https://assets.nexperia.com/documents/d ... C4245A.pdf
Octal dual supply translating transceiver; 3-state
"... designed to interface between a 3 V and 5 V
bus in a mixed 3 V and 5 V supply environment"

Data sheet also states, "TTL interface capability at 3.3 V"

VCC(A) supply voltage A -0.5 +6.5 V
VCC(B) supply voltage B -0.5 +4.6 V

So what's the point of powering the chip with two supply voltages? Seems to me that a plain old 'LVC245 @ 3.3 or 3.6V would do level translating, receiving and transmitting TTL compatible levels just as well as the specialized transceiver.

https://assets.nexperia.com/documents/d ... CH245A.pdf
"Direct interface with TTL levels"

How am I wrong?

_________________
"I am endeavoring, ma'am, to create a mnemonic memory circuit... using stone knives and bearskins." -- Spock to Edith Keeler

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

I think that the "TTL interface capability at 3.3 V" is the key. What they mean by this, as you can see from the Vɪʜ line in table 6 on page 5, is that 3.3V input will accept TTL levels for HIGH. So if you have TTL level inputs on your 3.3V side and need CMOS level outputs on your 5V side, this will do the trick.

A 74LVC245, on the other hand, though it will accept TTL level inputs at 3.3V Vcc (0.65 * 3.3 = 2.2 V), will generate only 3.3V outputs, which is enough for 5V TTL parts, but not 5V CMOS.

(Ok, BDD beat me to it, but I'll post this anyway in case it's marginally more clear. And there's a link to the 74LVC245 datasheet, if you need it.)

_________________
Curt J. Sampson - github.com/0cjs

Just to add to the explanations above.

I have mostly 3.3V LVC chips using TTL logic levels in my build. Except for the W65C816 that runs at 5V and uses CMOS logic levels. The 74LVC4245 up translates from the 3.3V LVC side to the 5V '816. Specifically that's the DATA pins (when reading) and PHI2, NMIB, IRQB, RESB, ABORTB and BE.

I suspect an 74HCT245 would work too but I haven't tried it. And I wouldn't because the 74LVC4245 has ridiculously fast propagation times whereas the HCT variant is slower. The Nexperia datasheet seems quite conservative compared to the actual propagation times I've seen.

... and whilst typing this 4 hours of loadshedding has begun and someone may have tried stealing cables out of the street mini-sub. I had a bit more to add but I should get some real work done but hopefully his is helpful.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

I have lived in central California for all 57 years of my life, and I am not aware of any systemic power problems in the past or present that could be directly correlated to the activities of short-sighted environmentalists. News bias could be at play here, as our sources of information clearly draw from different pools of "facts". What can't be disputed are the dates October 24, 2021 (the rainiest day in Sacramento's recorded history) and September 6, 2022 (the hottest day in Sacramento's recorded history). Records are meant to be broken, but these two fell disturbingly closely together, and disturbingly recently, at least in my humble opinion.

_________________
Got a kilobyte lying fallow in your 65xx's memory map? Sprinkle some VTL02C on it and see how it grows on you!

Mike B. (about me) (learning how to github)

Some interesting issues here. Mike, you've commented on the reality of climate change, and BDD seems to feel that upgrading infrastructure involves a lot of environmental red tape.

I expect other members will have some opinions to express in response!

Let me quickly point out that it's not an "either-or" situation. IOW, our efforts will be misdirected if we attempt to debate whether it's Mike or BDD who is "right." That is, one may reasonably conclude that climate change is real and at the same time believe that upgrading infrastructure involves a lot of environmental red tape. I hope the discussion doesn't become impoverished and polarized with nuance-less, for-or-against thinking.

-- Jeff

_________________
In 1988 my 65C02 got six new registers and 44 new full-speed instructions!
https://laughtonelectronics.com/Arcana/ ... mmary.html

Oh lordy, yes. Couple an unfortunately uneducated population with easily accessible high current carrying copper and... community policing is sometimes not a lot of fun.

I remembered that I wanted to come back and comment on the '4245 propagation times. Specifically because I thought someone said they read - or saw - that it was slow. Like on the order of a 100ns. I can't find that comment now so maybe I just imagined it or read it in another thread.

Either way though the 74LVC4245 datasheet shows a typical propagation time for up translation from 3.3V to 5V of around about 3.3ns. And that is ridiculously fast. And that's not the end of it either. If you're driving only one or two other LVC chips then your propagation times can be even faster. The 74LVC4245 datasheet shows a minimum propagation time of 1ns. I kind of interpret that as some engineer being asked to put a minimum number down and just throwing his hands up and going fine, whatever, 1ns it is.

In part I say that because there may be cases where its propagation is even faster than 1ns but how useful or usable that is debatable.

I'm actually using the 74ALVC164245, not the the 74LVC4245. It's got 16 data lines rather than the '4245s 8 data lines. And as an added bonus the '164245s datasheet lists the propagation time for up translation from 3.3V to 5V at only 2.5ns. And a minimum propagation time of 1.2ns. I mean that's fast. That's really, really fast. So fast in fact that it is basically irrelevant to - say - address decode timings on a 65x based system running at 20Mhz.

And the 74ALVC164245 seems to truly, stably manage that nanosecond propagation time. With caveats. I'm using a Siglent SDS1204X 200Mhz oscilloscope at 1GSa/s per channel to measure the rise and fall time of the '164245. That's not ideal because I'm getting maybe two or three digital samples around the high to low transition. And the analogue bandwidth SHOULD be attenuating the signal a lot. But it doesn't seem to be that bad. So maybe the baby Siglent is a good deal more capable than it's officially rated. (And good high frequency probes make a difference). And now with all of that said and also imposing some 'enthusiastic' interpretations on the dotty signal... I think I can see real propagation delays on a 74ALVC164245 of about 1ns when it's driving only a single Alliance SRAM IC on one side and a 65C816 on the other.

Why did I write all of this? I don't really know. It's just fun to mention that the '164245 is a very fast voltage translator with output enables. That's pretty cool. And I didn't even mention the 74ABT245 yet.