Next EEHack update

[steveo]

i finally have an ECM and a serial interface working on the bench. i have decided to stop bundling a flash tool with eehack and turn it into a separate (better) tool that'll come with EEHack... i will use this tool as a learning experience for my next project (another flash tool.) stay tuned.

[NomakeWan]

Oooooh I’m excited! Best of luck to you!

[steveo]

HERE WE GO AGAIN

god my brain hurts already

the beer on the right is the most important tool on this makeshift workbench.

IMG_1318.jpg

i keep saying to myself that good things will come of this, even though i never really plan on owning an LT1 again.

[steveo]

step one is to solve the bin scrambling thing that's been driving me nuts for ages.

i've traced the flash to 68k address lines. the following shows the re-ordered lines, so a bit of bit shifting in the indexes in my flash tool should easily convert bins between 'datastream flash' and 'hardware flash' bin formats. here's the t-side magic (lines that are normal not shown):

CPU_ADDRESS,EEPROM_ADDRESS
13,1
12,2
11,3
3,11
2,12
1,13

[steveo]

seems they were nice enough to make t-side and e-side the same, so that formula works for addressing on both sides.

i will now find a reasonably optimal mathematical function to derive the modified address scheme in either direction.

this reminds me of when kur4o gave me a bunch of examples for EE's programming key challenge/response, and it took me a ton of staring to realize that EE's secret programming key was just a NOT gate.

[steveo]

the bits in question we are swapping to derive the correct hardware addressing actually move a mask of 0xE from the low bytes to the high bytes (0xE to 0x000E)

i doubt it's an accident. i think we have found a hilarious GM circuit design easter egg.

[steveo]

i think this will translate 'hard' to 'soft' addressing in $EE bins:

( x & ~0xE00E ) | ((x & 0xE000) >> 12) | ((x & 0xE) << 12)

EDIT: this is not a working example, it fails to reverse the bits, will post a better and more human readable thing soon

[steveo]

looks like it works! so now we'll be able to use an eeprom programmer or the software flash routine interchangeably.

[In-Tech]

looks like it works! so now we'll be able to use an eeprom programmer or the software flash routine interchangeably.

Sweet, that is way cool. All the stuff you guys have accomplished is truly awesome. It makes me want to buy an LT1 Transam and play but I'll keep talking myself out of it and just follow along as best I can.

[NomakeWan]

looks like it works! so now we'll be able to use an eeprom programmer or the software flash routine interchangeably.

This is absolutely awesome. I have an experimental clip that I’m jerry-rigging into an in-place EEPROM programmer, so that’ll make things a lot easier. So would finding out there’s a spot on the PCM itself I could just plug into without my test clip though, haha.

Great work in such a short time! This has the makings of the best thread of the year.

[sherlock9c1]

This is absolutely awesome. I have an experimental clip that I’m jerry-rigging into an in-place EEPROM programmer, so that’ll make things a lot easier. So would finding out there’s a spot on the PCM itself I could just plug into without my test clip though, haha.

Didn't the guy in the 97 F-body thread say there WAS a location on the board to do this?

Steve, let me know if I need to send more beer money!

[steveo]

i have a very high rate of beer money consumption so every bit helps

there is totally a header that seems mostly pinned to the flash chip

flashing in-place on a non booting ecm is a high priority right now

in fact i might do that next

[sherlock9c1]

Send me a PM or email with your paypal address.

Dumb question, but did you already order the flash routine so it fills the BIOS (or equivalent) first, so that the amount of time that flashing is health-critical is quite small?

[steveo]

there is a gigantic 'donate to beer fund' button in the 'about' menu in eehack.

i have not done any reordering of the flash routine to increase the chances of it being bootable if it fails halfway, i do plan to work a bit on that, but i think i'll focus more on making it possible to 'keep trying' to flash as long as we don't reboot by losing ignition power to the ECM

[steveo]

i've started writing a small set of classes and subclasses for handling not just EE bins but bin files in general (this will allow the new tool work with any bin format in a generic way). the subclass for the EE bin format now handles standard (tunercat/eex), lt1edit backwards, and raw eside and tside with addresses scrambled

so first (small project) is a success

until i write a proper ui for it, if anyone needs a bin converted for their socketed ecm, just holler