LS/HCT/etc. logic familes

One chip programming in the production is zero cost, only tact time is affected (longer production time)

Programs that is properly documented, well I guess you don't interact a lot with software engineers. If you compare the design journals HW guys does to that of the SW guys, it's really insignificant

Cheers

Klaus

Any decent low price microcontroller has all those functions inside, so that does not count. You sometimes need a pull-up for the reset pin, but that's all

Cheers

Klaus

True. Most software guys don't document anything. "Just read the code." "The latest version is in the VCS repository."

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Yes,

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
Website: http://seventransistorlabs.com

And everyone knows schematics are self-documenting.

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
Website: http://seventransistorlabs.com

Really? They come with crystals and regulators and *reliable* reset controllers? So you can drive an MCU from 12 volts with no other external components other than a resistor? Wow, they've come a long way.

--

Rick C

The equivalent would be to say that netlists are self-documenting.

They are if sheet 1 is the block diagram and table of contents, and if there's an associated design notes document.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

..

I learned that trick from one of Don Lancaster's books when I was about 20. He called it "data selector logic". It was probably the CMOS Cookbook. And yes, in general you need both true and complement of the fourth signal. ( Of course you could use a 4067 and get one extra bit.) ;)

I actually used a 5-bit version in the failover logic of the first civilian DBS system, circa 1982.

Cheers

Phil Hobbs

It's nice

But, won't you get momentary shorting of the gate due to the "propagation delay" in the emitter followers? But I guess that part can handle a brief short of the output...

Cheers

Klaus

40109 level shifter

piglet

Won't you get "momentary shorting" driving a capacitance in the first place? Or a transmission line*? ;-)

(*No. A transmission line /is/ actually a resistance, until stuff reflects down it, that is. :) )

The "delay" of an emitter follower is very small, actually; at most, it's the forward recovery of the B-E junction, negligible to HC logic (fractional ns?). It manifests as an R || C, as far as the gate is concerned. :)

The interesting observation to make, is that the gate is ultimately a "complementary" pair of transistors driving a node. The limit is thermal only*. They're _absolutely fine_ driving a short for, eh... milliseconds at a time?

(In practice, you can get away with shorting logic outputs for indeterminate time periods. It's not /recommended/, but you can get away with it.)

Indeed, the follower /helps/ the direct-connected one, by forcing the load along. Where it would be dumping into a large (in my case 4nF) capacitance for hundreds of nanoseconds, instead it's only doing it for 50ns.

By the same reasoning: it's not recommended to parallel Schmitt trigger gates (HC14 in my case), but in practice, they're /monolithically matched/, so they'll all transition within fractions of a nanosecond, and even if they don't, they'll only be "fighting" each other for that same fraction of a nanosecond.

The underlying lesson being: digital theory** is great for static validation, but when it becomes necessary to take a hard, up-close look at it -- on scales similar to the rise or propagation times of the gates -- it fails to be representative of the real device physics, and we must go back to analog analysis methods instead. That is to say: I may be using 74HC gates, but it's an analog circuit. :-)

(**That is, the connectivity rules and event based time-domain analysis that most digital simulators employ.)

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
Website: http://seventransistorlabs.com

Oh why don't we just put a novel in there too! :^)

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Contract Design 
Website: http://seventransistorlabs.com

I sometimes use TLC59281RGE

It has 16 outputs, 17V open drain with resistor-programmed current limit. It's a LED driver, but is quite ok for driving logic.

Allan

I believe it's possible to get XOR/XNOR with the '151 using asynchronous feedback, without a race.

I'd want to prove that to be sure. ;-)

The issue is, who can find, and make sense of, and modify, a design two years after the author has vanished?

And when you open a file or look at a drawing, is your only possible response "what the hell is this?"

I have a 71 (so far) page design notes doc on a current project.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Well, if all you have is 12 you'll need to bring that down to 5 or 3.3. But yeah an 8 pin chip like the ATTiny45 will work fine with just a supply bypass capacitor. It won't run as fast with the internal oscillator as compared to a crystal, but 8-12MHz is good enough for some tasks. The Tiny series has a reset pin but it isn't required for most normal operation; its fuses can be configured to give you simply 6 GPIOs.

There's also often plenty of code data space to flash a bootloader into even the little chips so you don't have to muck about with high voltage programming; the chip powers up with the bootloader ready to accept a serial transfer for a few seconds, if it gets the handshake it expects you can send the new code over regular TTL serial and the bootloader will flash it into program memory for you on the fly

This is an 8 pin AVR programmer:

Plop chip in, connect to USB, hit Enter, wait 15 seconds, remove, plop into circuit

The Silicon Labs EFM series has build in boot loader over UART, precise oscillator at 29MHz (how they came up with that frequency up I don't know), precise reference, reset controller etc.

It does not have 12V supply, but you can drive it from anything from 2V to 5V, which is also seen in many other controllers. And then it's cheap

It's funny that 2 CD4000 IC's with mayby 100 transistors cost the same as a microcontroller with tens of thousands of transistors

Cheers

Klaus