5v-3v3 bidirectional buffer

[barnacle]

I'm sure I've come across such an animal: bidirectional buffer with +5 and +3.3v rails, but I'm blowed if I can remember a part number... my search-fu is weak today. It feels like it might be a VLC part...

Any thoughts?

Thanks,

Neil

[fachat]

Something like 74lvc4245 ?

[barnacle]

That's the puppy! Thanks...

I have a coprocessor running at 3v3, but I'd like to run logic at 5v for speed even though the 6502 is happy at the lower voltage.

Neil

[drogon]

The PiTubeDirect project uses a pair of 74LVC245A chips to interface the 5v BBC Micro to the 3.3v Pi. It would rely on the 5v side being able to treat a 3.3v signal as logic 1 but they work well in the Beeb at 2Mhz.

-Gordon

[fachat]

Good point. 3.3V CMOS outputs typically work driving 5V TTL(!) Inputs. But if there are CMOS inputs on the 5V side - and this is more and more common - you may get into trouble. It may work depending on load and other conditions - but no guarantees.

Edit: and of course the 3.3V side inputs must be 5V tolerant

[barnacle]

That's my concern. I've carefully arranged the connection to the 3v3 STM to use only its 5v tolerant pins, but it's uncertain just how much noise margin there is (n't) going to be to 5v cmos on the 6502 side. It's probably OK for a home test, but I'd like to get it right.

Neil

[Proxy]

to throw my 2 cents into the mix, i've been using the FT240X in basically all of my current projects, it only has 3.3V outputs but i've never had any issues when connected to a 65xx Bus or any 5V logic (PLD/CPLD/etc).
the maximum voltage it can give out is 3.6V and the minimum for the 65C02 (according to it's pessimistic datasheet) is 3.5V when using a perfect 5V supply.
though honestly i wouldn't be surprised if it accepted voltages as low as 3V.

so you could just try it and it would very likely just work even at decent double digit clock speeds. though if you want to be safe, use some level shifters like the TXB0108 (which is bi-directional) or a buffer like drogon suggested.

[sburrow]

Don't know if this is closer to what you are wanting or not, but I'll be using the TXB0104D:

https://www.ti.com/lit/ds/symlink/txb0104.pdf

It doesn't have 8 bits, but it is bi-directional and super flexible.

Thanks!

Chad

[barnacle]

Another couple of possibilities, thanks. I need to consider whether the processor and memory can both deliver 2mA/pin; that's a big chunk of current...

Neil

[speculatrix]

sburrow wrote:

Don't know if this is closer to what you are wanting or not, but I'll be using the TXB0104D:

https://www.ti.com/lit/ds/symlink/txb0104.pdf

It doesn't have 8 bits, but it is bi-directional and super flexible.

There is an 8-channel version. But all the TXB family are a little limited on power output, I think.

[Michael]

There are some nice 8-channel TXS0108E 'modules' available, too.

[BigDumbDinosaur]

Proxy wrote:

to throw my 2 cents into the mix, i've been using the FT240X in basically all of my current projects, it only has 3.3V outputs but i've never had any issues when connected to a 65xx Bus or any 5V logic (PLD/CPLD/etc).
the maximum voltage it can give out is 3.6V and the minimum for the 65C02 (according to it's pessimistic datasheet) is 3.5V when using a perfect 5V supply.
though honestly i wouldn't be surprised if it accepted voltages as low as 3V.

so you could just try it and it would very likely just work even at decent double digit clock speeds. though if you want to be safe, use some level shifters like the TXB0108 (which is bi-directional) or a buffer like drogon suggested.

A lot of these devices have clamping diodes on their inputs that may make the inputs okay with five volts, at the expense of some additional loading on whatever is driving the inputs.  There may be something about such a feature in the data sheet.

However, note that PLDs produce TTL-level outputs, not CMOS.  A PLD output that has been driven high will not exceed ~3.4 volts under very light loading.

[barnacle]

Yes. I want to drive a 5v address bus from 3v3, and read and write a 3v3 data bus... so even with 5v tolerant inputs on the STM I need to ensure that it's driving hard enough. I'm planning on three 74LVC4245 for simplicity (even though it looks like my package choice is either DIP or TSSOP - the latter being a bit tricky to solder).

Neil

[barnacle]

An aside: PS/2 keyboard or mouse use a 5v Vcc so I assume TTL logic levels, right? So another level converter required there, since I intend to use the STM to talk to it.

Neil

[AndrewP]

barnacle wrote:

An aside: PS/2 keyboard or mouse use a 5v Vcc so I assume TTL logic levels, right? So another level converter required there, since I intend to use the STM to talk to it.

I haven't read the PS2 spec (which I probably should do) but I tested two keyboards and they both run up to about 5V on signals.

('scope set to 5V per div, clock in yellow, data in blue)