Negative logic

I'd like a fast pulse from ground to -5V, is there anything wrong with running logic (74hc10) with Vcc=0 and ground at -5V? I'm not sure how I'll trigger it... maybe on the edge with an RC differential (high pass) and diode.

Reply to
George Herold
Loading thread data ...

No problem at all, it just looks funny on a schematic

Capacitive couple the signal like you suggested

Cheers.

Klaus

Reply to
Klaus Kragelund

Apart from possibly oscillating during transitions, that should be fine. I'd use a supply pour and good bypassing.

If there were still BFT92s, one of those makes a good level shifter--base grounded, emitter to CMOS via 1k, collector goes to 1k resistor to -5. An MMBT3640 is about the best you can still get.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

Phil Hobbs wondered

What's wrong with the BFR91?

5 GHz, have a plastic bag full. Many shops have those here, makes a great oscillator, amplifier, there is one in my antenna pre-amplifier for 70 cm...

formatting link
about 40 'merrican cents.

Reply to
<698839253X6D445TD

Thanks Klaus, Phil. I hadn't thought about a level shifter. With a 2n3906 it's a bit slow ~30 ns rise time. (1k ohm R's as you suggested) The RC diode thing works fine.

GH

Reply to
George Herold

That's fine. You can easily level-shift the input, with a bypassed zener or bandgap. Or as you say, AC couple it, maybe with a dc-restore diode.

If your pulse width is short, you could also level-shift the output.

If HC isn't fast enough, you can use one of the advanced families, AC or AHC or ACT or one of the Tiny parts. They transition in under a ns.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

It's an NPN.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

Yeah, the 3906 is a dog at low current.

Cheers

Phil Hobbs

(Who still hasn't forgiven NXP)

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

And obsolete.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Why do you want a hc10? That is a triple 3-input nand. There is no hysteresis to eliminate false triggering.

I'd suggest a 74AC14. There are 6 per ic. You can tie all the inputs together and connect 300 ohm resistors to the output. This will easily drive 50 ohms.

This is a place where LTspice is invaluable to sort out the parameters you need. Here is an example:

Version 4 SHEET 1 1204 680 WIRE 448 -64 368 -64 WIRE 480 -64 448 -64 WIRE 544 -64 480 -64 WIRE 368 -48 368 -64 WIRE 544 -32 544 -64 WIRE 480 -16 480 -64 WIRE 368 48 368 32 WIRE 480 112 480 48 WIRE 544 112 544 32 WIRE 544 112 480 112 WIRE 624 112 544 112 WIRE 656 112 624 112 WIRE 768 112 720 112 WIRE 816 112 768 112 WIRE 544 160 544 112 WIRE 544 256 544 240 WIRE 592 256 544 256 WIRE 624 256 592 256 WIRE 544 272 544 256 WIRE 544 368 544 352 FLAG 368 48 0 FLAG 448 -64 Vin FLAG 768 112 Vout FLAG 624 112 A1I FLAG 544 368 0 FLAG 592 256 VEE SYMBOL digital\\schmtinv 656 48 R0 WINDOW 3 8 54 Invisible 2 SYMATTR InstName A1 SYMATTR Value Vhigh=0 SYMATTR Value2 Vlow =-5V SYMBOL voltage 368 -64 R0 WINDOW 3 -161 149 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V1 SYMATTR Value PULSE(5 0 1u 1u 1u 1u) SYMBOL zener 528 32 M180 WINDOW 0 24 64 Left 2 WINDOW 3 24 0 Left 2 SYMATTR InstName D1 SYMATTR Value 1N751 SYMBOL res 528 144 R0 SYMATTR InstName R1 SYMATTR Value 1k SYMBOL voltage 544 256 R0 WINDOW 3 52 67 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V2 SYMATTR Value -5 SYMBOL cap 464 -16 R0 SYMATTR InstName C1 SYMATTR Value 1uf TEXT 384 -216 Left 2 !.tran 0 10u 0 10n TEXT 384 -248 Left 2 ;'Schmit Inverter at -5V TEXT 640 144 Left 2 ;74AC14

Reply to
Steve Wilson

The fastest rise/fall in logic is with Schmitt or flip/flop circuits. So, a tinylogic RS flipflop with no critical edge requirement on the inputs can be used at -5V, with PNP transistors to level-translate the /SET and /RESET signals from +5V logic.

NL17SZ74 would do, and gives both output polarities, guarantees 32 mA outputs

I'm inclined to disparage Schmitt solutions in CMOS, because it sucks lots of power supply current near thresholds, and that needlessly complicates the power filtering situation.

Reply to
whit3rd

some hct14 spec only a few 100 uA additional current in the threshold "window"

Reply to
Lasse Langwadt Christensen

Very few circuits spend much time at the threshold.

I really don't see how you plan to use a d-flop to accept a single-ended pulse input. Once you use the clock or set, you need some method to reset the flop to accept another input.

You can convert a D-flop into a single-shot, but this gives a fixed width output pulse. I'm not sure that's what Harold wants.

Reply to
Steve Wilson

The 2n4403 was worse.

I'm afraid I'm oblivious, having never been 'up close and personal' with 'em. My only comment is that technology is a wave, I find myself mostly surfing on the back end, which ends up being about 'life time buys' at some point. (or finding a new vendor, redesign..)

George H.

Reply to
George Herold

I interpret 'fast pulse' to mean a short rise time/fall time, so was looking to keep a slow input or level translator from causing slow-slewing output. For 'short pulse' you can do other things, like capacitor-coupling into a clamped input and applying gain.

Reply to
whit3rd

Right, sorry, my mis-type I meant hc14... I ordered a bunch (25) of the tiny parts from DK... maybe I should have gotten more? If I parallel three it's ~100mA.

George H.

Reply to
George Herold

It used to be common to drive the -V powered gate from a +Vdd gate through something on the order of 2x 10kR resistor divider. According to datasheet , typical Ci =3pF, so driving through a 5k effective source resistance ge ts you a nominal transition time of 30ns, which makes for a 10ns transition through the threshold region, worst case, and that is less than 1/2 Tpd, m eaning you can't make it go any faster.

Reply to
bloggs.fredbloggs.fred

One virtue of Schmitts is that they don't have a lot of shoot-through current, and they don't oscillate in the logic transition region. The small shoot-through current is only in the input stage, which is designed to operate between the rails; a non-schmitt gate that goes linear, or oscillates, can have full shoot-through in the big output fets.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

I bought a couple of reels of BFT92s, as well as about a dozen reels of BF862s and CPH3910s, so we're fine for internal use. Designs for customers and licensing need to be all current production, which is the problem.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

If you parallel sections for output, and there's any (delay or threshold) mismatch, there is shoot-through anyhow. My bad experiences with CMOS Schmitt devices are with ancient CD4584 and the like, with poorly guaranteed thresholds.

Reply to
whit3rd

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.