How can this quad NAND gate work??

[Chris Wilson]

SN7413 Positive quad input dual NAND gate in circuit attached. Pins 3, 5, 10, 11, 12 and 13 are floating, see image of PCB tracks. How can it work if one input is always zero (floating) please? Basically the board controls two big SCR's which adjust the mains input voltage level to a linear power supply's transformer's primary to reduce voltage drop across the pass transistors. It appears to work, but how....? Thanks.

Thanks.

[g0hjq]

Easy really ... unconnected inputs of 74 series TTL gates float up to logic '1' by design , not down to '0'.

I'm not sure that it's recommended practice due to noise immunity, but OK for non-critical applications with no track on the pin.

[skarecrow]

Easy really ... unconnected inputs of 74 series TTL gates float up to logic '1' by design , not down to '0'.

I'm not sure that it's recommended practice due to noise immunity, but OK for non-critical applications with no track on the pin.

Is that normally true? This is definitely not my area of expertise, but out of curiousity I decided to do some research before you answered. According to the TI datasheet (from 1983) I downloaded it says you need to use a pullup resistor.

Sent from my XT1565 using Tapatalk

[Chris Wilson]

Interesting, especially as this supply has the odd wobble where the input voltage to the pass transistor assembly goes to full secondary level, and with the 2N3055's that were used, one or more goes shorted. This seems only to happen using it to run a hefty 136kHz RF amp.... I was examining this area of the circuit, as best as my novice status allowed, and saw the lack of pin ID on the schematic and physically examined the SCR board, to find some are just floating. I wonder if the noise of this Class D amp (over 1kW very near the supply iteslf) is getting into this IC? Maybe a pull up resistor or two would be wise? Suggested value? 1K? 10K? Thanks!

[grumpydoc]

There's a good, clear explanation of how a TTL NAND gate works here http://www.learnabout-electronics.org/Digital/dig32.php and it is easy to see that with the inputs floating they behave the same as when a logic 1 is applied.

For bipolar TTL you have to work at pulling an input down to 0.

It is not considered good practice to leave inputs unconnected though - they should generally be tied high.

However for CMOS 74 series it is important that the inputs are tied either low or high because there's a fair chance that things will not work as expected if you leave them floating.

EDIT: 1k pullup for a TTL input is fine - you can connect several inputs to 5V via one resistor.

[Andy Watson]

Easy really ... unconnected inputs of 74 series TTL gates float up to logic '1' by design , not down to '0'.

I'm not sure that it's recommended practice due to noise immunity, but OK for non-critical applications with no track on the pin.

Is that normally true? .. the TI datasheet (from 1983) I downloaded it says you need to use a pullup resistor.

Yes, generally speaking. Although, as suggested in the other replies, it's a bit dubious to leave an input floating. If you look at the 7400 series datasheets the inputs are often depicted as open-emitters, multiple emiitter transistors in the case of and/nand gates. I.e. they are not active unless they are sourcing current. To be certain of pulling one of these input down (i.e. to 0) usually required sub-k-ohm impedances. Note we're talking about the original 7400 series. More modern variants (LS and whatever) can not be relied on to have the same behaviour. And CMOS should never be left floating.

[Cervisia]

Designing with TTL (https://www.fairchildsemi.com/application-notes/AN/AN-363.pdf) says:

TYING ALL UNUSED INPUTS TO A SOLID LOGIC LEVEL

Unused inputs on TTL devices float at threshold, anywhere
from 1.1V to 1.5V, depending upon the device and its family.
While this usually simulates a “high”, many application problems
can be traced to open inputs. Inputs floating at threshold
are very susceptible to induced noise (transmitted from
other lines) and can easily switch the state of the device. A
good design rule is to tie unused inputs to a solid logic level.
Inputs are usually tied to V_CC through a 1 kohm to 5 kohm resistor,
since tying them to ground means supplying the I_IL current
instead of the I_IH current. I_IL is several orders of magnitude
greater than I_IH. The resistor is recommended to protect the
input against V_CC voltage surges and to protect the system
against the possibility of the input shorting directly to ground.
A single 1k resistor can handle up to 10 inputs.

[Nusa]

Floating inputs are to be considered indeterminate, not high or low. Bad practice to leave them unconnected, even in those cases (like this one) where you can get away with it.

If you look at the datasheet for SN7413 (http://www.chipfind.net/datasheet/pdf/ti/sn7413.pdf), there's a schematic for the logic implementation. Note that there's a pull-up resistor on the input to the first transistor. Any one of the 4 inputs can pull it to a LOW level through the diode, which is how the AND function in implemented. It should be obvious a floating input can't pull anything LOW.

You can implement a X-input AND gate with X diodes and a resistor instead of an IC if you want. Ditto for OR gates by using them differently(reverse the diodes and pull-down instead of up). Add a transistor and another resistor to make it a NAND or NOR gate.

[Chris Wilson]

I will add a pull up to the floaters, one input has a long, un-terminated track right across the PCB to it, I suspect it could act as a rudimentary antenna, thanks for all the responses!

[skarecrow]

I will add a pull up to the floaters, one input has a long, un-terminated track right across the PCB to it, I suspect it could act as a rudimentary antenna, thanks for all the responses!

Keep us posted whether the resistors help or hurt in any way.

Sent from my XT1565 using Tapatalk

[Chris Wilson]

I will, although "hurt in any way" is a bit alarming..... I was going to use a single 1k resistor to the supply pin, fanning out to the floating pins. Mounted atop the IC which is socketed, soldered diect to the IC pins, as the PCB is not of modern quality and experience with them shows it doesn't take much for tracks to lift if soldered to. That OK? Spot of hot melt glue to stop things walking about, or maybe Araldite? Thanks.

[Nusa]

I will, although "hurt in any way" is a bit alarming..... I was going to use a single 1k resistor to the supply pin, fanning out to the floating pins. Mounted atop the IC which is socketed, soldered diect to the IC pins, as the PCB is not of modern quality and experience with them shows it doesn't take much for tracks to lift if soldered to. That OK? Spot of hot melt glue to stop things walking about, or maybe Araldite? Thanks.

My first question is why? It may be bad practice, but it doesn't mean the old existing product is actually broken. If it is, you're probably looking in the wrong place.

But if you're still going to, I'd recommend soldering to the IC pins over what you suggest. Almost all the connections you plan are going to pads that have no traces, so there are no associated tracks to lift. That means those pads themselves don't matter, even in the unlikely event that you destroy them. All that matters is that the wire connects to the socket pins.

[edavid]

I will add a pull up to the floaters, one input has a long, un-terminated track right across the PCB to it, I suspect it could act as a rudimentary antenna, thanks for all the responses!

There is no reason to change it, but if you really want to, it would be easier mechanically to tie the unconnected inputs to the connected inputs.