How much current at what voltage from digital?

Thread Starter

hp1729

Joined Nov 23, 2015
2,304
To answer a question often asked, how much voltage at what current can I get out of a logic gate? Well here's some answers. Rated values from the data sheet are quickly exceeded. Needless to say life expectancy beyond rated levels can't be predicted.
Inverters were used but the results might be the same for other normal output gates. Drivers were not done. Maybe next time I will get odd family members.
Outputs were tested at various loads. For high out the load was put between VCC and the output. For low out the load was put between the output and VCC. Measurements were taken quickly.
No smoke was released during any of these exercises. There is still the question of how long the output could be driven at these levels.

What level of pull up resistor is needed for what chips? In order for the input to see a valid high, just about any resistor will do. Less than 50K for 74xx, or 1M for CMOS. So put away your calculators on that question. Just about anything works.
 

Attachments

WBahn

Joined Mar 31, 2012
26,145
It's not at all surprising that you can significantly exceed the specs and that the part will still work. For a while, at least. In order for the specs to be valid over the range of conditions and process variation that they are guaranteed for, they have to be low for the typical part operating under typical conditions. But, as you note, how long your design will work when you exceed the rated specs is unpredictable. Also, the rated values are for a range of temperatures and supply voltages, not just the ones that you happened to use. Intentionally abusing parts on a one-off hobby project is one thing, but if you do this on something that you intend to last or sell or anything similar then this is extremely poor design practice.
 

Thread Starter

hp1729

Joined Nov 23, 2015
2,304
It's not at all surprising that you can significantly exceed the specs and that the part will still work. For a while, at least. In order for the specs to be valid over the range of conditions and process variation that they are guaranteed for, they have to be low for the typical part operating under typical conditions. But, as you note, how long your design will work when you exceed the rated specs is unpredictable. Also, the rated values are for a range of temperatures and supply voltages, not just the ones that you happened to use. Intentionally abusing parts on a one-off hobby project is one thing, but if you do this on something that you intend to last or sell or anything similar then this is extremely poor design practice.
Agree very much.
 

andrewmm

Joined Feb 25, 2011
574
Taking look at the 7404 data sheet as an example
compare it to your numbers,

http://www.ti.com/lit/ds/symlink/sn74s04.pdf

Volts Out Low ( Vol) is specified as guaranteed less than 0.4 v at a current of 16 mA.
0.4 volts is the Maximum level used to guaranteed to be a zero in TTL.
( its 0.6 v at the receiver, to allow for noise etc )

But Volts Out High ( Voh ) is specified as guaranteed over 2.4v at a current of 0.4 mA .
2.4v is the logic high min.

As you can see, the chip can sink 16 mA and still be a logic level, but only source 0.4 mA.
That's why historically LEDs etc have been pull down connected, as old TTL could sink a lot more than it could drive.

Your Vol at 16 mA and 0.12 volts looks respectable margin to me,

Your 15 mA at Voh of 2.36 V looks very impressive, Almost too impressive,
wonder if its not an old 7404 as such,

Later logic from different manufacturers , even if it is stated as a 'standard' 74 series, came with a symmetrical drive and sink of 25 mA and up, depending upon manufacturer.

Its fun when you have a circuit that has worked for years / decades, and some one replaces one of the chips with an 'identical'..
as ever, read the data sheets, keep within the limits and all is happy.
run by rule of thumb, and most of the time you will be happy, BUT
when rule of thump fails your in trouble de bugging ,,,,,

design conservatively, and you are looking for logic errors,
design rash, and your looking for logic errors and funnies,
and that can be a very long road.
 

Thread Starter

hp1729

Joined Nov 23, 2015
2,304
Taking look at the 7404 data sheet as an example
compare it to your numbers,

http://www.ti.com/lit/ds/symlink/sn74s04.pdf

Volts Out Low ( Vol) is specified as guaranteed less than 0.4 v at a current of 16 mA.
0.4 volts is the Maximum level used to guaranteed to be a zero in TTL.
( its 0.6 v at the receiver, to allow for noise etc )

But Volts Out High ( Voh ) is specified as guaranteed over 2.4v at a current of 0.4 mA .
2.4v is the logic high min.

As you can see, the chip can sink 16 mA and still be a logic level, but only source 0.4 mA.
That's why historically LEDs etc have been pull down connected, as old TTL could sink a lot more than it could drive.

Your Vol at 16 mA and 0.12 volts looks respectable margin to me,

Your 15 mA at Voh of 2.36 V looks very impressive, Almost too impressive,
wonder if its not an old 7404 as such,

Later logic from different manufacturers , even if it is stated as a 'standard' 74 series, came with a symmetrical drive and sink of 25 mA and up, depending upon manufacturer.

Its fun when you have a circuit that has worked for years / decades, and some one replaces one of the chips with an 'identical'..
as ever, read the data sheets, keep within the limits and all is happy.
run by rule of thumb, and most of the time you will be happy, BUT
when rule of thump fails your in trouble de bugging ,,,,,

design conservatively, and you are looking for logic errors,
design rash, and your looking for logic errors and funnies,
and that can be a very long road.
I didn't do a 74S04. I haven't seen a 74S04 in a circuit since early computers sat on floors. The data sheet quotes -1 mA high and 20 mA low for 74S04. A bit higher than 7404.
Yes, 7404 is old technology. Some had current limiting on the high side that allowed driving an LED using the current limiting as the load resistor for the LED, safely limited to 15 mA or so. No you couldn't expect to get a valid high out to drive logic also.
 

andrewmm

Joined Feb 25, 2011
574
I didn't do a 74S04. I haven't seen a 74S04 in a circuit since early computers sat on floors. The data sheet quotes -1 mA high and 20 mA low for 74S04. A bit higher than 7404.
Yes, 7404 is old technology. Some had current limiting on the high side that allowed driving an LED using the current limiting as the load resistor for the LED, safely limited to 15 mA or so. No you couldn't expect to get a valid high out to drive logic also.
Yes, 74S is interesting,but I htink you were talking about the 7404,
Have quick look at the data sheet,
I ts a TI one, it has the 70S04, 74LS04 and the 7404 you mentioned

sorry I did not bother re naming the link,
 

Bernard

Joined Aug 7, 2008
5,649
Thanks hp for broaching the subject of overloading IC's. Look at how often the 4017 is high side driving LED @ up to 20 mA & surviving. Just minutes ago the question came up, " Why cant I drive the LED's from the 4071" or something close. Probably could get maybe 2 mA, but output would no longer be a ligit. logic level if also connecting to another IC input, circuit might have eratic or no operation.
 

Thread Starter

hp1729

Joined Nov 23, 2015
2,304
Thanks hp for broaching the subject of overloading IC's. Look at how often the 4017 is high side driving LED @ up to 20 mA & surviving. Just minutes ago the question came up, " Why cant I drive the LED's from the 4071" or something close. Probably could get maybe 2 mA, but output would no longer be a ligit. logic level if also connecting to another IC input, circuit might have eratic or no operation.
Just don't try to drive other logic from the same 4071 output.
Then there is the question of long term reliability. I have used LED driven from CMOS for test fixtures but I don't know if would use it for a commercial product. If my test fixture dies I can fix it easily.
 
Top