0
\$\begingroup\$

I want to drive IV-9 numitron tubes with a shift register. These have 8 segments which each need about 4V/20mA DC to light up.

I’m considering to use the CD74HCT4094E shift register, but am unsure if I interpret the specifications correctly. In the data sheet, it says:

enter image description here

Does this mean all 10 outputs can output 20mA each, simultaneously? Or does it mean something completely different? In all other tables in the data sheet, typical currents mentioned only go up to 4mA.

\$\endgroup\$

2 Answers 2

1
\$\begingroup\$

The one you have circled is the diode current and refers to the pin clamp diodes, not something you would normally want to be forward biasing. So forget about that.

The line below that says you can sink or source up to 25mA per output pin without damaging the chip short-term (at room temperature) however the line below that takes a bit of that away since the current that goes through the outputs must enter or leave through ground/Vcc pins the total of all output currents must not be higher than 50mA again for short term at room temperature. A good design will never even get near the absolute maximum specifications in normal operation.

These specifications say nothing about how well the outputs will drive the current, they are about short-term chip damage. It is likely that operating near these limits would reduce the reliability of the chip because the on-chip current densities would be very high. This kind of failure is worse at higher temperatures.

If you read the datasheet specifications, when sinking current it is tested to work well (330mV drop) when sinking 4mA on a given output.

Bottom line, no it won't do a good job, use a chip designed for it (eg. 4511), discrete transistors with a decent SOA (eg. 2N4401), transistor arrays (eg. ULN2803) etc. If you want to stay with a shift register, TI has some nice power shift registers in their TPIC series (but take care as to layout when you mix fast clocked logic and power switching).

Keep in mind that the old Numitron tubes are direct-viewed incandescent filaments and as such have a surge at turn-on so you should have some margin on top of the 20mA nominal requirement. Dating myself, perhaps, but I have used SN7447 drivers (we repped a Japanese company that made clones of the RCA parts). Of course in 2018, the tubes and the chips are long obsolete.

\$\endgroup\$
2
  • 1
    \$\begingroup\$ Thanks for the comprehensive answer and all of the tips. It's it rather hard to find information on handling numitrons otherwise. \$\endgroup\$
    – Fii
    Mar 15, 2018 at 10:17
  • 1
    \$\begingroup\$ It is worth noting that the absolute maxima limits guarantee that the chip is not quickly damaged. These limits do NOT guarantee the normal behaviour, that is what the operating conditions section is for. \$\endgroup\$ Mar 15, 2018 at 16:26
1
\$\begingroup\$

No, it just means that you can load the protection diodes at the outputs with max. 20mA without damaging the chip permanently.

This may happen if you apply e.g. a voltage < -0.5V or > Vcc + 0.5V at the outputs (no good idea).

"Absolute Maximum Ratings" give only answer to the question about what you can do without damaging the chip. That doesn't mean that it works properly under those condidtions (e.g. that the output voltage level will be in the required range).

(That is obvious if you lokk e.g. at the DC supply voltage range Vcc=-0.5..7V. Of course the IC doesn't work properly with e.g. Vcc=0.2V; it's just not damaged)

If you want to know what you can do under operating conditions, you have to see the specifications under "operating conditions".
If you can't find the interesting values in the datasheet of the specific IC you will probably find it in a datasheet for the logic family.

\$\endgroup\$
3
  • \$\begingroup\$ 20 for diodes , 25 for driver and 50 for power pin total \$\endgroup\$ Mar 15, 2018 at 9:56
  • 1
    \$\begingroup\$ yes, of course all conditions have to be satisfied; not just one. \$\endgroup\$
    – Curd
    Mar 15, 2018 at 10:08
  • \$\begingroup\$ Thanks for the explanations. It's much clearer to me now. \$\endgroup\$
    – Fii
    Mar 15, 2018 at 10:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.