openecu.org

I am wondering just how many times it can be reflashed before it wont take a new flash. If there is a practical limit, will the ecu fail if you hit it or will it just refuse it? Also, if you dont know the answer, then how many times have you reflashed yours? I am around 8 so far.

Jeff

I think the failure rate would be closer to 1000 full reflashes of all memory blocks. I'm not saying it couldn't happen at 100, as some people say, but I just think its highly improbable.

Spiider wrote:I think the failure rate would be closer to 1000 full reflashes of all memory blocks. I'm not saying it couldn't happen at 100, as some people say, but I just think its highly improbable.

New here...and looking to get started on this as well...but what's the difference between a flash and a "full" flash? Any other info on this subject?

The flash memory is erased in blocks, and rewritten. EcuFlash examines the changes between what is currently in your ECU's flash memory and what you would like to have in the flash memory and only reflashes the necessary blocks. When tuning a car, you are typically only changing some data tables which reside in one area of ECU flash memory.

Since this may involve only one block, the reflash process can be done much faster, and without wearing out the other flash blocks. In reality, certain blocks are always the ones which get changed, and so these will fail before the others. I have yet to experience a failed flash block that didn't erase or rewrite. What is more likely is that the memory may get bit errors over time or temperature variation many years down the road. The chip manufacturers want these memories to last a very long time in harsh environments, so they are rated very conservatively.

cboles wrote:The flash memory is erased in blocks, and rewritten. EcuFlash examines the changes between what is currently in your ECU's flash memory and what you would like to have in the flash memory and only reflashes the necessary blocks. When tuning a car, you are typically only changing some data tables which reside in one area of ECU flash memory.

Since this may involve only one block, the reflash process can be done much faster, and without wearing out the other flash blocks. In reality, certain blocks are always the ones which get changed, and so these will fail before the others. I have yet to experience a failed flash block that didn't erase or rewrite. What is more likely is that the memory may get bit errors over time or temperature variation many years down the road. The chip manufacturers want these memories to last a very long time in harsh environments, so they are rated very conservatively.

Ok, that makes sense. Basically this is the first car I've owned that I could tune. Friend of mine has a 98 TransAm and uses HPTuners (OBDII, flash, scantool, logger, ect), has had excellent luck and results. I'm trying to piece together what I need to get started when I find time this summer to actually dial things in, and just being unsure/unaware, having read the "100" number other places...it makes someone uneasy

What would cause failure? It doesn't make sense to me how an IC could "wear out" -- is it electromigration or something else?

I think the high potential used to create tunneling to remove charge during the erase process can trap charge in the oxide, which is going to shift your effective Vt, amongst other things. Trapped charge == bad. I haven't done IC design in 10 years, so maybe someone else can give a better answer.

Colby

I did a lot of reflashing on varios applications (many PIC16c84 based) some time ago.
What killed the chips when they were in for a upgrade:
1: Careless tech personel
2: Static electricity and mechanical stress(some of the boards was delivered to me out of the casing )
3: wrong/unstable programming voltage and/or power supply. Mostly bad grounding and connectors!
4. Other faulty components, mostly capacitors.
5. x times reprogramings.

So I guess a part of the answer depends on how the programming voltage is done.. ...throw bad connectors in here..

What really will kill youre ECU is age, temperature conditions, moisture/dirt and power supply quality.
If youre ECU lives in a rough environment its much more likely to show its faults during a reflash.

Make shure youre ECU has a good life, ..dont stress it while reflashing.., and you can reflash maaany times

Does anyone have the model number for the flash? A little research could be done on the manufacturer specifications.

chuckdez wrote:Does anyone have the model number for the flash? A little research could be done on the manufacturer specifications.

yes, but the flash memory is integrated into the main CPU.. ..As for the JECS ECU no specs is aviable.
Dont know if these specs aviable at all for any Subaru ECUs..?

NoCtrl wrote:I did a lot of reflashing on varios applications (many PIC16c84 based) some time ago.
What killed the chips when they were in for a upgrade:
2: Static electricity and mechanical stress(some of the boards was delivered to me out of the casing ):

So for the personal car flasher (like me) this means don't allow static electric discharge?

NoCtrl wrote:3: wrong/unstable programming voltage and/or power supply. Mostly bad grounding and connectors!:

Basically a healthy wiring - power and ground, system on car?

Im using a 20' extension USB to reach my car in the garage, will this cause anything?

C6ect.ect wrote:Im using a 20' extension USB to reach my car in the garage, will this cause anything?

So, do you push the button and then run to turn the key on to flash or do you yell down the hall to someone to do so?

C6ect.ect wrote:So for the personal car flasher (like me) this means don't allow static electric discharge?

Basically a healthy wiring - power and ground, system on car?

Im using a 20' extension USB to reach my car in the garage, will this cause anything?

yep, yep, and ..hm.. well, yes and no.. (Read: I dont know)

USB extension: The logic in the Tactrix cable is inside the ODBII connector. And if I remember things right the programming voltage is 'made' form 12v supply there.
And I really hope there is some logic/filters inside the ECU as well..

Dont know how fault correction is done on the USB side of the cable,..

You will pick up noise/statics in the USB extension cable.
So, use a short as possible extension of good quality.
The safest bet is to use no extension at all.

cboles wrote:I think the high potential used to create tunneling to remove charge during the erase process can trap charge in the oxide, which is going to shift your effective Vt, amongst other things. Trapped charge == bad. I haven't done IC design in 10 years, so maybe someone else can give a better answer.

Colby

That is the appropriate answer cboles !

Renesas specifies that the flash can take a minimum of 100 reflashes., no
typical or maximum figures given.
100 is a very low figure compared to what other manufacturers guarantee, they are usually in the range 1000 - 10000.

JonnyM wrote:Renesas specifies that the flash can take a minimum of 100 reflashes., no
typical or maximum figures given.
100 is a very low figure compared to what other manufacturers guarantee, they are usually in the range 1000 - 10000.

Don't tell Cobb that. Those jokers sold more APs on the "flash limit/realtime map benefits" story then anything else I think.