LS/HCT/etc. logic familes

I'm puzzled, that looks like a buffer overpowering the output of the bottom chip.

I remember using a dozen computers networked to a server that consisted of a dual 8" floppy to provide storage. A whopping 400k per disc IIRC. Years later we were wowed by a 70M HDD.

If we went back to then and told them about the computers we have I don't think they's believe it.

NT

It's really just the output impedance of the emitter follower into the other gate. So for large signal slew, the emitter follower is boss, for low voltage, the other gate is the driving device

Cheers

Klaus

Yes, I realised after a bit of thinking. It's not a circuit technique I'm familiar with.

NT

I've never seen it in discrete form before, but I have seen it in ICs:

If you think of a gate as a pair of MOSFETs instead of a digital component, all becomes clear. :)

Tim

It was taught in our Junior logic design course ('73) but it was noted that it wasn't very useful. That is, something to note but one probably would never find a real-life use for it.

So someone did find a use for it. I never did, though I kept trying. ;-)

We had those in the conference rooms but they've all been taken out (scrapped, I suppose). It's easier to photograph a whiteboard with a phone than to keep those things working (they're antiques).

Sure, since everyone has one, phones are a lot easier. Everyone can make their own "notes".

Just copy them to disk drives. Keep a couple of copies. Cheap!

LM5112 is similar.

re...

1. 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 was born in '73, thank you for making me feel young :-) (sorry, couldn't resist)

Our youngest child was born in '72 ;-) ...Jim Thompson

Kidz these days! ;-)

Level shifters...

...Jim Thompson

Problem is, they are expensive. For normal jobs a Zetex (now Diodes Inc) pair behind a gate plus a resistor is less expensive. When every cent counts then it has to be two discrete BJT.

We're paying 73 cents each. We don't care about parts cost as much as assembly labor, which includes setup for pick-and-place. So for us, simpler is usually worth more than pure cheap parts cost.

We do tricks with quad r-packs to get various resistance values and divider ratios from one value of pack. That keeps the BOM line count down, and reduces feeder setup.

The LM5112 lets us have an output V- below ground, which is handy some times, and would take more discretes to do. And it has peak drive of

IXDN602 is cool too. Dual fet driver, 2 amps, 35 volts, 7 ns rise/fall, 77 cents. With a driver that brutal, sometimes you don't need the fet!

In your case that is certainly the right strategy. On many of my designs I have to pinch pennies. Once I designed a nice TDR for a client which ran under $10 in production and I felt real good about the results. Then the client called and asked "Well, it is a really nice design but, ahem, how shall I say, could you design another one that is maybe only 80% as good but costs under $5?".

Yes, though I am usually after more oomph. 10 amp range and Rdson values or equivalents in the 1 ohm vicinity.

New part:

Dual 5 amps, 20 volts, 7 ns r/f, about 1 ohm each driver, 44 cents in quantity.

Might be nice in a dc/dc converter too. Drive a transformer directly.

The IXYS drivers go all the way up to 30A. Maybe they have a bigger one now, too.

I used the 15A versions years ago to drive IGBT modules (+/-15V), no problem.

Tim

Buy one - get one free, and ohm-range for Rdson. Great! Thanks.

With 20V abs max for VCC that may be a nice use for it.