Don't know how. <--ME!!!


I would be glad to help someone develop something like this, if they had knowledge and I could be a helper.

 

[canadian]

realtime software

realtime software

seems to me like i'm duplicating work here, but i've been back on the EMInterface again. not trying to steal anybody's thunder here (especially bill) since i know how much work this stuff can be and Promgrammer does a lot of what i'm trying to do.

the only difference between my code and bill's (apart from the programming language), is that mine is basically generic. you can add/edit/delete any tables you want in a settings file.

if you've got a .bin file for a different car, just set up the setting file and go. menus are created based on the settings, etc and i'll be trying to get the ability to write through the emulator software to the ecm on the fly.

if anybody wants to lend a hand and knows visual basic drop me a line and i can share what i have so far. i'd say the functionality is 75% there, just need to clean it up a bit and figure out how to send stuff through the emulator (shouldn't be harder than to call the command line parameters).

dig, if you want to check out what i have now (doesn't save to a file yet, and it's strictly ide, no .exe), send me an e-mail (jabonie@hotmail.com) or pm and i'll fire it off to you.

 

[rarebit]

some rom ices can be updated on the fly just by saving a new .bin file in a directory. the software automatically looks for an updated file, loads in, and writes it to the rom ice.

if you could add data logging capability through the ecm serial port- coupled with the wide band o2 mod, its only a few steps away from a very powerful closed loop wide band system. it might not correct as fast as the speedpro system, but that might not be a problem.

too bad i dont know visual basic. ive done a lot of programming visual c++. ive seen a little bit of visual basic. i know you can call visual c++ through visual basic. maybe i could write a serial port interface for use with a diacom like cable.