[beagleboard] Interfacing a 7400 series with the BeagleBone Black

You'll have to take steps to scale the voltage down to the BBB's 3.3 V I/O level. The best way to do this is with dedicated voltage-level translator ICs. But depending on your application, you might be able to get away with current-limiting resistors in series from the 74xx output to the GPIO (you'll have to do some work to find the right value), and with the right 74xx selection, you won't need to put anything between BBB outputs and 74xx inputs (that will depend on what the 74xx Vhi threshold is).

Or maybe you can use 3.3 V 74xx variants.

You'll have to take steps to scale the voltage down to the BBB's 3.3 V I/O level. The best way to do this is with dedicated voltage-level translator ICs. But depending on your application, you might be able to get away with current-limiting resistors in series from the 74xx output to the GPIO (you'll have to do some work to find the right value), and with the right 74xx selection, you won't need to put anything between BBB outputs and 74xx inputs (that will depend on what the 74xx Vhi threshold is).

It's reasonable to do that when you have a 5 volt system with a number
of lines, and can't substitute the part (a 5 volt display, for
instance).

I would NOT use current limiting resistors, I don't thing that they
solve the problem.

As I remember it, and they could have changed things and this is not
as much of a problem now, the way that the pin is internally
constructed involves back biased diodes which connect to the power
supply rails. These back biased diodes, if forward biased, make an
SCR, which once turned on, stays on until the power is interrupted.

The voltage causes these SCR's to become functional, then fire. The
SCR then uses the internal power supply to provide enough energy to
destroy the I/O structure in the chip.

Current limiting resistors do not help, because it's not the current
coming in that does the damage, it's the relative voltage level.

This is called a "parasitic SCR" and has destroyed chips in the past.
Whether or not it is a problem still, I do not know.

You'd want to have chips (to connect the 5 volts to) that have 5 volt
tolerant inputs. These processors do not.

Better to use 3.3 volt logic everywhere possible, and level translate
only when talking to a 5 volt system that cannot be redesigned.

Talking, in this case, means signals going either direction.

Or maybe you can use 3.3 V 74xx variants.

This would be my preferred solution, yes.

Harvey

Thank you for all your advices.

There shouldn’t be any problem if i use a bidirectionnal voltage shifter ?

Thank you for all your advices.

There shouldn't be any problem if i use a bidirectionnal voltage shifter ?

I use such a thing to go from 3.3 to 5.0 volt systems any time I need
to deal with that kind of thing. I typically use the 74LVCH8T245PWR
for 8 bits of data, and the 74LVC1T45 for single bit shifting. These
packages are surface mount, though, and you may not be able to handle
them easily, although there are surface mount to dip package adaptors.
Harvey

Hello,

What I’m missing is if the 74LS02 is a fixed requirement, or if there is an option to use the 3.3V sibling. Thus shifting the problem to the xx74xx02 inputs. Because if you kill the 74LS02 with your experiments it might hurt less than if you kill the BBB.

I would NOT use current limiting resistors, I don’t thing that they
solve the problem.

This has been done for decades. I don’t see why it should not work for the Sitara.
If the I/O voltage is 3.3V and we assume the clamping voltage 0.3V the overvoltage is 5V-3.6V = 1.4V. As an indication of clamping diode capability can we take the pin current capability. Eg. 4mA (varies over pins). Than we would only need a 350Ohm series resistor. → Take 10k and you cover all the wrong assumptions ;-).
You can take even higher resistors (eg. 100k). But you will limit your maximum frequency because somewhere you will not charge the capacity of the Sitara input buffers fast enough. High resistors values can also increase EMC susceptibility.

Further more pins USB0_VBUS and USB1_VBUS should withstand 5V. But I assume you need USB and you need 4 inputs.

Chilli

he said he wants to test ttl ics.

I would get a low cost TTL chip tester off of ebay. cheaper then
replacing a few BBB boards

I overlooked Hadriens intermediate post. And you are absolutely right. Not only regarding killed beagles but also regarding the time needed to set up a simplified TTL tester.
On the other hand it seems Hadrien ist just looking for a ‘project’ to make the first steps. I think if he connects all GPIO with 100k resitors to a bread board and the supply with a 50mA fast blow fuse he can’t do too much wrong.

Chilli