An other Microchip screwup on teh 18F14K22

She may decide to use a given pin as an interrupt rather than an ordinary input pin, or to bit-bang communication rather than using internal hardware to meet an oddball break requirement etc. etc.

Best regards, Spehro Pefhany

Yes.

Yes. I hate big meaningless blocks for high density logic. I use homogeneous blocks and separate functions and power. This almost always changes from design to design so the library changes with it. It's *not* that hard to do.

The library part reflects its function. If the function changes then the library part changes along with it. If it changes from hardware based communication to bit-banged the function didn't change.

On a sunny day (Fri, 07 May 2010 21:37:22 +0100) it happened nospam wrote in :

Do you not think it is strange that it took 8 revisions for it to work (IF it works in A8)? In my view Microflip should focus a bit more on quality and not so much on putting out yet an other 'market segmented' version of the same thing. What does a new mask cost these days? Must be very expensive, especially if after 7 silicon versions with the same problem not a single customer will buy these chips. No return on investment.

When I bought the Z80, in the 19 eighties, a chip that has a much more complicated instruction set, but much easier to use, it worked on all advertised issues. Indeed hardware like this is becoming like Microsoft software, to actually use it you need the next service pack, and that then breaks some things usually. If John L gets his way, and we get RAM (say FPGA) on chip to for the microcode, then next thing is 'live updates' for your embedded system. I am out of here :-) LOL

e IP

or such

yep, programmable logic takes space, is not easy to design, and the software is not easy to get right, some might say that the big 2(3) are still trying :)

-Lasse

works in A8)?

The datasheet was issued in Feb 2009 and is still marked preliminary. It is a new part. I think they didn't make and ship a new silicon revision every

advertised issues.

A Z80 system with the features of the PIC18F14K22 would have taken a square foot of PCB stuffed full of chips, take 100 times more power to run at less than 1/4 the speed. Silly comparison.

Considering that the alternative is far more difficult *and* expensive... Nothing worthwhile is easy.

works in A8)?

buy these chips.

Why do you think they focus on "an other 'market segment'"? Masks (and more so the people that go with it) are expensive.

advertised issues.

It might be a "more complicated" instruction set but you're not having trouble with the "instruction set", now are you?

things usually.

If your widgets are as perfect as you claim, why would they need updates? You do test them, right? You never add features to your products?

On a sunny day (Sat, 08 May 2010 15:26:43 +0100) it happened nospam wrote in :

works in A8)?

Well, how come then they have 7 silicon eversions????

advertised issues.

No it is not silly, is is comparing quality of chip design.

Did you look into the IMX series from Freescale? Or the ARM SoCs from Samsung? Very cost effective.

advertised issues.

The Z80 did have at least one bug incorrectly setting the parity flag for some instructions when an interrupt acknowledge cycle occurs.

Dividing the complexity of the silicon by the number of defects the PIC design looks about 10 times higher quality.

On a sunny day (Sat, 08 May 2010 17:23:23 +0100) it happened nospam wrote in :

advertised issues.

The Z80 worked as specified, the PIC does not works in its much advertised mode of SPI. And I2C. Maybe you think SPI and I2C are new, well, I was one of the first users of I2c in those same eighties. Philips had no problem implementing I2c in hardware at that time.

Do any of the Samsungs have GbE? I'll have my guys look into them.

Marvell is coming around, after some pressure from the distributor and a lucky contact at the Embedded System Conference. But their eval board is $4K, presumably mostly earnest money.

John

$4K?! They really don't want business.

I remember when the Asian chip makers wanted $20K for a development system (and 50,000 pieces as the rock-bottom minimum order). That was back when $20K would buy several new cars..

Best regards, Spehro Pefhany

Sure, we used to buy Intel, Tek, and HP development systems that went higher than that. A lot of things have changed in the last 30 years. Everyone who wants business now gives this stuff away. I have to refuse a lot of dev kits that I know I'll never have the time to even play with. At the *most* these dev kits are couple of hundred bucks, complete with what used to be a few thousand bucks worth of software and licenses thrown in (I tell the Altera sales guy all about the new Spartan 6 features ;-).

The Samsungs have no internal MAC so you can attach any MAC-PHY you want. There are very few SoCs with GbE.

Way too expensive. Select a SoC for which $200 eval boards are available so you know you'll have some community support.

Seems like it's not that uncommon anymore.

It's apparently not expensive relative to the predicted lost sales of delaying the part by, e.g., a year. :-(

Unfortunately it's very unlikely that the bugs in your 18F14K22 would have stopped anyone who was planning to buy a million of the things per year from continuing to use the part...

I don't know if Microchip is doing it yet, but many microcontrollers today are designed in Verilog or VHDL, and this seems to have brought with it the mindset of "all designs have bugs" that permeates the software industry today. (Note that an x86-compatible CPU today from AMD or Intel often has dozens or even hundreds of errata... also note that for years now Intel and AMD have make some of their CPUs "soft" so that microcode updates applied at boot by the BIOS can fix some of the problems...)

---Joel

IP

for such

PALs and small FPGAs have been around over 30 years. If you limit to not much more tech, all the patents should be expired. Atmel, Freescale, Microchip, etc., all have nice multiplexing tech for the multifunction pins.