• TI Thinks Resolved

SN65HVD72: No load supply current at up to 128kBaud datarate

Part Number: SN65HVD72

I'm investigating using this component for a low power sensor interface. The plan is to have no termination resistors (i.e. short stub length). I'm referring to Figure 6 of the datasheet. Unfortunately there is no 'Supply Current vs Signal Rate' graph for the unloaded case. For example a competitor's product showing this is the Renesas ISL32603 in figure 27. Can You provide similar data for the unloaded case? Alternatively, is there a formula for calculating this (e.g. quiescent power + Cpd * Vcc^2 * fsw).

Thanks and best regards,

Stefan

  • Stefan,

    You can refer to this blog for how to estimate the power consumption. BTW, I think your equation is correct.

    What's the capacitive load of your application? Is it close to 50pF like ISL32603's data sheet? If the estimation is not enough, please tell me the information, I could get some measurement for you with a typical test condition (room temperature, Vcc=3.3V).

    Regards,

    Hao

  • In reply to Hao L:

    Hi Hao,

    thanks for Your reply! I read through Your (very helpful !) blog and made some comparison to Figure 6 of the data sheet.

    By doing so, I come to the conclusion that increasing supply current over data rate at 50pf load must by far be dominated by components not covered in the blog (like cross conduction).

    Here is what I calculated:

    ICCic = 0.75 mA (typical spec from datasheet)

    + Idc_noload = 0 mA (because no load)

    + Iac = a few uA (negligible for low data rates and small C load)

    Itotal_noload = 0.75 mA (no load case)

    Then for comparison with figure 6 I add the load current for the 54 Ohm load:

    Idc_54Ohmload = 2V / 54 Ohm = 37 mA

    Itotal_54Ohmload = 37.75 mA (54 Ohm load case), almost independent of data rate (low C load and data rate assumed)

    Figure 6 instead shows a slope of increasing supply current over data rate, which cannot be explained by capacitive loading. Unfortunately, for my application (no load) I'm worrying about the same slope over data rate (without the 54 Ohm load current of course). Therefore only a no load plot over data rate could help clarify this.

    So may I ask You to provide some data as You offered for the following use case:

    Vcc = 3.3V, room temperature, no resistive load, no capacitive load (or 50 pF if You want, but this should make little difference), data rates 9.6k, 38.4k, 115.2k and 250k bits per second.

    From this I could draw the slope.

    The same question and request applies to my other post for the THVD1410.

    Another question is: If these devices do not meet my low power requirements, does TI have alternative parts for low power differential interfaces? These must not necessarily be RS485.

    Thanks and best regards,

    Stefan

  • In reply to Stefan Olejniczak57:

    Stefan,

    Your calculation looks correct to me. The non-linear curve is due to some non-ideal behavior of the IC. I'll take the data and post them in a few days.

    Regards,
    Hao
  • In reply to Hao L:

    data rate (kbps) Icc (mA)
    0 0.256
    9.6 0.27
    38.4 0.307
    115.2 0.403
    250 0.57

    Stefan,

    Please find the bench data. The test condition is Vcc=3.3V, room temp, DE=H, REB=H, no load. Please let me know if you have more questions.

    Regards,

    Hao

  • In reply to Hao L:

    Hi Hao,

    thanks for Your measurements! The supply current looks surprisingly low to me, which i would never have expected from the datasheet. So this helps a lot and yes, this device would meet my low power requirement. However I notice that You had the receiver disabeled for the measurement. May I ask You - if You didn't already - repeat the same measurement with same parameters, but with the receiver enabled? This could assure that there is no hidden penalty with the receiver power consumption (although my application could disable the receiver for quite a time). I would then order two evalboards and benchmark with my communication protocol.

    Thanks and best regards,

    Stefan