74LS90 power-on auto reset

[Torben]

Hi all,

OK, hopefully this is the last question I'll need answered on my TTL foosball scoring circuit. Thanks to those who've helped so far!

Last thing to do is to have it reset to 0 automatically when first powered on. I *also* need it to have a manual reset switch.

For various reasons (mostly room, actually--and I want to get this thing done tonight if I can, for another), I don't want to go with an IC-based Schmitt trigger, which I've read here is the preferred solution. I've tried a couple of basic RC nets on the reset pins but can't get the thing to count with it attached (i.e. it keeps the reset pins off ground). I tried sticking a 2n3904's emitter to ground, collector to the reset pins, and base to a RC net designed to delay the base input by ~20mS. This configuration let me count but wouldn't reset the circuit on power-on.

Seems like there *must* be a way to do this, given the number of references to doing it with only an RC net I've found. But I'm just a hobbyist with no calculus and a very, very basic understanding of what's actually going on in the RC net, so I am not sure what I need to change to get it working.

I'm going about it fairly emperically right now (just researching/googling, trying different things, failing, reading more, trying again....) but I'm hoping somebody has a) the answer and b) the urge to write up a quick answer/explanation/pointer to solid information on this. Seems like a pretty basic thing, but I just haven't learned it yet.


Thanks (and sorry for the wordiness),

Torben

 

[kchriste]

Seems like there *must* be a way to do this, given the number of references to doing it with only an RC net I've found.

If the reset pins on the LS90 were active LOW, this would be true but they are active high. I think your transistor method will work if you use a bigger cap. I assume you have something like this:

 

[audioguru]

Cmos logic ICs are SO much easier to work with than power-hungry old TTL.

 

[Torben]

If the reset pins on the LS90 were active LOW, this would be true but they are active high. I think your transistor method will work if you use a bigger cap. I assume you have something like this:

I do indeed have something like that, except I missed a resistor. And my cap value was way lower. This does the trick.

I do have a couple more problems but I want to keep working them before asking about it.

Thanks a lot for the help!


Torben

 

[Hero999]

Cmos logic ICs are SO much easier to work with than power-hungry old TTL.

But it's LS TTL which is so much more efficient and no TTL is easier to work with because it's far more resistant to ESD. The slightest overvoltage on the gate of a MOSFET in a CMOS chip is enough to zap it but a bjt just goes into avalanche breakdown and as long as the current isn't too high for too long no harm will be done.

I've had a quick glance at the datasheet and MR1 and MR2 both need to be high to reset the counter so connect them both to +V via a 1nF capacitor and to 0V via a 100k resistor.

 

[audioguru]

Hi Hero,
I have used thousands of CD4xxx Cmos logic circuits without damaging a single one with ESD, even though there is plenty of static during cold winters in Canada. I don't use a ground strap, but I am a little careful.

The inputs have a series resistor and protection diodes to both supplies.

 

[audioguru]

I've had a quick glance at the datasheet and MR1 and MR2 both need to be high to reset the counter so connect them both to +V via a 1nF capacitor and to 0V via a 100k resistor.

It is power-hungry TTL (LS-TTL). Each input sources a max current of 0.4mA and you want the low voltage to be 0.4V so a resistor to ground for each input should be only 0.4V/0.4mA= 1k ohms max.Then use a huge reset capacitor.

To power-up-reset a Cmos counter you could use a 10M resistor if you want and a tiny capacitor.

 

[Hero999]

Of course CMOS is more effiecient than LS TTL at DC (I wasn't sugesting otherwise). However CMOS is only more effiecient than LS TTL at low frequencies, at higher frequencies the power consumption of CMOS exceeds LS TLL considerably (I realise that you probably weren't suggesting otherwise).

I still refute the fact that CMOS is easier to use, apart from ESD, CMOS prone to latchup and TTL isn't.

Here's a nice little experiment.

Build a simple NOT gate oscillator, using both TTL and CMOS gates.

Now try the following:

• Take the input on an unused gate, above and below the power supply by 0.6V.
• Zap it using a piezo igniter from a cigarette lighter.

90% of the time TTL will survive this abuse while the CMOS version will self-destruct almost 100% of the time.

When I was a nube I built a TTL astable multivibrator and ran it off an unregulated 4.5V mains adaptor, it survived even though I later measured the voltage to be 7V.

 

[audioguru]

I wouldn't zap a circuit with a piezo lighter, but I have operated a 74HCxx Schmitt trigger oscillator with only a 1.25V supply, and a CD4xxx oscillator with a 20V supply (Texas Instruments Cmos has a 22V max supply rating).

 

[kchriste]

Thanks a lot for the help!

No problem...

I do have a couple more problems but I want to keep working them before asking about it.

That's the spirit!

When I was a nube I built a TTL astable multivibrator and ran it off an unregulated 4.5V mains adaptor, it survived even though I later measured the voltage to be 7V.

The spec for the 74Lx series power supply is a max of 7V, so you were on the hair of the hairy edge. As a young noob I ran a TTL LED chaser (Cylon eye for a Halloween costume) from 4 AA batteries with no problem. They are great chips for learning about gates and power supply bypassing.

Zap it using a piezo igniter from a cigarette lighter.

Is chip torture covered under the Geneva convention?

 

[Torben]

No problem...

That's the spirit!

Thanks! And it appears that the main problem I was still having was just related to the score and win triggers (more 2n3904s), which seem to have been pulsing the power supply enough to trigger the power-on reset circuit...but only when the lines meant to run to the indicator lamps were attached. Meaning that if you scored, the thing would instantly reset unless you had the lamps unplugged. Not nice.

Bypass caps on the score and win circuits seems to have taken care of it, though. I'm still poking around to try to figure out if I can explain the exact mechanism of the problem, but at least it's solved--although I don't trust any solution I cannot explain .

The spec for the 74Lx series power supply is a max of 7V, so you were on the hair of the hairy edge. As a young noob I ran a TTL LED chaser (Cylon eye for a Halloween costume) from 4 AA batteries with no problem. They are great chips for learning about gates and power supply bypassing.

Is chip torture covered under the Geneva convention?

I think you were replying to Hero here. *I'm* still a noob (though no longer so young).


Thanks again for the help!

Torben

 

[ljcox]

The power on reset circuit is very crude. I suspect the reason it was triggered spasmodically is due to the electroylic cap. These have a relatively high ESR and high impedence at high frequencies.

I would have used a Schmitt Trigger, eg. a 74LS14. Connect a resistor and a capacitor (say 100 nF) to the input. The top end of the res goes to +5 Volt and the lower end of the cap goes to gnd.

There would need to be a second one to invert the signal if you want an active low reset signal.

 

[Torben]

The power on reset circuit is very crude. I suspect the reason it was triggered spasmodically is due to the electroylic cap. These have a relatively high ESR and high impedence at high frequencies.

Cool--I'll try some different kinds in there. What I have here are electrolytic, mylar, and ceramic. And a couple I can't identify just yet.

I would have used a Schmitt Trigger, eg. a 74LS14. Connect a resistor and a capacitor (say 100 nF) to the input. The top end of the res goes to +5 Volt and the lower end of the cap goes to gnd.

There would need to be a second one to invert the signal if you want an active low reset signal.

I've found the Schmitt trigger solution many times as it's obviously the preferred solution--and while I earlier stated that I didn't want to go that route for various reasons, I'm now thinking of just redesigning my "body board" to include the IC. (I think of the board with the counters and input debouncer/timers as the brain board, and the board with the power supplies and score/win triggers as the body board.)

I'm not sure if you're saying to connect the Schmitt trigger and then use the RC net you describe, or to just use the net right on the 74LS90 reset pins.

Anyway, the first part of my project went faster because I designed everything in Eagle for the brain and display boards. I'm finding weirder stuff with the power/output board, I suspect, largely because I jumped right in on the perfboard and started soldering.

So I think tonight my project is to take my little body board, capture it in Eagle, and go carefully over the resulting schematic to see what I really have here. The whole is working...erratically. So I've probably got something riding right on the edge of its spec, squishy power, or parasitic capacitance somewhere (each one of those is something I've already run into with this project, which is why I suspect them again).

Thanks for the hints!


Torben

 

[audioguru]

TTL eats power. Even LS TTL. So a 0.1uF ceramic disc capacitor is a must across the power pins of each IC to stop the circuit from behaving erratically.
Make the 0V wire pretty big on the circuit board.

 

[Torben]

TTL eats power. Even LS TTL. So a 0.1uF ceramic disc capacitor is a must across the power pins of each IC to stop the circuit from behaving erratically.
Make the 0V wire pretty big on the circuit board.

I've got 0.1uF ceramics across the power pins on the 74LS90s, but not on the 74LS47s on the display boards. (Actually I do have it on one display board, but not the other.)

I'll try adding one to the other board and see what happens.


Thanks!

Torben

 

[ljcox]

I've found the Schmitt trigger solution many times as it's obviously the preferred solution--and while I earlier stated that I didn't want to go that route for various reasons

Sorry, I did not notice your earlier comment.

(I think of the board with the counters and input debouncer/timers as ....

You may be able to use one of the input debouncers or timers for the power on reset (if there is a spare section in an IC, eg. one Schmitt spare in a Hex Schmitt or if you are using a 555, use a 556 and use the spare section for the POR.

I'm not sure if you're saying to connect the Schmitt trigger and then use the RC net you describe, or to just use the net right on the 74LS90 reset pins.

You could connect it directly to the 74LS90 reset pins if you wish, I prefer the Schmitt solution as it gives a "clean" pulse to the 74LS90 reset.

 

[ljcox]

TTL eats power. Even LS TTL.

Yes, I gave TTL and LS away as a bad joke years ago. I prefer the 4000 series CMOS and the 74HC series.

 

[audioguru]

In the '70s I used TTL LS logic ICs. Then I tried Cmos and never used another power hungry TTL again.