4538 monostable Cext pin connection

Discussion in 'General Electronics Chat' started by szabikka, Mar 22, 2016.

  1. szabikka

    Thread Starter Member

    Sep 3, 2014
    77
    1
    Hello everyone!

    I have a question regarding one of the 4538 monostable ic's pin connection. I have read two datasheets for the 4538 (included in links below) and both of them states, that pin 1 or pin 15 (doesn't matter which, since it's a dual monostable in one package) apart from connecting to the timing capacitor needs to be connected to ground. However, on page 5 of NXP's datasheet there is no connection to ground and in all of the online available circuit designs there is no connection from pin 1 to ground. On the contrary, on page 2 of Fairchild's datasheet, there is a connection going from pin 1 to ground. I have constructed the circuit on a breadboard each way and I found that both one is operational, although when I add ground, timing is not that accurate. My questions are: Is pin 1 internally grounded so there is no need for external grounding and I misunderstood the "always connected to ground" statement? Can it be harmful for the circuit if I include/do not include an external ground on pin 1?

    Links:
    http://www.nxp.com/documents/data_sheet/HEF4538B.pdf
    https://www.fairchildsemi.com/datasheets/CD/CD4538BC.pdf

    Thank you for the answers in advance!
     
  2. Dodgydave

    Distinguished Member

    Jun 22, 2012
    4,969
    744
    The capacitor pin1 is to prevent the capacitor discharging on power down, so accurate timing is always available, its connected internally via a transistor.
     
  3. szabikka

    Thread Starter Member

    Sep 3, 2014
    77
    1
    Thanks Dodgydave! So I don't have to connect it externally to ground.
     
  4. MrChips

    Moderator

    Oct 2, 2009
    12,421
    3,356
    Monostable multivibrators have a tendency to timeout early in noisy environments.

    Here are some tips:
    1. Eliminate all bipolar 555 timer ICs such as NE555, LM555 from your design. Use CMOS versions such as LMC555, TLC555 instead.
    2. Make sure you have a 100nF decoupling ceramic capacitor between VDD and VSS close to the 4538 IC pins. Keep the capacitor leads as short as possible.
    3. Add an additional 10μF electrolytic capacitor across VDD and Vss.
     
    szabikka likes this.
  5. dl324

    Distinguished Member

    Mar 30, 2015
    3,235
    619
    The NXP datasheet is the only one I've seen that didn't show the ground connection on the caps. FWIW, NXP datasheets are some of the poorest I've seen.

    I always connect pin 1 and/or 15 to ground as every datasheet I had seen before the NXP indicated.
     
    szabikka likes this.
  6. szabikka

    Thread Starter Member

    Sep 3, 2014
    77
    1
    Thank you for the answers!

    MrChips, I always add decoupling caps through the power rails. Those are the first things I actually add, I also like to add voltage regulators, to ensure that momentary changes in power supply voltage don't interfere with the circuits operation. I never use any bipolar 555s anymore as they were the cause of double clocking in one of my counter circuits.

    dl324, in this datasheet they actually contadict themselves by stating one thing in the text and another in the circuit schematic.
     
  7. dl324

    Distinguished Member

    Mar 30, 2015
    3,235
    619
    The NXP part may have the internal VSS connection as they show in their logic diagram, but that makes them non-standard because National and Motorola show an external VSS connection. If you use NXP parts, other parts can't be substituted if the external VSS connection isn't used.
     
    szabikka likes this.
  8. szabikka

    Thread Starter Member

    Sep 3, 2014
    77
    1
    Okay. This is what I found out about the chip I use. It was manufactured by Phillips and their datasheet was hard to find, but I finally succeeded. This paper also states that pin 1 is always connected to ground. On page 3 figure 3 I can see that Ctc pin (pin 1) is connected to the source of the same n-channel mosfet as Vss. From this I presume I don't need to add an external ground. Page 4 which shows the connection of the cap and resistor further confirms this. Am I right?

    Link: http://www.datasheetspdf.com/PDF/HEF4538BP/78494/3
    http://www.datasheetspdf.com/PDF/HEF4538BP/78494/4
     
  9. hp1729

    Well-Known Member

    Nov 23, 2015
    1,940
    218
    When you added the ground timing was not accurate???
    Off by how much, which way? Did it matter who's chip you used?

    (edited to add ...)
    TI data sheet does not show an external ground example on their data sheet either. The data sheets do show the pins internally connected to ground.
    Does the external ground just bypass the internal resistance to ground for when the capacitor discharges?
    Various 4528s and 4538s show anything from 2 ohms to 10 ohms between pin 1 and 8.
     
    Last edited: Mar 22, 2016
  10. szabikka

    Thread Starter Member

    Sep 3, 2014
    77
    1
    I only have a Phillips chip, so I only have experience with that one. I used a 100 uF capacitor and 1M resistor (if I remember correctly), which should give 100 seconds high output when triggered. It was steadily 100 seconds when pin 1 was not externally grounded and varied between 105 and 100 seconds when external grounding was added.

    So on the TI datasheet that ground symbol represented internal grounding. That would be the best scenario as every schematic I have seen so far omits external grounding.
     
  11. hp1729

    Well-Known Member

    Nov 23, 2015
    1,940
    218
    Interesting.
     
  12. dl324

    Distinguished Member

    Mar 30, 2015
    3,235
    619
    Connect pin1/15 as indicated by manufacturer datasheet for the part you're using. Usually manufacturers try to be socket compatible; apparently this is not the case for Phillips. Maybe they thought an internal connection would make their part more desirable. Personally, I don't think it's a good idea.

    Regarding timing accuracy. Caps of the size you're using usually have a 10-50% tolerance. If you want timing accurate to more than your component tolerances can provide, you need to have a way to adjust for tolerances.
     
    szabikka likes this.
Loading...