Using 4000 series logic to drive....?

Wow! You are so far from functional with MCUs. If you think it is hard to write working code for MCUs, please post your reply to comp.lang.forth and see the sorts of replies you get there.

Even a "few" logic chips need to be tested thoroughly before committing to production. The cost of a bug is even higher.

Not sure what "sometimes" means. Nearly every product shipped these days uses an MCU rather than logic.

Den tirsdag den 1. august 2017 kl. 18.48.03 UTC+2 skrev rickman:

while

ard, but that's far from true for the rest of us.

its of MSI are often pretty useful. I've recently used MSI in frequency syn thesizers and lock-ins, for instance. When you need an accurate two-phase c lock, a Johnson counter is good medicine. Also, as JL says, you don't have to program, document, and maintain a 74HC163.

ort

t's

have you done anything with Xilinx recently?

No, it's not hard to write working code, there are millions of Arduino-hackers writing working code day in and day out. Writing well-documented maintainable code that conforms to some kind of standard is significantly more time consuming.

Design for lil widget is done, the approach to doing what needed to be done with some chipz was obvious, the widget will be built, tested, shipped to DOZENS AND DOZENS of pleased users, money shall be deposited, and I'm sure I'll do something with a uP again eventually. And it was a fun way to spend a couple weekends. LIFE. WILL. GO. ON.

There is an ongoing debate here as to whether Altera or Xilinx have the more annoying tools.

Having a couple of ARM cores on the same chip as logic fabric has some huge benefits, but the price is complexity. We've done several good ZYNQ/DRAM/Linux projects, with impressive timings, but I don't want to learn the tools myself. Somebody's got to do architectures and board design, and I never really learned to type.

But really, we are packing so much functionality into products these days, you can't expect it to be simple.

I wonder what's next in FPGA architecture. DRAM on chip would be nice.

It's gonna be fine, brah. Nah nobody uses Forth. Nah I don't wanna spend $10 on some Sparkfun board I get my own board made for a couple bucks each.

The tools to compile for AVR/ARM targets with C/C++ are fine; on Linux you can just use vim and compile/upload from the command line, if you're hardcore. I wrote a custom allocator so I can use dynamically re-sizable std::vectors and smart pointers, etc. on an ATTiny, even, if the application requires that. EVERYTHING IS FINE

If you think it is any harder to document code that does the same thing as a few MSI chips, then you obviously don't document your MSI designs. There is no reason why one would require more documentation than the other.

I think you also aren't really interested in learning how to write good code or you would have crossposted to comp.lang.forth.

te:

s?

ou can

lution

nd in

just

doing

as they

old the

bug in

to be

ship

ified

any

rd

rth

d is

s a

is

ode

Dinosaurs Rick, simply Dinosaurs. Sure if you are programming really comple x algorithms on FPGAs it does get tough. Not for basic logic gates though. That's what they were designed for in the first place. Even first year stud ents at Uni get this one. The world has changed. Fair enough though if some body has a hobby and wants to use those old chips just as some use Vacuum t ubes.

Anything that relies on 'code' follows a variation of Murphy's Laws... it will fail, and the failure will be in such a way as to guarantee the worst possible outcome. ...Jim Thompson

Sorry to interrupt but I'm just curious - what does Forth have to do with anything?

"Dinosaur!" say some folks to a 36 y/o while talking about Forth.

PS: the audio folks pay good money for tube designs nowadays. I've got something in mind along those lines as well. shh. don't tell anyone.

You're cracked. There's a design document that explains how the unit works, but you really don't have to document how a 74HC163 works. It's in the datasheet, unlike custom-written code.

Where I come from, we know a trick worth two of that one.

I got into Forth in 1984 when I got my first and only Macintosh. It was a lot like programming HP calculators, which I'd been doing for years.

It was the integer-only math and the write-only quality of the code that made me ditch it all those years ago.

Cheers

Phil Hobbs

cros?

You can

solution

, and in

Ps just

opy

ff:

ow

on doing

es as they

o hold the

r a bug in

ble to be

o, ship

verified

, many

hard

.forth

dard is

g as a

ere is

d code

plex algorithms on FPGAs it does get tough. Not for basic logic gates thoug h. That's what they were designed for in the first place. Even first year s tudents at Uni get this one. The world has changed. Fair enough though if s omebody has a hobby and wants to use those old chips just as some use Vacuu m tubes.

Code, that mimic a 74HC163 and some glue logic is going to be so simple you would have to be very "talented" to mess that up

Advantage of the microcontroller is that it can so much more, so functional ity can be added with extreme low additional cost

Cheers

Klaus

Yes, I know... CHEAP ;-)

The only appliance problems we have here at our house are those pieces-a-crap that need a reboot >:-}

The latest fiasco was the supposedly pro-grade Kitchenaid double wall oven which somehow locked the doors without it being in self-clean mode.

No amount of touch screen maneuvering would unlock it. Had to go out to the entry panel and turn off its breaker... waited 10 minutes then re-powered... no joy, still locked.

But I got it loose by then commanding 'self-clean' from the touch panel, then 'cancel'.

I'd like to return to real KNOBS ;-) ...Jim Thompson

ile

d, but that's far from true for the rest of us.

s of MSI are often pretty useful. I've recently used MSI in frequency synth esizers and lock-ins, for instance. When you need an accurate two-phase clo ck, a Johnson counter is good medicine. Also, as JL says, you don't have to program, document, and maintain a 74HC163.

uPs are not just about typing. If you want to use the processor right for a optimimum system design, it's about learning the architecture and using it best possible. For example using self running DMA loops to have high throu gh put without CPU intervention and tweeking the design if needed using SW mode shifts

Cheers

Klaus

ile

d, but that's far from true for the rest of us.

s of MSI are often pretty useful. I've recently used MSI in frequency synth esizers and lock-ins, for instance. When you need an accurate two-phase clo ck, a Johnson counter is good medicine. Also, as JL says, you don't have to program, document, and maintain a 74HC163.

Even our trainees with less than 1 year experience in electronics know how to use the tools. The most known compilers from Keil or IAR are dead simple , you really just click the download and "play" button

Cheers

Klaus

For very loose values of "mimic". HCMOS has picosecond jitter. It's very hard to do that in an MCU even if there's nothing else going on. Good luck doing a high performance synthesizer that way. That's the last thing I used one in, about 2 years ago.

Of course. I'm doing a Cortex M4 design right now, in collaboration with the hunchbacks. The Cortex M's more-or-less deterministic interrupt latency (12 cycles coming from mainline code, but 8 coming from another handler) is a help.

Cheers

Phil Hobbs

Sure, the jitter is only going to be as good as the update rate of the DMA, so sub 100ns is doable. Otherwise, you could go for a Cypress PSoC that has glue logic on chip, so that is going to be pretty fast, maybe at the same level as the discrete logic

Cheers

Klaus

Yes. "Technology" has become synonymous with "coding."

Which makes the value of real electronics go up.