Bi-directional i/f to 5v logic
Harley
Posts: 997
I just realized I can't use the usual single resistor to i/f with a 74LS245 on 5v.
I've looked through the HOW TO SAFELY INTERFACE A 5V SIGNAL TO THE PROPELLER? forum. Wow, Beau's approach uses 2 transistors and 4 resistors per line. For a 8-bit i/f this is horribly bulky (16 Qs and 32 Rs)!!
Is there no other way?
To me it appears the 'culprit' is driving the -245 input low. One data sheet shows -0.2 ma @ 0.4v on the -245 input for the LO state and 20 µa @ 2.7v for HI. That LO low input state drive from the Prop means the resistor is fairly low; 2K if the Prop could output 0v at 0.2 ma). Has anyone found a better solution for this tri-state interfacing?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Harley Shanko
h.a.s. designn
I've looked through the HOW TO SAFELY INTERFACE A 5V SIGNAL TO THE PROPELLER? forum. Wow, Beau's approach uses 2 transistors and 4 resistors per line. For a 8-bit i/f this is horribly bulky (16 Qs and 32 Rs)!!
Is there no other way?
To me it appears the 'culprit' is driving the -245 input low. One data sheet shows -0.2 ma @ 0.4v on the -245 input for the LO state and 20 µa @ 2.7v for HI. That LO low input state drive from the Prop means the resistor is fairly low; 2K if the Prop could output 0v at 0.2 ma). Has anyone found a better solution for this tri-state interfacing?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Harley Shanko
h.a.s. designn
Comments
Also, I believe other users have mentioned using a 74HC245
Use a 74HCT245 instead.
-Phil
Harley was concerned about the bidirectional lines. To protect the Propeller when they're outputting, you need a series resistor. But this same resistor limits the drive current available from the Propeller to pull the TTL line low, when it's being used as as an input. While you may indeed have a working system, it's possible that you're operating closer to the margins than Harley is comfortable with. I suggested HCT logic instead, since the input currrent requirements are virtually nil, and the inputs are compatible with TTL voltage levels. By contrast, with HC logic running at 5V, a 3.3V input barely qualifies, if at all, as a "high".
If what you have works, by all means, use it. But next time 'round, you might want to give HCT devices some consideration. For me, at least, they've been the most compatible with 3.3V logic levels when they have to run at 5 volts.
-Phil
Hopefully it all continues to work with the 6502 accessing memory at high speed as well. At least the 6502 is not "seeing" the Propeller current sink directly; it's acting through the transceiver.
I've used HCT before and have no problem with them; I just like to build with LS because it's more "old school". I even buy plain regular TTL over LS when I can find it.
When the propeller lines are set to be outputs and the 244 tristates are gated on, the Schmidt inputs on the 244s translate the 3.3V propeller to 5V on the 244 outputs (the 244 inputs are supposed to be switching high at about 1.6V, then low at about 1.2V, so the propeller should be switching the 244 with plenty of margin). The 2.2k resistors play no role in this mode.
When the propeller lines are set to inputs and the 244 tristate outputs are set to high-impedance, the external 5V HD logic drives the propeller inputs thru the 2.2k resistors. The 244s play no role in this mode.
The only reason I used 74LS logic was because I had them in the junkbox, and they worked. I don't know any reason why any 5V Schmidt-input 3-state logic, like maybe HCT, would not work with this same configuration, but I haven't tried them.
It seems to work fine. It does take an additional propeller output pin to operate the 244 tristate gates, so it's not ideal, but since it only needs 2 chips and 16 resistors (which can just straddle the chips), it's not too hard to work with.
Thanks. Yes, that should simplify that i/f to just the 'series resistor' only scheme. Very good call.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Harley Shanko
h.a.s. designn
I say:
No no no no no........
The prop is well designed - Yes. Well tested - Yes. But to knowingly operate it outside its absolutes·- I disagree completely. Under your condition @5v it survived - great -·I'd be thankful and ensure that I didn't try that one again.
Phils approach is the way here .....
Post Edited (QuattroRS4) : 2/12/2007 5:45:25 PM GMT
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer
Parallax, Inc.