• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 18 subscribers
  • Views 2795 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

SN74HC595: Source/sink current of 74HC595@3.3V and maximum recommended length of daisy chain?

Tuomo Latto
Prodigy 40 points
Part Number: SN74HC595

Two questions:

  1. What sort of current can an SN74HC595 source/sink at Vcc=3.3 V, assuming a voltage dip/rise that is somehow sensible?
  2. What is the maximum recommended length of an SN74HC595 daisy chain at that (or other) voltage(s) and at all rated temperatures?

Having read the posts tagged with this part number, I guess the answer to the first question one can expect is that one should look at the datasheet for 2 V and arrive at the conclusion of 20 μA with ΔV of 0.1 V max. Then again, 2 V is the lower limit of Vcc and at 4.5 V we are looking at 4 mA with max. 0.33 V. Considering 3.3 V is not at the lower limit but actually quite near the half-distance between those two voltages, one could assume the actual figures would be somewhere near 2 mA with max. ΔV around 0.2 V, but since you can't simply assume a linear relationship, I'm left wondering what the actual figures are.

What if an acceptable maximum of ΔV is 0.4 V or even 0.6 V? How much current can you push or pull through an output pin then?  If there were typical (and maximum) characteristic curves on the datasheet, you could at least come up with some realistic approximations for different situation, but there aren't any and I'm stuck asking here. Now, Nexperia (NXP) provides a difficult-to-use SPICE model library-thingy for their HC parts, but I don't know how accurate it is. So, does anyone actually know what one can expect at 3.3 V?


As for the second question, I read the thread about the 16 daisy-chained 74HC595's. That's a bit much for me. I'm only looking to chain 4 of them, probably at max. 2@MHz with a shared clock from a microcontroller, but I think the question is a valid one: What is the length of the daisy chain when you have to start thinking about the propagation delay and delaying the clock, if you want reliable shifting (at full temperature range)?

  • 0
    TI__Guru* 96591 points

    Hi Tuomo,

    To answer your questions directly:

    1. What sort of current can an SN74HC595 source/sink at Vcc=3.3 V, assuming a voltage dip/rise that is somehow sensible?

    The maximum output current of this device is listed in the Absolute Maximum Ratings table as "Continuous output current" (35 mA).  There is a caveat to that in the same table -- the total maximum output current must be limited to 70mA through Vcc or ground -- ie you could be sourcing and sinking 70mA total at the same time, but you can't source or sink 71mA and remain within the datasheet specs.

    The current values you are referring to in the Electrical Characteristics table are test conditions for VOH and VOL which aren't hard limits - they just give you an idea of the output impedance in the worst case. (using Ohm's Law, we can easily find the output impedance from the given values).

    2. What is the maximum recommended length of an SN74HC595 daisy chain at that (or other) voltage(s) and at all rated temperatures?

    There is no length limit, however there are practical limits in a system.

    Think about how exactly the daisy chaining works with just 2 devices. Data from the last bit of one device is shifted into the first bit of the next device at a clock edge. There is no interaction from that bit to any subsequent stage of the daisy chain, so as long as you can time your clock to arrive at every device simultaneously.

    The timing is where things get complicated -- real systems have physical size, and the clock has to travel from chip to chip.  You _could_ match the length of every line to ensure that the clock signal reaches everything at the same time... but then you will run into capacitive load issues at large bit counts.  Then you have to start doing fanout techniques to get enough current to drive the capacitive load, which adds delays, skew, etc.

    The short answer is - you can make the chain as long as you want, but you need to match the timing for the clock pulses as closely as possible to get good results.

    Let me know if I can be of any further assistance!

  • 0 in reply to Emrys Maier
    Prodigy 40 points
    Sorry about the silence. I've been thinking about this and it doesn't seem quite right.
    You seem to be saying I can simply assume ohmic resistance/impedance for the voltage drop/rise.
    Fair enough, but the impedances calculated this way are (respectively)
    Typ 25°C, Max 25°C, Max full range:
    at 2V @Iol/Ioh=20μA: 100ohm, 5kohm, 5kohm
    at 4.5V & 6V @Iol/Ioh=20μA: 50ohm, 5kohm, 5kohm

    at 4.5V @Ioh=6mA: 33.33ohm, 86.66ohm, 110ohm
    at 6V @Ioh=7.8mA: 25.64ohm, 66.66ohm, 84.6ohm

    at 4.5V @Iol=6mA: 28.33ohm, 43.33ohm, 55ohm
    at 6V @Iol=7.8mA: 19.23ohm, 33.33ohm, 42.31ohm

    There is still quite a variance in impedances and a huge difference between the maximum impedances for 20μA and 6mA currents. If I'm interested in 1-2 mA @ 3.3 V, I'm looking at maximum voltage drop between 110-220 mV and 5-10V.
    You're right, this is easy, but unfortunately it doesn't seem very useful.
    What am I missing?
  • 0 in reply to Tuomo Latto
    TI__Guru* 96591 points

    The 20uA numbers are not the most accurate - they give a lot of head room (thus why you see 5kohm values there). Most parts will list 100mV for the voltage drop on the FET at 20uA, but in reality the drop will be far less.  In typical operation, you will see ~Vcc or ~gnd on the output for such a low output current (as is shown in the 'typical' VOH/VOL numbers). 

    It's unlikely that you would actually see 5kohm of impedance at 2V, but I wouldn't be surprised to see quite a bit (even  500+ ohms). Since there aren't higher current numbers at 2V, I would ignore those numbers for your estimation.

    We usually use linear interpolation for finding this type of value -- but since there isn't a good 2V value to use, we have to just use the 2 available.  With those, I see ~130 ohms over temperature, which sounds right.  That means the short circuit current would be ~25mA, and at 1mA you'd see around 130mV of voltage drop in the worst case.

    Can you tell me about the project that has you so interested in the current capabilities of this 30+ year old device?

  • 0 in reply to Emrys Maier
    Prodigy 40 points
    This is a slow-moving personal project where I'm just trying to utilize 7-segment displays, etc. in an outdoor environment. I want to minimize the voltages and current draw (and obviously balance them against the brightness). The current capabilities are of interest since I would prefer to drive the LEDs directly with the 595s instead of adding 20+ transistors and whatever additional resistors they require.

    Linear interpolation in terms of voltages would then mean ~130 mohm high and ~65 mohm low @ 3.3V worst case and best case ~40/36 mohm? That is certainly a more workable amount of variation.
  • 0 in reply to Tuomo Latto
    TI__Guru* 96591 points
    I think you may have missed a 1000 in there - it should be ~130 ohm to ~40/36 ohm. That sounds very reasonable for a logic device (I didn't go back and check the math).
  • 0 in reply to Emrys Maier
    Prodigy 40 points
    Right. I have no idea what I was thinking. I was focusing on the numbers, so most likely I wasn't.