It goes back to that dreadful example with ATMEL to which I refer from time to time.

They produced an EVB for their ARM7TDMI board on which nearly nothing at all worked. Their SDRAM interface didn't work. Their boot code didn't work. Their VERY costly SRAM board was underdesigned, so you couldn't run their software on it. Many of their on-chip features didn't work ... it wasn't even possible to give the thing a "trial run." BTW, it was, by far, the most costly eval kit of its type, by a factor of about 3x.

We settled on a much less costly eval kit for an ARM9 chip that cost somewhat more, but which functioned very well and, while the chip cost more, and worked, the eval kit cost 20% of what the ATMEL kit cost, and worked perfectly as advertised as did all their software.

Their French design team tried to blame me for the fact nothing on their product works. It was, of course, eventually withdrawn from the market.

As a consequence, I don't use anything from ATMEL. My clients don't use anything from ATMEL. I don't even consider them, and I tell the story to anyone who will listen. It wasted over a year of our time, and produced absolutely nothing. You get the picture. It clearly illustrates their attitude toward their customers and, hence, I'm not one of them.

RE