74LS14 testing circuit pls

[Ian.M]

74ACH..../74AHCT.... are excessively fast, which causes problems if you have long wires or don't have good grounding and decoupling - which is a typical problem when breadboarding.

74HC..../74HCT.... are a bit slower (similar to 'classic' or LS TTL)  and more forgiving.  They are far more breadboard friendly.   
 
Only get the 74xxCT...   TTL logic level compatible input version if you *NEED* a low threshold voltage, either for compatibility with bipolar TTL outputs, or 3.3V CMOS logic outputs etc.   Otherwise the 74xxC... families have more or less symmetrical input thresholds and significantly better noise immunity.   All 74xxC... families drive rail-to-rail (except open drain parts or if excessively loaded), but for many, they are still better at sinking than sourcing, so higher current LEDs are best wired active low (between the output and Vcc, with a series resistor).

[KrudyZ]

Ok, here is the working one gate of one of the two ic's.


If I try all the other input to feed with the same signal and setup like this working one
nothing is happening or the outputs are only HI or LOW.

I assume my IC's are bad and I have one IC with only one working gate.
This is what I'm think and I cant get believe this because the IC's are brand new!

The rise and fall times of that purple trace are VERY slow.
Look at the time base of the scope...

[Chriss]

Yes. The purple trace is the output from the photo interrupter which is connected to the input of the gate.


Sent from my GT-I8260 using Tapatalk

[Ian.M]

So you've only got about 3V input amplitude, which means that if you are running the logic from 5V, you *NEED* real bipolar TTL or LS TTL or 74HCT not 74HC logic.

It may be worth using a comparator (e.g LM339), with resistors to set the threshold and hysterisis.  As that comparator has an open collector output, don't forget the output pullup resistors.

[ebastler]

Yes. The purple trace is the output from the photo interrupter which is connected to the input of the gate.

Huh? I thought you had a pull-down resistor and a switch to +5V connected to the input?! 
(Your schematic in reply #5?)

It's hard to do remote troubleshooting when we don't have complete and consistent information about the boundary conditions. Your photo interrupter's output might just be too weak for the LS14 input.

[Chriss]

In the first version I had that with the switch, after I found that good gate I connected also my photo interrupter to the
74LS14 good gate and made the measurements...

I know it is hard to remote troubleshooting...

I was thinking also about the LM339 to use, but I found more simple to use a 74LS14 in my circuit than the LM339.

I tested in my circuit that one IC 74LS14 which have a good gate and it is doing the job well.
I also connected it to my uC and started to count the triggered signals and works great.

I don't know, maybe I will give a try to another IC but now from the 74HCT version and something not from the year '85. 

I just was wondering about my situation with this IC's because I was playing before also whit such of IC's and I never had a burnt one
or situations like now, except if I repaired some electronics with a bad Schmitt trigger...

So, back to the situation, I assume more or less my schematic shouldn't kill the IC's.
Maybe that version where I connected the LED to the IC.
But I have also a bad IC on which was the LED not connected.

I will try to get a new ic but now one from the 74HCT version and make a new test..

I don't know, maybe I messed up something.

Just one more question:
If I don't connect the unused pins to gnd, can that kill the IC?

[rstofer]

Some of MY rules for logic (regardless of family)

• Apply Vcc and Gnd to the proper pins with short leads from the PS
• Add 0.1 ufd capacitor between Vcc and Gnd at EVERY chip and keep the leads as short as possible.  There are sockets with built-in capacitors - that is short!
• Always pull inputs high and switch them low.  The size of the pull-up resistor depends on the logic high threshold and the pin input current.  If I don't care about power consumption, 1k is the right answer.
• Never pull a load high, always pull it to ground.  This is important for TTL and less so for CMOS but I do it that way anyway.

[rstofer]

Just one more question:
If I don't connect the unused pins to gnd, can that kill the IC?

It can be for CMOS.

You need to treat the inputs and outputs properly.  You can pull inputs to a solid ground but I would never pull them to a solid Vcc.  I ALWAYS use a resistor to limit the logic high current (others will have a different opinion).  ALL inputs on ALL devices need to have some kind of connection.  Up or down, your choice but they can not be allowed to float.

Outputs:  Never pull an LED up, always pull it down, read the datasheet for maximum output current - both high and low.  Connect the resistor between Vcc and the LED anode, connect the LED cathode to the output pin.

ETA:  You can tie multiple inputs to the same pull-up resistor but you have to make sure the resulting pin voltage is higher than the logic 1 threshold with all inputs sinking their maximum logic high input current.

[TomS_]

74ACH..../74AHCT.... are excessively fast, which causes problems if you have long wires or don't have good grounding and decoupling

Sorry for the hijack, but this is burning a hole in my understanding - which was that the AHC/AHCT versions "simply have a lower propagation delay" (thats the way TI marketing material positioned it anyway). If the HC/HCT versions are a bit slower propagation wise, how does an input transitioning an output a bit quicker cause such significant issues? Or perhaps I have missed/do not understand another detail... 

Re using xHCT parts, I have been looking to stock these specifically because so many microcontrollers and other logic devices seem to be 3V3 parts these days, particularly at the mid-high end, and the xHCT parts are more compatible due to lower logic high requirement. So its more or less about stocking parts which can be as universal as possible.

Thanks!

[Ian.M]

Its all down to risetimes and load capacitance.   The faster the risetime, the bigger the current spike drawn from Vcc or Gnd to charge or discharge the load capacitance when an output switches and the greater the resulting threshold shift experienced by other gates on the same chip caused by the disturbance at its supply pins.   Then its a matter of noise levels on the inputs and the probability that the threshold shift will momentarily erode the logic input threshold margins enough to cause a glitch.  Its a square law effect - the current spike amplitude increases in proportion to the speed of the transition, but its duration also reduces proportionately, so a 20% reduction in risetime makes it 44% harder to decouple the supplies adequately and provide a reliable ground (stiff enough at DC and low enough impedance up to an equivalent frequency to double the rise time)

With very fast logic you get into the situation that you need to source terminate longer lines to control the risetime to tame the spikes.

[Zero999]

So you've only got about 3V input amplitude, which means that if you are running the logic from 5V, you *NEED* real bipolar TTL or LS TTL or 74HCT not 74HC logic.

One way round that is to add a diode to the pull-up resistor. It's a bit of a crude hack and only works at fairly low frequencies.

[Ian.M]

The other dirty trick is to lift the ground of *ALL* the 3.3V logic by one diode drop, which can work reasonably well with faster signals.  However it barely makes the typical 70% guaranteed logic '1' threshold, so isn't great for noise immunity.   Also, beware of 5V logic with 80%/20% thresholds, as even with a diode drop boosting it, the 3.3V logic wont reliably cross the 80% threshold unless you overvolt the 3.3V logic and/or undervolt the 5V logic.

[joeqsmith]

Hi!
Can somebody share with me a simplest testing circuit for a 74LS14 Schmitt Trigger pls.

I have some strange situation with this ic's and I cant figure it out.

I give power to the ic as in the datasheet described.
I put a LED through a 330Ohm resistor to the Y1 and
set a +5v to A1 and the led is all the time on.
No matter if I set a +5v or GND to the A1.
Pin. Also to other gates too..

So, does I mess something or are my ic's bad?
Some of them with the same setup didn't lite up the LED.
I just swap the ic in the same circuit...

Picture of your setup may have helped.   Could be wired wrong.   Some time ago, I played with a bunch of old DIP inverters.  I started out trying to replicate someones setup using some whiteboard. 

https://www.eevblog.com/forum/projects/waveforms-in-a-74ls04-ring-oscillator/

[Chriss]

Ok.
I have some questions:
1. Is my ic fake or not?
2. Is it worth to buy the HCT version and use that new technology for upcoming projects or can I use also the old 74LS14 series?

I digged through tons of data sheets, resources on the net but found millions of variations of circuit made with the 74ls14 which are actually crossing between the theory and real world hardware circuits.

So I found also a blog where somebodi I think from Porugal wrote about a light detector circuit where actually my circuit was used. I mean the guy who wrote the text on that blog didn't copy - paste my circuit but it is the same meaning how they made his electronic light switch.
If my circuit could kill the ic an/or are wrong then millions of other circuits are still wrong in the world, what are not impossible.

No one datasheet show any so to say "tipical" schematic how to use the ic.
Can we figure out a tipical wireup for this ic what could work in real world?

What a kind of crappie ic is this when it can't handle a LED diode.
Because of 20mA max current they will be killed....
Than it is a "must have" to put a transistor to the output. Without the teansistor a such of ic can be really easy killed.

Alk these is to shitty.

At the end I come to a mind not to use this crappie ic's and interchange them wit an LM339 and finishing the job.

So, what is your opinion about a circuit how to use this crap inverter to switch the input from a photo interrupter and the output should send the signal to a uC.



Sent from my GT-I8260 using Tapatalk

[ebastler]

Ok.
I have some questions:
1. Is my ic fake or not?
2. Is it worth to buy the HCT version and use that new technology for upcoming projects or can I use also the old 74LS14 series?

Ad 1.
It seems unlikely that it is a fake, but given its age and not-so-mainstream origin and package, maybe it has not aged well and has degraded over time.

Have you tested it with (a) clean power supply, decoupled with suitabe 100nF capacitor, (b) strong, well-defined input test signal from pullup resistor and switch to GND, and (c) checking the output on an oscilloscope, without a load connected? It is not clear to me from the thread if you have done tests where all three conditions were met. If the anser is "yes", ad the chip still did not behave, then it is probably bad.

Ad 2.
A key benefit of an HCT chip would be that you can get one from current production, ruling out the possible degradation of your Hungarian LS chip. Also, the higher input impedance of the HCT may be helpful if you feed it a wimpy signal from a photosensor. So yes, I would give it a try.

[rstofer]

There is nothing wrong with 74LSxx devices.  It is up to you to read and understand the capabilities and limitations as given in the datasheet.  Every single number printed in the datasheet means something.  Sometimes it is a factor in your design, sometimes it isn't.

All TTL logic is asymmetric in terms of pull-up and pull-down current.  That's why loads are always pulled down.  This isn't true for CMOS...

For the 74LS family, I_OL (pull down capability) is 8 mA - hardly enough to light an LED but it will work.  OTOH, I_OH (pull up capability) is only 400 uA and is incapable of driving an LED. 

If you look at I_OL and I_IL you will see where the device output can drive 20 inputs.  Same for I_OH driving I_IH. I wouldn't push the 20 gate fanout but I would certainly do 10 or 15 gates.  The point is, the 74LSxx devices are perfectly capable of doing performing their intended function in their intended environment.  It's up to you to design around the capabilities and limitations.

http://ee-classes.usc.edu/ee459/library/datasheets/DM74LS14.pdf

[Chriss]

ebastler:

Have you tested it with (a) clean power supply, decoupled with suitabe 100nF capacitor, (b) strong, well-defined input test signal from pullup resistor and switch to GND, and (c) checking the output on an oscilloscope, without a load connected? It is not clear to me from the thread if you have done tests where all three conditions were met. If the anser is "yes", ad the chip still did not behave, then it is probably bad.

Yes, I done these tests several times gate by gate and only one gate had a good signal of the two IC's.
That's made me crazy because I bought 2 IC's and I tried them and they are looking bad.

I tested before too in the past this type of IC's with my scope without loads and all they was ok, with a nice inverted signals,
but these two ic's didnt do anything. Or non stop low, or non-stop high, or non-stop nothing. Just one gate is acting well.

Here is two circuits what I was used also before for testing and button debouncing.
When I say "before" I mean years back...



Is the 1. circuit good enough to just test the gates on this 74LS14?
I ask because I done this type of connections in my projects and they still working even after 7 years.

Also the 2. circuit, is that good enough because that is also working years back till now...
This circuit I used for button hardware debouncing for uC's...

When I tested the 74LS14 in the past, I never pulled up the not used inputs, and I never faced this kind of problem like now.
I should maybe avoid this type of wiring, to let the inputs floating...

[Ian.M]

For the second circuit, it works with a switch, but if you attempt to drive it from a transistor, unless you reduce the 1K resistor, you may not meet the guaranteed logic '0' threshold.   

e.g. the OnSemi 74LS14 datasheet says it can take up to 0.4mA to pull the input down.  Therefore there can be up to 0.4V across the 1K resistor and to get the input under the 0.6V guaranteed logic '0' threshold (parameter: 'Negative-Going Threshold Voltage') the left end of the resistor has to be below 0.2V.  However the typical output '0' voltage is 0.25V so you are already S.O.L. if you want to reliably drive a low pass filter like that from another gate on the same chip.   Keep that resistor under 500R and you'll get away with it.

RC debouncing for 74HC/HCT is much easier as their input current is negligible so you can use much lower value capacitors and higher series resistors.

[ebastler]

Chriss - when you say „this type of IC“ has worked for you in the past, do you mean the exact batch of Hungarian ICs you are using now, or some orher brand? If you were successfully using something else before, it seems pretty clear that the Hungarian chips are bad - toss them!

[amyk]

Not fake, but certainly very old and possibly defective.

Does the bottom look similar to this?
http://www.silirium.ru/pictures/ussr-Electronpribor-KR531IP3-brown-back.jpg

A quick Google suggests these parts were found in Commodore C64s

http://cbm-hackers.2304266.n4.nabble.com/CBM-s-CP-M-Z-80-carts-for-the-64-td4656117.html

I have seen lots Commodore systems that used
EL7406, EL74LS14 and others. Those have no makers logo, come in a brown
casing instead of the usual black and are, as far as I know, russian in
origin.

https://www.forum64.de/index.php?thread/45601-west-trifft-ost-russ-ramchips-im-c64/

Not sure what the EL manufacturer prefix is either, but Elektronpribor is a good guess.

[Chriss]

Ian.M:
Thanks for this nice explanation, I really appreciate it.
I will definitively switch to the 74HC/HCT series IC's.

ebastler:
When I say „this type of IC“ has worked for me in the past, I mean this hexagon Schmitt trigger IC's but from other manufacturer.
I never faced with this brow casing IC.
I have experience with 74xx series IC's but from other manufacturer.

amyk:
No, the bottom didn't looks like on the link.
The  bottom is simple brown plastic like the upper side.

[Chriss]

Ok, today I got new IC's but now SN74S14N in black package...

Please give me suggestion how to correct breadboard this IC.
In the past I used the 1. circuit from my schematic above.
I read somewhere on the net does all the not used input pins should be pull-up or down.
I never did that before and I never killed any IC this type of Schmitt trigger.

But now I'm a bit afraid and wish to change things if I need to make my circuits better.

What would be your simplest setup for a quick test on breadboard the SN74SL14N?

Thank you.

[iMo]

Based on the package and type of font used it is an old Soviet chip (most probably an export marking)..
The former eastern countries produced 74xxx series as well, they used black packages (with different form and fonts used), however.

This is the simplest test you may try:

http://www.learningaboutelectronics.com/Articles/7414-schmitt-trigger-oscillator-circuit.php

You may use any resistor 1k-100k and capacitor 10nF-10uF with your test. You may wire the inputs of the other free inverters to the oscillator's output and observe their outputs - you have to see pulses as well..

[Ian.M]

Are you sure about that part number? Google doesn't return *ANY* datasheets for  SN74S14N or even 74S14.  If that's what's on them, odds are they are fakes.   IMHO its more likely you've got SN74LS14N chips.

Your circuit n.o. 1 or my circuit from reply #21 will do to test them, one gate at a time.   Don't leave them running for a long time on my circuit - it pushes them too close to their limits.

Bipolar TTL (all types) inputs float high well enough that you can leave unused gates' inputs floating when breadboarding.  However the recommendation for the final build is to pull up to Vcc via a 1K resistor.  You can often group several unused inputs to one resistor.   Strictly speaking its only original 7400 series TTL that you *had* to use a pullup resistor for to avoid latchup if the input experienced a spike on the Vcc rail.   One generally avoids grounding unused inputs unless its required for the logic function, as that consumes extra current.  Don't leave spare inputs of used gates floating - that can result in big trouble with glitches when another pin on the same chip changes state.

CMOS is *MUCH* easier - just tie the input direct to either rail AT THE CHIP - it wont waste current and it wont latchup.

[PA0PBZ]

In the same post he calls it a SN74SL14N so my bet is a keyboard problem.