-
What to do with an unused monostable? (CD74HC123)
Posted by
Faringdon
on 30 Jun, 2022 18:26
-
-
#1 Reply
Posted by
RJSV
on 30 Jun, 2022 18:44
-
Yeah, I'd bet money there is an energy waste penalty, to drain batteries faster, if don't follow data sheet recommendations, (I would just connect unused inputs t gnd). That could be, also, spurious pickups, from surroundings, or even the power rail could be coupled capacitively. Data sheet, or TTL book.
-
#2 Reply
Posted by
ebastler
on 30 Jun, 2022 18:49
-
The datasheet might include a subtle hint, right there in the first section (Description):
If either Mono is not used each input on the unused device (/A, B, and /R) must be terminated high or low.
-
#3 Reply
Posted by
Gyro
on 30 Jun, 2022 19:19
-
I think you'll find that Cx and CxRx are (must be) outputs, otherwise how could it work as a monostable?
-
#4 Reply
Posted by
tggzzz
on 30 Jun, 2022 19:33
-
-
#5 Reply
Posted by
jpanhalt
on 30 Jun, 2022 19:36
-
-
#6 Reply
Posted by
Faringdon
on 30 Jun, 2022 20:37
-
Thanks yes, it was Cx and RxCx that were of most confusion....something discharges the C so thats an output...but also, the voltage of the RC point has to be sensed.
-
#7 Reply
Posted by
Faringdon
on 01 Jul, 2022 09:07
-
Thanks, ive actually seen that with 74HC123, it is edge triggered, so the input pulse being longer than the monostable interval (400ns) shouldnt matter?....ie, If A goes high for 1us, the Q will go high for 400ns, then go back low again, even though A is still high for another 600ns?
I mean, you never know, the datasheet doesnt cover this event in the waveforms section....and "edge triggered" doesnt neccesarily mean "doesnt keep triggering if input pulse stays high after the edge"
-
#8 Reply
Posted by
RJSV
on 01 Jul, 2022 09:31
-
Hello. I was confused by question, Edge Triggered is exactly that, and if that input stays high for 1/2 second, so what ? I mean, you get, say 400 nSec result on the output, then the output goes low: meanwhile that long 1/2 second input just sits there, then goes down, at 1/2 second.
Maybe you meant to ask the reverse case, where a new trigger comes, BEFORE current output cycle finished ?
-
#9 Reply
Posted by
Zero999
on 01 Jul, 2022 12:23
-
Use the 74LVC1G123 which is faster, uses less power, is single and takes up less space.
The 74x123 is not like the 555 timer. Whether the trigger is high or low makes no difference to the state. It's only when it changes from high to low, or low to high, does it trigger. It's also retriggereable, which means the cycle will start again, if another trigger pulse is applied in the middle of a cycle.
-
#10 Reply
Posted by
Faringdon
on 01 Jul, 2022 15:19
-
-
#11 Reply
Posted by
Simon
on 01 Jul, 2022 17:04
-
inputs tend to be high impedance so can take little to flip one way or the other by say capacitive coupling in the chip when an output switches. Power is consumed to switch an output so making sure unused sections are statically locked in a stable state reduces spurious switching that will consume power.
-
#12 Reply
Posted by
Zero999
on 01 Jul, 2022 18:53
-
https://www.eevblog.com/2022/07/01/odysee-exclusive-how-i-make-videos/
Thanks, ill bear that in mind...for this, i prefer the SOIC footprint with wider pitch (easier for my rubbish soldering).
Also, for this i dont need high switching speed, just as long as it can get a 400-500ns time for the pulse.
Why have you posted a link to one of Dave's blogs, rather than my post?
I wouldn't say it's too small. If you must use the larger package containing two, the general rule is to connect all unused inputs to 0V or +V. The RC circuit connection should be left open circuit as one side connects to the output of a logic gate.
-
#13 Reply
Posted by
tszaboo
on 01 Jul, 2022 18:59
-
Can't you solve any problem with just a tiny little bit of self-sufficiency?
-
#14 Reply
Posted by
Zero999
on 01 Jul, 2022 19:24
-
Can't you solve any problem with just a tiny little bit of self-sufficiency?
That would be asking too much. Nowadays all one has to do to pass many courses is just tick a few boxes. Gone are the days when students actually had to prove they knew the topic.
Unfortunately how to read even the simplest datasheet is no longer taught.
If either Mono is not used each input on the unused device (A, B, and R) must be terminated high or low
https://www.ti.com/lit/ds/symlink/cd74hc123.pdf
-
#15 Reply
Posted by
ebastler
on 01 Jul, 2022 19:34
-
Unfortunately how to read even the simplest datasheet is no longer taught.
If either Mono is not used each input on the unused device (A, B, and R) must be terminated high or low
https://www.ti.com/lit/ds/symlink/cd74hc123.pdf
Unfortunately how to read even the shortest post is no longer taught either.
See reply #2 above...
-
#16 Reply
Posted by
Zero999
on 01 Jul, 2022 20:20
-
-
#17 Reply
Posted by
Faringdon
on 01 Jul, 2022 22:22
-
Why have you posted a link to one of Dave's blogs, rather than my post?
...Woops!...sorry about that, must have got stuck in the laptop buffer.
Ref CD74HC123, i wrongly assumed that the C is connected to ground...and that theres a shorting FET across the C....(~10R in series with the shorting FET so it can be quickly discharged)...and that when it gets triggered, the C charges from Vcc through the R...and that a comparator reads the RC point (a ref is into the other comparator input), and when its 0.45*R*C after the shorting fet was opened, the Output pulse terminates due to comparator trip......its the "comparator input", if it exists, that worries me, concerning leaving it open.
People often tie unused inputs to rail....and if that was done with the RxCx pin, then the discharge fet (i am assuming there must be one) could short the rail to ground, if it ever somehow got activated.
-
#18 Reply
Posted by
Faringdon
on 02 Aug, 2022 15:19
-
Hi, The CD74HC123M monostable is not working....may i ask do you know why?...
The attached is a synchronised PFC controller using UCC28070A.
The Monostable (CD74HC123M) is not working, ie, its not providing narrow pulses to the UCC28070A. Do you know why?
The UCC28070A now seems to have powered up OK, and the UCC28086D is correctly providing a square wave to the monostable.
At first power up, the UCC28070A Vcc pin was stuck at around the turn-on threshold of some 10.4V. ….And it had 0V on its VREF pin indicating it hadnt woken up. I then changed R7 to a signal diode (anode to CD74..) and then the UCC280780A powered up fine (nice 6V on its VREF pin). However, even though a Square wave (~25% duty cycle) is correctly being sent to the CD74, the square wave is not going back down to 0V at the CD74 side of R3…..it is being held up at around 4.2V. -So the waveform seen at the “1B” pin of the CD74HC123M is just going continuously from 5V to 4.2V.
This must mean that the “1B” pin of the CD74HC123M is broken, because its supposed to be an input, and shouldn’t be able to source voltage.
Do you agree that the CD74HC123M must have somehow gotten damaged, eg, ESD, or perhaps something to do with R7 when it was in the first place, a 100R resistor? ..Also wondering about differences in supply assertion time between the 12V and 5V supplies. UCC28086D and CD74HC123M are supplied by the same 5V rail. UCC28070A is supplied by a 12V rail. (5V rail actually gets derived from the 12V rail so 12V asserts first).
CD74HC123M datasheet:
https://www.ti.com/lit/ds/symlink/cd74hc123.pdf?HQS=dis-mous-null-mousermode-dsf-pf-null-wwe&ts=1659373327659&ref_url=https%253A%252F%252Far.mouser.com%252FUCC28070A datasheet:
https://www.ti.com/lit/ds/symlink/ucc28070a.pdf?ts=1659390653180
-
#19 Reply
Posted by
Faringdon
on 02 Aug, 2022 18:01
-
......now the post just above this is part-fixed..as follows IYAI........I just took off R7, and replaced the CD74HC123M, and it is now giving exactly the right output.........
.so the next thing is to connect it back to the UCC28070A..... the first time i powered up, a voltage
must have coupled through the UCC28070A pin capacitance-to-VCC, and then blew up the CD74.....so , yes, i will use a
1N4148 in there (in place of R7).....must admit i would still like some resistance aswell.....but the
track goes on an inner layer after R7, so not so easy to add an R.......maybe i can do a "wigwam"
type thing with a 1N4148 and a
1k resistor in series between the vias either side of the R7 pads.....
Must admit, i would also now like a signal diode on the CD74 output pin, (just in case it rings below ground) but i'll struggle for room.
I never realised how utterly delicate these CMOS chips are (CD74HC123M). Just because the 12V powered up first...and then capacitively coupled through the vcc-to-pin capacitance of UCC28070A, ignored the R7/input_pin_capacitance time constant of CD74, and then killed the CD74, with what was probably a tiny overvoltage for a tiny amount of time!
*********************************Now not working again*****************************
Just put 1k and a 1n4148 (anode to CD74), instead of R7, , and it now doesnt work......or should i say, when i probe at the mid point between 1k and 1n4148, it pulses for a second or two after power-up, then just goes flat at ~5V.
So it now goes CD74.......1K......1N4148.......UCC28070A
So i am wondering about this now.....maybe i just need to change R7 for 10k?
*********************still not working*********************
However, with 10k instead of R7, the CD74 doesnt work at all.......just get a flat 5v5 output from it. ...So in fact now i am wondering if the UCC28070A is damaged.
Any ideas why the CD74HC123M isnt working?
*********************still not working*********************
Now disconnected R7 from ucc28070a RDM pin, and it (CD74HC123M) still doesnt work.....though every now and again, it does work, for about a second or two after power up....then the CD74 output goes flat at 5V.
*********************
Now just changed the CD74HC123M timing capacitor to fresh one, same value....now it works but only when not connected to the UCC8070A.
-
#20 Reply
Posted by
Zero999
on 04 Aug, 2022 13:10
-
Is the 74HC123 blown again?
Try a Schottky diode such as the BAT54.
Adding some overvoltage protection to the 5V rail is also a good idea. You could try a 6.2V zener diode. Improving the decoupling will also help to absorb voltage spikes. Try adding a 10µF tantalum capacitor, with a voltage rating of at least 10V.
-
#21 Reply
Posted by
Faringdon
on 04 Aug, 2022 13:44
-
Thanks Zero999, sorry i should have come back, it was actually the UCC28070A which was dead...and back-feeding voltage to the CD74HC123M......the CD74 is workign fine now on its own.
The UCC28070A was taking 25mA when doing nothing,...above its abs max.....its soft start pin was also dead....so too gate drivers.
I changed the UCC28070A and its now fine.....so it was likely esd damage possibly during re-reeling.......i have the CD74 oscillator working now, but am not going to bother using it any more......to be honest, its only really needed for the next cct which will be a quad interleaved boost PFC.
The clear tube shown in the top post might actually be from TO247 resistors or diodes, which are , AYK, not needing of ESD protection.....and if the clear plastic tube protects it from any external ESD overvoltage, then why does it need to be in an ESD bag? (it did come in an ESD bag).
AYK, there is a world shortage of FETs/chips/etc....i reckon when you get a FET...theres a good chance it may have been re-reeled/re-tubed multiple times by disty staff not using ESD protection.
Very often, AYK, when you buy small qty of TO220/TO247 FETs, they often come in a wee box with black ESD foam in it, and this is inside an ESD silver bag......so it shows that the clear tube is not the "absolute" method at least......i reckon the "clear tube" is a bit diffy on ESD precautions.
