3V3 in, 5V out buffer driver

Of course you can't. Google groups are way too complicated to use.

;------------------------------------------------------------------- From: Vladimir Vassilevsky - view profile Date: Sun, Nov 26 2006 2:14 pm Groups: comp.arch.embedded, sci.electronics.design

Jim Thompson wrote: >>>>>>Incredible. Designing complicated circuit instead of simply using >>>>>>74HCT14 / 74HCT04 / zillion other HCT74xx and AHCTxx ICs. >>>>>You've just proved yourself an incredible dumb-s*it ;-)

Nothing happens. The 2.4V reads as a clear 1, the threshold is at 1.4V

subject: "3.3v 5v interfacing @ 15Mhz"

some said it would it would work but use a bit of current and went through the datasheet but Jim didn't like it and thought it was a hack, then everyone started yelling ;)

-Lasse

Interesting, but that must not be the link, because Jim wrote, "Go back and read all the way through the original thread on this topic. There you will find that another poster likewise criticized my opinion, THEN went and MEASURED a 74HCTxxx part at 3.3V input with VDD=5V, and measured the ENORMOUS overlap I warned about."

On a quick scan, I didn't see the current-measurement post Jim mentioned.

That's a good part, using two supply rails, working without current-consumption compromises, and providing 6ns delay for 2.7V to 5V translation: doing very well indeed. Still, it's a relatively-large 24-pin 8-line bidirectional, and we'd love to see something smaller. Another part in the 74LVC family?

Philips Semi, 74HCT: For dual supply systems the theoretical worst-case (VI = 2.4 V; VCC = 5.5 V) specification is: DICC = 0.65 mA (typical) and

For 3.0V input typical value is a HUGE 0.1mA.

I think your thermometer did not have enough resolution to measure temperature rise due to the HUGE 500uW power dissipation. :)

Win,

Of course it is good practice if within spec, which it is. I wrote I would look into changing HCT wherever applicable without realizing Jim was suggesting not just a better family - same as the HCT, but better input characteristics - but some other hoops to go through. For TTL levels, this might be worth it - if 2V high is really expected and if consumption really matters (as in the case I remembered). But from static 3.3V CMOS outputs, where you do get just 3.3V, well, it can hardly be justified, the current _must_ be much less than 4-5 mA at 3.3V if it is that at 2V.

Jim,

this has worked for ages. The "T" is for TTL compatible, i.e. they are specified to work with Vih down to 2V. I cannot remember now where and how, but I am sure I have used this a number of times without a problem. I am not sure how much they'll heat, but I am positive it will be less than standard series TTL, not to speak about S, AS, or F.

Yuriy,

This agrees more with my memories. Well, at least I know I never burned my fingers while debugging _that_ board :-). I thought I had discovered some 150 mW to be had - and I will look into it, I can make sure all inputs are practically all the time at 0 by software, and see if there is a measurable consumption difference when I have the chance.

Dimiter

------------------------------------------------------ Dimiter Popoff Transgalactic Instruments

------------------------------------------------------

W> >

Discussion here about how it should be done...

Also found the post with the 4mA...

From: Vladimir Vassilevsky - view profile Date: Tues, Nov 28 2006 10:48 am Email: Vladimir Vassilevsky Groups: comp.arch.embedded, sci.electr> I understand what you are saying, but your first few posts were not at

I have to admit that I didn't expect the cross conduction to be that high in the HCT at 3.3V input levels. Actually it made me rethink my position. The HCTs can be used as 3.3V -> 5V level translators, however the user must be aware of the power consumption.

BTW, the maximum of cross conduction appears to be at 0.9V input (5Vdd), and it is about 4mA.

VLV

...Jim Thompson

Very nice, I like that. The question is where to find such a circuit in ordinary packages, Also, Jim, do you think that's the away TI does it in their 74LVC4245A level translator?

Thanks for looking that up Jim. Sorry, but I'll have to call you on this one. We really want the current at 3.3V, not 0.9V. We all know the cross-conduction current peaks near the logic threshold, and can be pretty high for an instant as the signal passes through that point (it's interesting the value Vladimir measured is 8 times higher than TI's datasheet plot for their 74hct14 inverter, which was 0.48mA when running at 5.5 volts). But hey, what we care about is the standing or static current when the gate input is sitting at 3.0 to 3.3 volts, right? In TI's case that's about five times lower than the peak, or about 0.1mA. Unattractive, or even ugly, but hardly enough to condemn this suggested use of these parts as a hack. I agree that if there's a better IC it'd be good to use it instead, but we haven't yet turned up such a part.

Wait a minute, hold it!! I just found some proper two-rail parts from TI, their 74LVC1T45 single-gate and 74LVC2T45 dual-gate ICs, with less than 4 and 10uA of standing current. And the txb0104, a 4-line bi-directional IC with auto-direction sensing. Wahoo!

Likely. My Intel stuff is so old, it's in archive. Maybe I can retrieve it after Christmas and show a full device-level rendition.

I misread... sorry about that, but static current should be even lower, as you found...

Yep, the 74LVC1T45 looks like a winner.

But I'm still surprised at the 4uA and 10uA... I'd expect leakage.

...Jim Thompson