Driving 74LS from 3.3V uP

I should add to be clear that there's no output on the display when the bench supply goes into overcurrent, the open-collector output drivers are not turning on.

Double-check connections? Measure power supply voltage to the chips themselves?

Are the ATTiny's pins 5-volt tolerant, when the chip is operated from a 3.3-volt supply?

Based on what I see in the 74LS74 data sheet, the flip-flop's input pins are almost certainly going to be trying to pull the Tiny's output pins up to around 4.3 volts. This may very well be turning on some overvoltage-protection diodes in the Tiny. The 74LS74 data sheet suggests that Iil is limited to around 3.2 mA per pin, worst case, so 80 mA surprises me.

... Looking at the ATTiny85 data, it appears to me that it is *not* tolerant in this way. Absolute maximum voltage on any pin is limited to -0.5V, to Vcc+0.5V. So, you're overdriving it.

You'll need a different sort of driver circuit to do this safely.

Ah, thank you. Seems like I should be able to swap in a 74HC4511 and run it off the 3.3 instead?

Some of those old LS parts may draw significant crowbar currents as the supply is ramped up... choking the "eBay special". ...Jim Thompson

It's an LS47, not an LS74. But that doesn't matter.

An LS input pulls itself up to about 1 volt, and weakly at that. That shouldn't bother the uP.

I'm currently rewriting the code to test if it might be a "sequencing" issue. Hold the outputs tri-state, bring up the boost supply via the enable pin, then set the pins to output and latch the data.

You're right, I mis-read. But, as you say, it doesn't matter.

The TI datasheet for the 74LS47 shows the equivalent input circuit as having a resistor (nominal 20k) tied to Vcc, and then fed through a zener to the input pin. It looks as if it'll try to pull up to "Vcc minus one diode drop" - not all that strongly, but that may be enough to pull the ATTiny's pin up above its absolute-maximum limit.

The O.P. may want to measure the Tiny's output-pin voltages, with the

Assuming that the problem isn't due to a short circuit or miswiring somewhere, the idea of switching to a CMOS 7-segment driver and running the chip from the 3.3-volt supply would not be a bad idea... this would honor the Tiny's absolute-maximum voltage ratings. If I understand correctly, the 7-segment drivers are open-collector / open-drain, and so the 7-segment display itself could be run directly from the LiPo battery. This would avoid the need for a boost converter, and should increase electrical efficiency - fewer unnecessary volts to dissipate in the 7-segment current limiting resistors, and so smaller resistors could be used.

That's not a zener, it's a Schottky diode. But the resistor-schottky node is clamped to ground through two more diode drops, so it can't go very positive. TTL and LSTTL don't pull up very far, around +1 volt.

I suspect something else is going on.

MCU I/Os are typically protected by an ESD diode which will typically handle the current from 5 volts through a 20 kohm resistance which would be less than 100 uA. That's the real issue, the current through that diode. Too much and trouble. Until it blows, it won't let the voltage rise above the "danger" point.

But your are right, this is not a good idea. Better to use one of many alternative solutions to the problem.

What about a driver chip that runs from 3.3 volts? TLC5916-Q1/TLC5917-Q1

There's no problem with a 3.3v cmos uP driving 5 volt LS logic.

You might run into the same problem again when they do, due to filament cold resistance. You could soft start them.

NT

You are facing that eternal cosmic question: where the hell is the current going?

It's only one tube. I tested the tube/converter circuit in isolation and it appears to light all the digit segment combinations of one tube OK

ts

lls

he

say

om

's

ing

which at 85% or less efficiency (Chinese special) = 1.65W or more from t he LiPo or 1.65/3.5= 0.5A *average* with possible 1A peaks for the boost. Ckt is not the only thing hopelessly underpowered around here.

The bench supply is reacting to instantaneous current, so 2 segements is 40 mA x 5/3.5/0.85= 67mA and possible 100mA current ripple peak in the boost inductor. Use a bigger PS or adjust that current limit upwards to 1A. Ther e is no problem here, other than underpowered PS.

All segments from a cold start?

I did some more experimenting this afternoon and I think something along these lines is the cause of the issue.

Had the driver and incandescent display in isolation running from the bench supply at 3.7 volts. If I connect up the bench supply and then turn the power supply on from the switch, the circuit latches up and the supply goes into protection.

If I allow the bench supply to start up, with the leads to the protoboard unconnected, and then connect the leads, the display and