MrDude_1

jsut a off the wall question..


using HPtuners on LS1 PCMs, it writes out to the data table areas.... you CAN write out the whole thing, but 95% of the time, you can just write out the data.

is there anyway with the LT1 PCM that you can write everywhere EXCEPT over the bootloader? so even if it fails, so what? you can write over it again.....






anyone?

junkcltr

iTrader: (1)

Also, keep in mind that the ALDL runs at 8.192kbits per second. The PC runs around 8.228k bits per second. This is over 4% of error which yields a fairly high probability of error. To do it right you need some DOS code (windows won't let you get at the baud rate generator directly) or special hardware.

There are two 64k by 8_byte chips ( 2* 64 = 128k by 8) BYTES. Or written another way. There are two 512kbit chips (2*512k = 1024k BITS.

you are confusing bits with bytes. All 128k BYTES are used for storage. Not sure what info is stored where. You could dump the memory and disassemble.

oh yeah, if Tuner has a low volatge setting I would bet that low voltage wasn't the problem and that it was simply a serial comms error. Did you try reading the PCMs after the error occured? That would tell you if it was just a coms error or a bootload wipeout (I think).

11sORbust

I think it was low voltage. Yes I tried reading the PCMs after the error. Every time the pcm would not respond. The first crash happened after 99% was programmed. That one wouldn't read or write but the ALDL datastream would display through datamaster afterwards. Then, the second one was read twice before programming, no communication after the crash. Check the voltage and it was 10.9 . Third one was exactly the same as the second. It even crashed in the same exact spot (right at the start of programming). Only difference was that I charged the battery for the 3rd PCM. BUT I guess it's possible for a drop of .5-1.0v with key on/ engine off. He has a TV and some other stuff in the car. TC never suggested why the low voltage error didn't pop up.

FYI, I'm dumping both chips. Still learning how to disassemble and stuff. No public hacks for the lt1 that I know of..

junkcltr

iTrader: (1)

You already have the contents of bith chips if you read out the PCM first. To get you started on the dissasmbly you could start with where the data tables are located. That way when you read through the code you can get some info on what tables it is going to. The software you bought gives you the location of the tables.

CrazyHawaiian

First time I've ever read about this and I'm getting scared (hahah!). Does this only apply to the LT1 ECM's? Or does this only apply to certain tuning hardware? I was planning on setting up a romulator with a prominator (730/749 SD), but if a voltage drop from fans coming on or something would damage the hardware, it dosnt sound like a good idea. If thats the case I would prefer to set up a bench and not chance it.

11sORbust

So both chips= 128k? Guess tunercat takes both chips and assembles them to one file? I can't just load the TC dump to one flash chip, right? If not, then how would I split it up to load it into two chips?? I was talking about raw dumps off each an28F512, but you are saying that's the same thing,huh?

Yeah, I know the locations are somewhat "out in the open". The bootloader is not though. Guess the bootloader location will be displayed at first part of the code. Still trying to figure out how the program works. From what I understand, bootloader is just code that'll load up the program(from PROM to ram) for the MCU to operate. All MCU's have this bit of code. Bootloader is suppose to reside in the PROM. Not sure how the bootloader can load the calibration data to a chip, considering flash chips erase fully before the rewrite? I can only speculate that bootloader is moved from the rom to ram when the ignition is turned on. Then it works from the ram, communicates with the serial port and reflashes new calibration data. Even then, it's stored into the ram so I don't get how a programming error would prevent a reflash in the event of a failure. Sure if the ram is volatile and the key gets tunred off, but I still don't get the details...

11sORbust

Don't worry, your ECM is not even close to a PCM...

junkcltr

iTrader: (1)

This does apply to the LT1 PCMs. It doesn't apply to the 730/749 ECMs.

junkcltr

iTrader: (1)

Yes.
Yes.
Yes.
N/A
The 128k file has everything that both chips contain or should anyway. From what I have heard is that the info from each chip is scrambled into the 128k file. So yes, reading each chip seperately would help in figuring out the scrambling.
Don't worry about the bootloader thing. It was GMs first attempt at in-car-reprogramming. Just like other things I read about on this board......it was the learning curve stages for them. It works sometimes, just not all of the time.
Good news is that most EEPROMs today are +5V programmable. And no, I have yet to find a drop in replacement for the LT1 EEPROM. Looking to set up my emulator for the LT1 this summer though.

11sORbust

Thanks for all your help! What about the AT29C512? It comes in PLCC form....

So, it sounds like IF a 5v flash chip will work AND bootloader is modified, there will be ALOT smaller chance of a programming error?

11sORbust

WTH does that mean?

Maybe that's why the junkyard PCMs didn't work?



.........


FYI, the reman PCM did read on the bench this morning. Guess if I can read the PCM, the cable is fine? Maybe bootloader checks the VIN before writing to the flash chip?

RBob

iTrader: (1)

Where are you quoting this stuff from?

RBob.

HalfInchWrench

Would a buffered connection help or are most of these problems voltage related?

I know I've built a buffered JTAG for other purposes using an SN74HCT244N and it works flawlessly. My old unbuffered JTAG would sometimes cause RS232 communication problems.

funstick

from experience most of the flashing errors and trouble ive seen are cuased by low battery voltage.

11sORbust

Sorry, TC help file. Maybe if I match the BIN's VIN to the PCM's it wouldn't have locked up? The first crash was on the original PCM. I think the voltage dropped right at the end of programming(no low voltage error). Second and third crash was different. Both was out of a roadmaster. Both failed to program right after TC said the memory was erased and just as it started to program.

funstick

mind sending me a copy of that datasheet. never seen one for that chip.

11sORbust

Check your email....

junkcltr

iTrader: (1)

The VIN doesn't matter when programming. I bought a used roadmaster and loaded it with a camaro BIN.

Make sure you swap knock modules. Camaro is setup for one knock sensor. The cop cars and roadmasters use two knock sensors. Keep the right knock module with the right engine. LT1 tuners play games with the knock module (LT4 and stuff) but I don't think it works out correctly, well, electrically anyway.

11sORbust

One last question. I can read PCMs just fine. Datamaster works great. The cable passed DM's cable/com test. Does this mean the cable is fine? Reading the PCM works but programming failed (3 times).....that really sounds like low voltage. I'm going to assume incorrect msg sent means the message sent back from the PCM doesn't match the original msg sent. Guess when bootloader haults programming, it'll send back a different message..or something like that.

One thing that doesn't make sense,

The TC program checks voltage before programming. I know that my voltage was way too low for one of the attempts. Not one low voltage error. I really hope the program doesn't check voltage after erasing the memory.

HaulnA$$

The DataMatser test only checks the PC end of the cable. It checks the TX/RX functions in a loopback fashion (the MAX232 performs the loopback). The test has no way to check the PCM end of the cable. The TX portion of the PCM end of the cable could be causing your problem. Only way to know for sure is to try a known good cable. HTH

11sORbust

Reading/communicating with the PCM doesn't use TX

HaulnA$$

Yes, but reliably transmitting three or four bytes to connect to the PCM and reliably transmitting 64K bytes to flash it are two different things. HTH

junkcltr

iTrader: (1)

Really?? What is the difference? This is RS-232 right? It is supposed to re-sync for each character......stop/start bits.

HaulnA$$

Gee, are you the only person in the entire world who has never had a Modem or data terminal lock-up? I don't wish to get into a pi$$ing contest. Just trying to help the guy out, thats all. Antagonism never helped anyone.

junkcltr

iTrader: (1)

???? What are you talking about. I don't get it. Anyway, this topic is dead to me now.

11sORbust

IMO, he could have worded it differently but it wasn't a cut at you. Maybe you should hold your cards a little closer. He might know something you don't. If you are right, then my cable could be at fault. If you are wrong, then the worst is that you'll learn something new. So....why do you say there is a difference between communicating and programming a large file?




That hurts I couldn't find a whole lot of info on the LT1 when doing research, vendors didn't help much with my problems. Figure this thread might become a decent souce for people wanting to dig. Just got both chips (T and E side) and plan to solder the sockets in tonight or tomorrow. Still waiting on the PLCC-DIP adapter. I'll have raw dumps of both chips if anyone wants em. Guess I'll know soon enough if my cable is bad.

junkcltr

iTrader: (1)

http://www.geocities.com/MotorCity/S...24/95cam2.html

funstick

just finished reading the data sheet. wow what a stupid move by gm. there were 5volt flash parts avaiable at the time. what was the engineering department thinking. key on engine off most battersy hang around 11.5-12.2 volts. good ones anyways. yeah its does take 12volts to program the chip. in thoery this is probolay the most likely culprit for flash errors. wonder if some simple hardware and code mods could be done to set it up for a 5v flash chip. most of the programming logic should stay the same. on the mcu in the ecm though there is some is usually some flash rom.i wonder if its avaiable on the mcu if we could rewrite the writing routine to store the bootloader into the flash rom so in the instance of a programming error it could be restarted and reloaded. ill ask around and see if this is doable. i know a few ee's

junkcltr

iTrader: (1)

All that you said is true and I tend to agree with it being a low voltage problem in this case.

CATS Tuner might set their lower voltage limit threshold low in order to try and program the PCM. I don't think they would want to have their software saying "low voltage" every time one tried to program a PCM. People would think their software is bad. Just a guess. You could ask them the lower limit for allowing programming.

Another thing is that the "two transistor" ALDL is used by some people. This is a hack and does not meet the RS-232 standard. Therefore, it is never guaranteed to work. Only use a device that contains a real RS-232 standard chip (.e.g MAX232). I know some use the two transistor for ALDL, but why cut cost when dealing with a 500 - $10,000 engine. The difference in the ALDL cable is like 10 bucks. I am sure some will be all over me that the two transistor is fine. Anyway, if programming a PCM don't use the two transistor ALDL.

funstick

yeah so the use a battery charger stupid thing ive been saying for the last 3 yrs might make some sense huh ?before i came here. lt1 edit i used it locked a few pcms. got the low voltage error. been using a battery charger since.

HaulnA$$

Please accept my sincerest apologies. Had a bad day. There are a lot of things I don't know, like as much about PROM tuning as I would like, but I am learning as much as I can as fast as I can. What I do know is Electronics and it just seems that very time I have posted lately, others have been quick to pounce and I got rubbed the wrong way. I'll try to not to let it happen again. I only post where I know I can help. I'm not one of those who repeats some rumor or half truth from some other questionable source as reflected in my low post count . As for your cable, asynchronous communications, especially RS-232, which is an ancient serial protocol, are not exactly known for their reliability. Just look at all the posts on this board about it. There are a bunch of ALDL cable designs out there and most are questionable when you need reliability. The transmit portion of the circuit is most critical. It is much easier to recieve reliably than transmit. I could very easily be wrong about your situation. I am just raising a possibility. Good luck. HTH

junkcltr

iTrader: (1)

Sometimes I take things the wrong way and sometimes I have bad days too. You raised a good point when talking serial comms stuff. The standard RS-232 protocol tries to defeat frequency error by sending short frames. That is, start and stop bits per byte. You tend to get freq. errors with a start, then a lot of bytes, and then a stop.

One other thing. We have been talking about +12v here. But the true absolute minimum at the chip is +12 - 5% or +12 - .6 = 11.4 volts.

11sORbust

I have everything to socket and program the flash chips. Will know what's wrong soon enough! Hope to have both boards socketed tomorrow. Any tips?

HaulnA$$

I have one tip. When you go to solder the PLCC carrier sockets to the surface mount board, be careful how much heat you use. Surface mount traces lift a lot easier than plated through PCB's. It would be a bummer to successfully burn the bootloader but destroy a PCM or two. Keep us posted. HTH

3.8TransAM

To all whom concerned:

I would like to thank you for everyone here at the board for resolving your squabble peacefully. :-)

I just caught it reading thru this post. It has been informative and inciteful as well. I'm reading along trying to cover all the bases and pitfalls of LT1 flash tuning. Going to be investigating the software/hardware for flash tuning on GM LT/LS engines soon and want tos tart out succesfully.

later and thanks
Jeremy

11sORbust

.....and I want to thank you (and others) for keeping this thread open. Like I said before, I have gotten very little help from the vendors. That is in no way a slam towards them though. I'm sure Andrew and TC is very busy. There is very little "real world" information out there (that I could find). I knew that with all my friends here, the problem would get resolved. It's great to get more info out there, for others searching for insight. That is why I posted this instead of working through PMs.

I think this board has matured greatly in the past year or so. IMO, it all boils down to how one deals with a particular situation. Given my past history here, you could say I have learned alot on that subject.


......... http://www.geocities.com/MotorCity/S...24/95cam2.html
That is a great link for this board to have archived!

11sORbust

I have applied hear directly to the chip, enough to burn my finger. Still didn't lift. I'm not using brute force but am prying it fairly hard. Don't want to lift the traces. Any more sugeestions?

jwscab

get your self some solder wick, available at radio shack, its basically copper braid. Place it between a soldering iron tip and each pin, and it will wick away most of the solder.

then, really, a special heat gun(hot air) with a special tip on it is what should be used. but you can probably get away with a regular heat shrink type heat gun, but you have to be very careful.

good luck with it!

HaulnA$$

I have a question and one suggestion. Are you using a baffle to direct the heat from the gun? If not, you might try that. You should'nt have to pry too hard to lift the IC. Solder quickly turns from a solid to a liquid once the melting point is reached (approx. 400 deg. F). I also suggest that you pick up a piece of scrap electronics to practice on before you attempt to rework the PCM again. HTH

junkcltr

iTrader: (1)

I don't have a heat gun. I use a regular fin-tipped soldering iron. I always use the "destructive" method for surface mount devices.

I cut all of the leads with a small set of electronic wire cutters so that I can remove the chip with the leads still soldered to the PCB. I then remove each lead one-by-one. This destroys the chip of course but uses the least amount of heat to the PCB.

It all depends on how important to you the existing chips are.

11sORbust

Thanks for the tips. I'll have some results in the nxt day or two. Must say that a PLCC chip is nothing like a DIP, in terms of soldering.

11sORbust

Right now I'm stuck because my heat gun is bogus. Need to get a good one....one with the adapter to apply heat directly to the chip. BTW, I am practicing on an old sound card. Must say that desoldering braid works great!...

Craig Moates

I've used a Mapp gas torch in the past with good results. Not really endorsing it, just saying it can work quite easily. 10 seconds and a quick rap on the bench, and you're done.

11sORbust

I thought about this. There is a map torch laying around. Going to check funds and see what I can do. Dont think one could cut the legs off a surface mount plcc chip.

Craig Moates

You can, they're only attached at the top of the chip.

On the Mapp gas torch, I'd suggest heating just the chip and only briefly. The chip will get hot enough to melt the solder at the pads, and then it'll come right loose. Just be careful not to melt everything else around the chip. Seriously it only takes about 2-3 seconds of heat.

junkcltr

iTrader: (1)

I am curious. How did you make out the chip removal? I wouldn't try re-using the old chips if you heat them with MAP gas.

11sORbust

Decided against using the map gas. Figured that I have very little exp with circut boards and one f-up will cost at least $100. Couldn't afford a real heat gun right now. Didn't have mini dikes. I ended up using the method you suggested. The cost of a chip was much less than risking the circut board. The though of desoldering some legs sounded reallu good(and easy). In the end, I used a dremel to cut off the legs. Right now one board is done and it looks perfect! Plan to socket the other board in the morning. Then I can put it on the bench. For some reason I think my cable will turn out to be the actual problem.

junkcltr

iTrader: (1)

I haven't messed with the LT1 stuff much yet besides just programming it a few times. I did take a look at the BIN I read from the PCM when I got it. I was comparing it against the data from that website with the link above.

I did a few lines of hand disassembly of the file I have and compared to the files from the link above. Unfortunately, those disassemblies are very wrong.
I also noticed that the E-side of the code resides in the lower 64k bytes and the T-side in the upper 64k bytes of the file I read from the PCM that was 128k bytes total. I did not see any signs of address swapping. Although, I only spent about an hour on it.

How are you going to program the new chips? I am guessing that you could split the 128k file in two and program each chip using the newly created 64K byte files. I am not sure which PCB is the E-side and which is the T-side. Remember, I only looked at it quick so please don't take this as fact.

junkcltr

iTrader: (1)

Could you shed some light on this?

3.8TransAM

junkcltr,

The mixed/scrambled address lines is explained some here: http://www.geocities.com/MotorCity/S...24/95cam2.html

Its the link from earlier in the article that I was reading, I can hang with it, but the stuff is curretnly over my head still.

later
Jeremy