| tonitze |
Is it advisable to ground all unused input pins of the 74ac02?
Please advice.
Toni. |
|
|
| macboy |
| Yes, but keep in mind that with both inputs grounded, the output will be high. Ground that, and you are asking for trouble. Generally, you can ground all unused inputs; leave unused outputs unconnected. Some logic types may have pull-up resistors on the inputs (or even the outputs) which is a consideration when determining whether you want to pull-up or pull-down the unused inputs. With AC series though, you should be fine either way. |
|
|
| Jocko Homo |
You would be better off avoiding AC logic. HC logic should be good enough; if not try VHC. Most anything but AC.
Jocko |
|
|
| tonitze |
Thanks Jocko and Macboy. I will try out the VHC on my next clock circuit.
Toni. |
|
|
| Guido Tent |
| quote: | Originally posted by tonitze
Thanks Jocko and Macboy. I will try out the VHC on my next clock circuit.
Toni. |
Hi
Firts follow Jocko's advise. Then connect all unused inputs to Vcc and ground all unused outputs. After you've done that figure out why.
cheers,
- |
|
|
| seoman |
Ground all outputs :xeye:
Ok I dig the part of connecting to VCC.
Aspecialy the old TTL ports use less power that way.
But i would never ever connect an output to a power line.
That's asking for smoke!
PLease explain
Simon |
|
|
| Guido Tent |
| quote: | Originally posted by seoman
Ground all outputs :xeye:
Ok I dig the part of connecting to VCC.
Aspecialy the old TTL ports use less power that way.
But i would never ever connect an output to a power line.
That's asking for smoke!
PLease explain
Simon |
Hi
The goal is lower inductance and damping in the ground circuit (off chip - on chip).
Why the smoke ?
cheers |
|
|
| seoman |
Well
If you make a mistake the chip will draw a lot of current from the supply.
Ok some people won't make a mistake but......
People reading this posts could think they always should do this trick.
And thats tricky!
Regards Simon |
|
|
| tonitze |
Today just follow Guido's advice by connecting unused inputs to vcc and outputs to gnd. No smoke at all and the 74hcu04 runs at normal temperature. Improvement, wasn't sure but somehow sound smoother. Lower noise I guess.
Toni. |
|
|
| BrianL |
Guido, et al., I agree with seoman
I really question the suggestion to ground the unused
outputs.
It's always good practice to pull up all unused
inputs (with resistor, such as 1K unless the input is in
a gate you're using which might require pulling low)
If the unused output is naturally going low as a result
of your input setting, you still have the potential for
pulling current through the output and I don't see
that as a good or safe thing. I can't see how
connecting the output to ground will have any
salutary effect as the only think it can do is
cause more currents (hopefully DC) to circulate.
I also agree with Jocko to avoid AC logic (and "FACT' as well) |
|
|
| tonitze |
| Any specific reason why we should avoid the AC family? |
|
|
| Guido Tent |
| quote: | Originally posted by seoman
Well
If you make a mistake the chip will draw a lot of current from the supply.
Regards Simon |
Many mistakes are accompanied by smoke...... |
|
|
| Guido Tent |
| quote: | Originally posted by BrianL
Guido, et al., I agree with seoman
I really question the suggestion to ground the unused
outputs.
It's always good practice to pull up all unused
inputs (with resistor, such as 1K unless the input is in
a gate you're using which might require pulling low)
If the unused output is naturally going low as a result
of your input setting, you still have the potential for
pulling current through the output and I don't see
that as a good or safe thing. I can't see how
connecting the output to ground will have any
salutary effect as the only think it can do is
cause more currents (hopefully DC) to circulate.
I also agree with Jocko to avoid AC logic (and "FACT' as well) |
Hi
I suggest you draw out the equivalent circuit that arises when you follow my suggestion. Include buffer output R and leadframe inductances.
Ofcourse no DC currents will circulate when you connect a low output signal to ground.
The reason I propose this is to lower the inductance between internl reference and PCB ground, and to add damping, to reduce induced noise in your in and outputs.
best regards |
|
|
| gmarsh |
| quote: | Originally posted by Guido Tent
Ofcourse no DC currents will circulate when you connect a low output signal to ground. |
Which negates the whole purpose of connecting low outputs to ground? :)
Depending on logic family, you'll probably still have some quiescent current - the output complimentary pair is often biased to make it run a bit faster. I know F-series TTL does this, i'm not sure about different CMOS families.
Also during turn-on/turn-off when the supply rail is out of limits, the output of the logic might be momentarily high. I've personally encountered this... now suppose the circuit it's in has 10,000uF of decoupling on the logic supply and it's unplugged from the wall, that 10,000uF at 2 volts or so might end up almost completely being dumped into the output transistors of a part.
I wouldn't do it.
As for AC logic... if it was a bad idea to use it, they wouldn't make it... I've used it without any problems, but keep in mind that it's noisy - the output slew rate is fast so series output termination is almost always needed, and it will smack its power supply all over the place unless you put excellent decoupling on it. |
|
|
| Guido Tent |
| quote: | Originally posted by gmarsh
Which negates the whole purpose of connecting low outputs to ground? :)
Depending on logic family, you'll probably still have some quiescent current - the output complimentary pair is often biased to make it run a bit faster. I know F-series TTL does this, i'm not sure about different CMOS families.
Also during turn-on/turn-off when the supply rail is out of limits, the output of the logic might be momentarily high. I've personally encountered this... now suppose the circuit it's in has 10,000uF of decoupling on the logic supply and it's unplugged from the wall, that 10,000uF at 2 volts or so might end up almost completely being dumped into the output transistors of a part.
I wouldn't do it.
As for AC logic... if it was a bad idea to use it, they wouldn't make it... I've used it without any problems, but keep in mind that it's noisy - the output slew rate is fast so series output termination is almost always needed, and it will smack its power supply all over the place unless you put excellent decoupling on it. |
Hi
What's the fuzz on DC ? When it comes to digital audio, only the switching characteristics are important.
You may want to ignore that, it doesn't bother me but you'll miss the point when it comes down to jitter......
regards |
|
|
| gmarsh |
| quote: | Originally posted by Guido Tent
Hi
What's the fuzz on DC ? When it comes to digital audio, only the switching characteristics are important.
You may want to ignore that, it doesn't bother me but you'll miss the point when it comes down to jitter......
regards |
I understand your point - I've seen programmable logic which has 'user programmable grounds' which perform the same function. When you have huge numbers of simultaneous switching outputs on a CPLD/FPGA, ground bounce becomes an issue.
But these chips are designed for such things, 74-series logic is not. So give it a try - but don't be surprised if your chip fails. That's my point. |
|
|
| Guido Tent |
| quote: | Originally posted by gmarsh
I understand your point - I've seen programmable logic which has 'user programmable grounds' which perform the same function. When you have huge numbers of simultaneous switching outputs on a CPLD/FPGA, ground bounce becomes an issue.
But these chips are designed for such things, 74-series logic is not. So give it a try - but don't be surprised if your chip fails. That's my point. |
Hi,
Thanks for feedback
A single inverter is enough to create significant groundbounce. The crosscurrent may peak to 10mA peak in ns, multiply that by a few nH and you're in tens of mV (at the most important part of te transition)
I wouldn't suggest it if I hadn't tried it.......
cheers |
|
|
| Jocko Homo |
I would not advise quiting my day job and starting an audio company because I know many who have tried. The names have been changed to protect the innocent.
Jocko |
|
|
| gmarsh |
| quote: | Originally posted by Guido Tent
Hi,
Thanks for feedback
A single inverter is enough to create significant groundbounce. The crosscurrent may peak to 10mA peak in ns, multiply that by a few nH and you're in tens of mV (at the most important part of te transition)
I wouldn't suggest it if I hadn't tried it.......
cheers |
I'm tempted to verify how much of an effect this makes. I'll hook up a 74HC04 chip, drive a fast-edged square wave into a single inverter and tie the rest of the inputs to +5V. I'll measure one of the "low" outputs with a scope, and I'll start tying the rest of the low outputs to ground and look for a difference.
Now, If ground bounce is a big concern in your designs, I hope you're using "just fast enough" logic in TSSOP packages decoupled with an 0603 SMT ceramic on each side. :D |
|
|
| Guido Tent |
| quote: | Originally posted by gmarsh
I'm tempted to verify how much of an effect this makes. I'll hook up a 74HC04 chip, drive a fast-edged square wave into a single inverter and tie the rest of the inputs to +5V. I'll measure one of the "low" outputs with a scope, and I'll start tying the rest of the low outputs to ground and look for a difference.
Now, If ground bounce is a big concern in your designs, I hope you're using "just fast enough" logic in TSSOP packages decoupled with an 0603 SMT ceramic on each side. :D |
Hi
The groundbounce depends on Ldi/dt. The better the external decoupling, the higher di/dt may develop......
cheers |
|
|
|
