74HCT74 NXP vs Texas

[Jan Audio]

Hi, merry cristmas.
I have another problem :

When i change my 74HCT74 from NXP to Texas instruments it dont works anymore.

I have a square wave to 8K2 resistor going into a 2n2 capacitor,
the capacitor out goes with a 10K and diode to ground and to the CLK input of the 74HCT74.

Why is it different ?
thanks

btw : it does work somehow only it toggles twice really fast so the output looks like always on state.

[rhb]

Perhaps you should post a schematic?  That way readers would know what you're trying to do.

[DaJMasta]

Is the TI part also an HCT series chip?  Some different families require different voltages or have different response characteristics (or are open collector outputs, like the F series).

Hard to tell exactly from your description, but is your clock input signal AC couple but not biased to stay above negative voltages (can cause erratic behavior)?  Also could be that the TI part is configured to use any old edge trigger whereas the NXP part is only rising or falling edge - giving you 2x the switching rate - the datasheet should be able to tell you if that's the case, though usually these are different part numbers on the package (the 74 model should be positive edge triggered only).

[Jan Audio]

Yes the Texas is the SN74HCT74N.

[podericof]

Hi there,
can you also tell us the complete PN for the 2 ICs?

[Jan Audio]

Hi, sorry what is a PN ?

Texas is : SN74HCT74N
NXP : 74HCT74, i have to look.

[Jan Audio]

The NXP is : 74HCT74N

[wraper]

If it does not work and IC is not faulty, there is something wrong with your circuit. Just happens to be that one of them marginally operates.

[Jan Audio]

Thanks, i will re-check my circuit again, hope the Texas works since i have 100 piece.

[wraper]

I have a square wave to 8K2 resistor going into a 2n2 capacitor,
the capacitor out goes with a 10K and diode to ground and to the CLK input of the 74HCT74.

You shall not input signals with slow rise/fall time into CMOS ICs to begin with. The only exception are ICs with schmitt trigger inputs. You have not described diode orientation as well. As people already said, post the schematic.

[Zero999]

What are you trying to achieve the the resistors and capacitor.

[Bassman59]

Thanks, i will re-check my circuit again, hope the Texas works since i have 100 piece.

Post a schematic.

[David Hess]

For the same part from different manufacturers, some clock inputs are implemented as schmitt triggers and some are not but this is not always reflected accurately in the datasheets.  This may be one of the later cases and it is exactly what I would expect of the 74HCT74.

The only way to be sure is to either qualify the parts or include your own schmitt trigger before the clock input if you are going to use different manufacturers.

[Jan Audio]

Here was the NXP schematic :

                                              2N2
[ 12v waveform ]-----[8K2]-----||------+----[10K]---ground;
                                                          +----|<4148]---ground;
                                                          +----[CLK]

I fixed it with this for Texas :

                                              2N2
[ 12v waveform ]-----------------||------+----[1K]---ground;
                                                          +----|<4148]---ground;
                                                          +----[CLK]

Maybe it is a bit short circuit like this ?

You shall not input signals with slow rise/fall time into CMOS ICs to begin with. The only exception are ICs with schmitt trigger inputs.

Ok, thank you i did not know.

[wraper]

[ 12v waveform ]-----[8K2]-----||------+----[10K]---ground;
                                                          +----|<4148]---ground;
                                                          +----[CLK]

I fixed it with this for Texas :

                                              2N2
[ 12v waveform ]-----------------||------+----[1K]---ground;
                                                          +----|<4148]---ground;
                                                          +----[CLK]

Some piece of nonsense you made there. If you need to reduce voltage from 12V, why not use resistive divider of resistor + zener diode? And why don't you draw normal circuit?

[Jan Audio]

I have no zeners in house, another parts series to buy ?, do i really need them...
If the NXP was still available i would buy them.

I dont have software to render schematics, got any free ones ?, they should build it in this website one that locally generates on simple info.

[NiHaoMike]

I dont have software to render schematics, got any free ones ?

DaveCAD might be really limited in features, but is definitely the easiest to use.
Fork: https://github.com/scanlime/davecad
Original: https://github.com/EEVblog/DaveCAD

[wraper]

I have no zeners in house, another parts series to buy ?, do i really need them...

2 resistors for divider sounds too difficult? What's the point using capacitor to begin with?

[Jan Audio]

The saw wave i put in now is from 6,6v to 11,5v.

[SiliconWizard]

Hi, sorry what is a PN ?

Texas is : SN74HCT74N
NXP : 74HCT74, i have to look.

Just take a look at the datasheets. The NXP one has a schmitt trigger in the clock input (actually most NXP logic parts have schmitt trigger inputs AFAIK), wheras the TI most likely doesn't (the DS doesn't make any mention of it).
NXP: https://assets.nexperia.com/documents/data-sheet/74HC_HCT74.pdf
TI: http://www.ti.com/lit/ds/symlink/sn74hct74.pdf

The lack of a schmitt trigger in the TI part makes the internal clock signal bounce in your case (it's not a digital signal with "clean" edges but an high-pass filtered signal.)

Don't remove the series resistor. You're going to kill the 74HCT74 eventually. If you didn't kill it yet, it's thanks to the high-pass filter which limits the duration of the 12V pulse and the internal clamping diodes. You may end up frying one.

With this kind of circuit, your best bet is to use a 74HCT74 with a schmitt trigger clock input such as the NXP one (I'm sure there are other vendors with this).

You could also add a schmitt trigger buffer in front of it, such as the SN74LVC1G17 (it's a single schmitt trigger buffer, the LVC series should be ok for your requirements and it's easily available whereas the 74HCT1G17 is not.)

If this was meant to be in a product, the second option would probably be better, because using a "standard" part (a reference available from many vendors) and requiring specific features in it from a specific vendor is a recipe for potential supply disaster later on.

[Jan Audio]

Ok, many thanks.
Why did NXP quit ?, and why dont Texas has schmitt-trigger ?

I will take a look for other vendors, its for personal use only, thanks for the advice.
I also dont like to use more space on my circuit for a schmitt trigger.

Maybe Texas themself has also with schmitt trigger, i,m gonna find out.

[Zero999]

Use a potential divider:


R1 = 4k7
R2 = 3k3

I used the site linked below to calculate the resistor values, which I divided by 10, to give more speed and noise immunity.
https://www.random-science-tools.com/electronics/divider.htm

[Jan Audio]

You cant divide 6,6volt off state and 11,5volt on state, the zener is needed also.

[Zero999]

You cant divide 6,6volt off state and 11,5volt on state, the zener is needed also.

You're right, I missed the 6.6V and 11.5V part. The resistor values could be changed, but it would be marginal.

If you knew that, then why didn't you mention the zener diode before?

Change R2 for a 4.7V or 5.1V zener diode and keep R1 the same.

I have no zeners in house, another parts series to buy ?, do i really need them...

Zener diodes are not expensive.

[Jan Audio]

The zeners might be usefull, i go look for those.
I think the chip wont break since the difference is about 5 volt.
+ the saw is also a slow ramp.

As long it works im happy for now, going to solder a new circuit.

[IDEngineer]

the saw is also a slow ramp.

I have a remarkably similar situation in one of my existing designs, and I specified the 74AHC74 specifically because it has a Schmitt Trigger on its CLK input. There is also a 74AHCT74 version if you need the TTL compatible logic levels. A quick glance at DigiKey shows this part stocked in many thousands of pieces and available from at least two manufacturers. Cost starts under $0.25/100's.

[Jan Audio]

Thank you Engineer, usefull information.

[jmelson]

If you really need to work with 12 V signals, maybe you should use a 4000 series part.  I think the equivalent is a 4013, but the pinout will not be identical to the 74xx74.
These are much slower parts, and might work without additional components.  But, if a slow transition on the clock is the problem, then a Schmitt trigger on the clock is the very best solution.

Jon

[David Hess]

Or use 74Cxx which is high voltage CMOS like the 4000 series but with 74xx pinouts and functions.

[IDEngineer]

Or use 74Cxx which is high voltage CMOS like the 4000 series but with 74xx pinouts and functions.

I had the exact same idea, but then I looked at specs and availability... the 74C74 is listed as "obsolete" from many manufacturers and the only one in stock (at DigiKey anyway) wants $7 each in low volumes. Also, the spec sheets I checked do not mention a Schmitt Trigger input on the clock, nor any other input.

There were other CMOS-based 74-series parts back in the day, so perhaps one of those is available and would work, but I'll leave such research as a worthwhile exercise for the OP. If he's got a $7 budget to solve this problem, my 25 cent recommendation and a couple of FET's would allow him to use 3.3 or 5V logic to control a 12V environment and still leave enough money for lunch. 

[David Hess]

Or use 74Cxx which is high voltage CMOS like the 4000 series but with 74xx pinouts and functions.

I had the exact same idea, but then I looked at specs and availability... the 74C74 is listed as "obsolete" from many manufacturers and the only one in stock (at DigiKey anyway) wants $7 each in low volumes. Also, the spec sheets I checked do not mention a Schmitt Trigger input on the clock, nor any other input.

Clock inputs often use Schmitt trigger operation to avoid the exact problem you had but like I said, this is not always reflected accurately in the datasheet and it varies between manufacturers for the same part.

There were other CMOS-based 74-series parts back in the day, so perhaps one of those is available and would work, but I'll leave such research as a worthwhile exercise for the OP. If he's got a $7 budget to solve this problem, my 25 cent recommendation and a couple of FET's would allow him to use 3.3 or 5V logic to control a 12V environment and still leave enough money for lunch. 

Yea, availability and pricing will be a problem for such an old part and technology.  I do not know offhand what is still considered current.

It is too bad because high voltage 4000 and 74C series CMOS are very useful in mixed-signal applications which now have to be redesigned to use a lower voltage logic family and level shifting.

[vk6zgo]

IC manufacturers are not always innocent when it comes to strange things happening when using supposedly identical devices only separated by a fair few years between manufacture.

I had to work on a piece of equipment which used a mono stable IC in the 4000 series, where it was used to provide a quite long "on" time.

Unfortunately, one of these devices became lost.
"No worries, there are plenty in the Store!"
When I tried to use the new one, it kept timing out early, so the equipment didn't work.
The same result with all the new ones in the Store.

It turns out that the manufacturer originally specified the mono for a very large range of "on" times, but some users had problems with them with short
"on" times.

The answer was to change the design, so that the original part number was good for fairly short durations, & introduce a new part number for people who want long durations.

We had to specially order the "long duration" version, replace the originals with the new type in both the units using them, generate an internal Tech note, & change the original schematic.......Aaaarrrrgggghhhh!

[Jan Audio]

So i build the new circuit, now i have something weird that wont happen with the NXP :
If i touch with the scope or even a loose wire the output of the SN74HCT74N, all working good, soon as i release the wire, the thing stops working.

In my circuit i attatched the output signal to the base of a transistor with a 470K resistor.

I hope the SN74AHCT74N dont have that problem.
Can someone explain ?
thanks

[IDEngineer]

If i touch with the scope or even a loose wire the output of the SN74HCT74N, all working good, soon as i release the wire, the thing stops working.

Classic behavior of a very high impedance node. Try lowering that 470K value, paying attention to the output current capacity of the '74.

[Jan Audio]

Thanks, i go read some datasheets.

With a 10K pulldown resistor also did not work.

[IDEngineer]

Wow, do I wish we could see a non-ASCII schematic. Even pencil-on-napkin would be helpful.

[Jan Audio]

Ok, i will download DaveCAD somehow.

I cant find any software on this link :
https://github.com/EEVblog/DaveCAD

Where do i download ?

[SiliconWizard]

Ok, i will download DaveCAD somehow.

I cant find any software on this link :
https://github.com/EEVblog/DaveCAD

Where do i download ?

 

[IDEngineer]

Wow, do I wish we could see a non-ASCII schematic. Even pencil-on-napkin would be helpful.

Suggestion: Just sketch out your schematic on a piece of paper, scan it, and post it here. Stay focused on your actual project for now. Learning any CAD-style application will take non-zero time that you could instead be using to answer your circuit questions.

[Jan Audio]

Thank you again IDEngineer, the AHCT works perfect.

[IDEngineer]

Thank you again IDEngineer, the AHCT works perfect.

Glad to hear things are going better! Sometimes "older" parts are still good solutions for certain kinds of problems.

[Gyro]

...and sometimes, understanding exactly what is going on, and not relying on marginal factors, makes reliable products!

[Jan Audio]

Oh wait, i had to change the 8K2 to 5K6, i take 4K7 to be sure.
So there still is a difference between Texas and NXP.
The driving of the 470K resistor was ok.
Just to make this story more complete.