MCB900, FLASHMAGIC and LPC922 case-study

[Hitesh]

Hi friends,

About six month ago when I started using LPC9xx chips with MCB900 and FlashMagic, I had taken good help from many of you. Since then I have been using these two tools to program LPC922 CPUs. I have observed some typical problems. And some clue-less solutuons too.

I have P4 machine and windows98 OS
MCB900 v4 on COM1
Flashmagic V2.07, DLL v1.74
Target CPU (LPC922) connected through header, chip on MCB900 removed.

Problem: Sometime all of a sudden FlashMagic stops communicating to target CPU. After many retries it never succeed. Just like CPU suspected failed. Repowering of MCB900 or reloading of FlashMagic donot help.

Sol-1: After complete reboot of system, it works normal. Sometime this also donot work.

Sol-2: When sol-1 donot worked, I removed failed target CPU and connected a fresh-one, FlashMagic loaded successfully. After that again I connected suspected-failed CPU, that also worked. And I could reprogram it many times, and still its working.

This behavious I couldnot understand. But still, it works. So I am sending this case study to forum.

With warm regards.

Yours truly,
Hitesh Mahant
www.vital4all.com

[Andy Ayre]

Interesting. Please take a look at the Flash Magic application note "what to do when ISP does not work" and refer to the "last resort" section. That will tell you how to generate debug files.

Can you please repeat these steps generating debug files. Keep renaming them, so you have one debug file for Sol-1 when it fails, another debug file for Sol-2 when it fails and yet another when Sol-2 works again. Email them to me. My email address is at the bottom of this posting, in the help about window of Flash Magic, etc.

It's important that you create seperate debug files for each specific point, otherwise I will have a hard time figuring it out! Thanks.

[Andy Ayre]

But before you do that, have you tried another PC? Do that first, because it may be a faulty COM port, perhaps.

[erikm]

I have P4 machine and windows98 OS


Andy,
could it be that you test with something current and there is a glitch when using '98

Erik