24V signals to logic level

Discussion in 'General Electronics Chat' started by panic mode, Mar 3, 2013.

  1. panic mode

    Thread Starter Senior Member

    Oct 10, 2011
    hi everyone,

    i'm looking for ideas for interface ICs because after lengthy search i'm mostly hitting the same part numbers.

    i've been working on couple of projects requiring digital I/O interface to 24V. now i'm doing another one and i'm trying to go away from conventional or customized level shifting using resistors and clamping diodes or zeners or optocouplers etc. while it has all flexibility one would need, is just too bulky and it takes too many parts. i'm looking for something more integrated (fewer parts), and able to handle multiple channels.

    it could be very simple, something like ULN2002 which i would use too if the Vce_on was not that high. there must be some hopefully single chip solution not costing an arm and leg that integrates all the jazz (voltage dividers, clamping diodes and level shifter) with 8-16 channels. can you offer some part numbers worth checking out? i'm just checked 74HC07 which has clamping diodes built in on each input (so only needs resistor network for non-isolated interface).

    for outputs i've been using UDN29xx and VNQ810 but would be interested in something that has true PNP open collector - something that does a bit of both (like UDN but with low Vce, not Darlington emitter follower).

    what do you think
  2. SPQR


    Nov 4, 2011
    Ok, Let's see if I understand.

    You have some device that outputs 24V, and you want to lower that voltage to 5V.

    How fast does it have to go "off and on" - is it just a switch, or are you cycling it at a high rate?

    Some open collector buffers/drivers are
    Open drains:
  3. panic mode

    Thread Starter Senior Member

    Oct 10, 2011

    there is a bunch of signals i need to bring down to MCU. i don't need isolation in this case and don't like to put bunch of
    - resistors and diodes (diode gong to +5V to limit voltage seen by input) or
    - resistors and zeners (similar thing, zener clamps voltage to something input will handle) or
    - resistors and optocouplers (no need to isolation, optos take more room) etc.

    for few i/o or boards with more room this was not an issue but i really don't want to increase board size for something like this.

    inputs will never exceed 1kHz. two outputs may go up to 400kHz in worst case but will probably be under 50-100kHz ( still needs to be seen).

    good stuf, thank you.

    yes 74x07 goes to 30V or more, but 74x17 only to 15V and therefore it is not suitable.
    moreover both are open collector/drain ("sink" type outputs) and therefore they are unsuitable for my outputs. they also don't have built in voltage clamping devices at their inputs (or datasheet does not mention them) so can't be used to condition inputs either.

    CD74HCT541E has tristate outputs (low voltage) and has no input clamping so it also is ruled out.

    i mentioned 74HC07 which according to datasheet does have diodes at the inputs (+/-20mA max, see attachment). so bringing higher voltage input should be simple (just adding resistor to each input or one resistor network for several inputs). the downside is it is relatively low density device (6 channel instead of 8 or 16, but not too bad). also the output (open drain) is not high voltage but since output is sink, it is not going to fit the outputs job (but it can be used to replace bunch of diodes when used to condition inputs or to drive internal loads - though i use ULN2803A for that).

    UDN2982 and VNQ810 or VNQ860 are high side switches (source outputs). VNQ's are nice but have low channel density (but they do have feedback and are high current outputs which can be handy sometimes). UDN2982A is high density source (complement of ULN2803A) but the outputs are not open collector/drain, they are emitter follower. in many cases this is not an issue and a bit of protection is a good thing but sometimes the excessive voltage drop could be a problem. for those few cases one can make own driver with a pair of transistors - NPN drives PNP (both common emitter) so that Vce on outout is only 0.1-0.2V. darlingtons and emitter followers like UDN have voltage drop that is an order of magnitude larger (1-2V) and if i could find something that addresses this, i would keep it in mind.

    here are some details from 74HC07 and 7417 datasheet. note that 7417 input does have one diode (against reversed polarity) but since this is not a zener, it does nothing for case where input voltage is above Vcc (so i would have to add those externally - something i would like to avoid). in the old days it was not hard to find SIP diode networks with 8 or 10 diodes, but modern ones are usually just 2 diodes in sot package. there must be ICs for appliances or similar that do this all in one chip. and i think i've seen some a while back but can't find anything now.:rolleyes:
  4. Papabravo


    Feb 24, 2006
    I would use a comparator like an LM393, LM339, or the single in the SOT23-5 package. The threshold input needs to be set away from the rails to avoid the common mode voltage restriction. The output is open collector so it is compatible with any micro processor system.