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.

SN65HVS882: 20 mA current draw on input

Part Number: SN65HVS882

I've an odd situation where the input seems to be drawing nearly 20mA (~22V across a 1k2). I've., at least temporarily, run out of things to check and am wondering if anyone else has seen something like this. I have ~23V5 at the input to the resistor R82 of the switch input filter and a little over 1V on the other side of the resistor. I've checked two boards, both are behaving in the same fashion. As a last stab in the dark I removed the capacitor (C70) on the input filter in case the caps were all bad. As expected that didn't help. The power pad is connected to the ground plane. DB0 and DB1 are floating (read about 4V5).

I don't really expect that there's an obvious chip issue but in case I've missed something or this triggers some recall by someone I thought I'd ask before I dig on the board some more.

Power input

Switch input & LED

GND is 0V above. GND2 is 0V/2 above

Robert

  • Hi Robert,

    The current is limited by the Rlim resistor. What value of Rlim are you using? and what is you current limit?

    While the device is specified for a current limit of 3.6 mA, (via RLIM = 25 kΩ), it is easy to lower the current limit to further reduce the power consumption as shown in the datasheet "Digital Inputs" section.

    Thanks

    Saminah

  • The Rlim is as shown on the schematic (36k) so that should be around 2.5mA. Are you suggesting there's another way to measure the current limit?

    Robert

  • Hi Robert,

    Can you share details on what is exactly connected to the Switch and Input you are showing Switch input & LED schematic - How is it connected to the SN65HV

    Looks like the switch node is getting pulled down by GND as you are seeing 1V on the other side of the resistor.

    Regards

    Saminah

  • Saminah,

    The input labeled switch is connected to +24V. It's actually connected by a solder bridge for testing. I've measured that potential as well.

    One side of the resistor is measuring at a nominal 24V (As I recall it's actually more like 23.5) and the other at 1V. Thus my comment about drawing 20mA. Given the size of the resistor it's getting hot to the touch as well, so that's further confirmation of the draw.

    Robert

  • Added 4K75 resistor in series to the 1K2

    • Vin = 23V68 4.973 205
    • Between 4K75 and 1K2 = 4V973 --> (23.68-4.973)/4750 = 3.9mA
    • Between 1K2 and pin 4 =  0V205 --> (4.973-0.205)/1200 = 4mA
    • Sanity check (23.68-0.205)/(1200+4750) = 3.9mA
    • Equivalent R drop Pin 4 to Gnd = (0.205/0.039) = 52R

    Added a 910R resistor in series to the 1K2

    • Vin = 23V61
    • Between 910R and 1K2 = 13V80 --> (23.61-13.8)/910 = 10.8mA
    • Between 1K2 and pin 4 =  0V576 --> (13.8-0.576)/1200 = 11mA
    • Sanity check (23.61-0.576)/(1200+910) = 10.9mA
    • Equivalent R drop Pin 4 to Gnd = (0.576/0.011) = 52R

    Without the 910R resistor in series

    • Vin = 23V6
    • Between 1K2 and pin 4 =1V06 --> (23.6-1.06)/1200 = 18.8mA
    • Equivalent R drop Pin 4 to Gnd = (1.06/0.019) = 56R

    I thought that the second two equivalent resistance drops looked suspiciously similar so I added the first to confirm.

    Confirmed that not only was there no resistor from pin 4 to gnd and no resistor is measured at power off.

    It almost looks like the inputs are simulating a resistor,

    Robert

  • Saminah,

    That drops the input voltage even further below the detection threshold. It makes that problem worse, not better.

    Robert

  • Robert,

    you mentioned the input labeled switch is connected to +24V. - what is "Input" connected to?

    Does the input connect to "IPx" of SN65HVS882?

  • Saminah,

    Yes, the block I'm testing connects to connects to IP0 (pin 4). It's a hierarchical schematic and I left out the breakdown of switch input block (8) #0 which just repeats the above circuit 8 times (once for each input).

    Robert

  • I'm at a loss here. As near as I can tell the connections match the schematic and the non-switch signal levels seem consistent.

    So right now all my measurements and checks actually seem to point to the '882.

    Robert

  • Hi Robert,

    To help debug:

    • Have you checked when you remove the 882 and short the signal path? Does the issue still exist?   In most of the cases, the IC is damaged as it exceeds abs max conditions.
    • Does removing the IC and replacing it with a new one solve the issue?

     You can of course lower the input voltage to reduce the power consumption, but  that is something you have to see for your application

    Regards

    Saminah

  • Saminah C said:
    Have you checked when you remove the 882 and short the signal path?

    Umm, there's no signal then, what's to check?

    Saminah C said:
    Does removing the IC and replacing it with a new one solve the issue?

    I've checked multiple ICs on multiple boards. All the same. Replacing the IC is decidedly not easy since it's a powerpad device. I'm not sure my equipment will remove it.

    Saminah C said:
     You can of course lower the input voltage to reduce the power consumption,

    Power consumption is a symptom, not the problem. The current is supposed to be limited to 2.5mA and then the voltage on the input will rise above the detection threshold. That is not what's happening. That you ask this question raises a concern.

    Saminah C said:
     In most of the cases, the IC is damaged as it exceeds abs max conditions.

    This is a 24V device. Abs max is 36V on any non-logic input.

    Robert

  • I can check at a lower voltage but given the resistance measurements I already made I'd be be surprised if it had any effect.

    Robert

  • As expected with supply set a little above 16V

    • Vin = 15V38
    • Between 1K2 and pin 4 =0V651 --> (15.38-0.651)/1200 = 15.4mA
    • Equivalent R drop Pin 4 to Gnd = (0.651/0.015) = 43R

    The current mirror does appear a little different but nothing substantial appears to have changed.

    Robert

  • Hi Robert,

    How about the REx pin voltage in these tests? Does the 20mA happen on multiple devices? 

    Also, what is the voltage measured on the RLIM?

    Regards

    Saminah

     

  • Saminah,

    Saminah C said:
    How about the REx pin voltage in these tests?

    ~0V. The LEDs do not light.

    Just checked, the actual reading is ~20mV

    Saminah C said:
    Does the 20mA happen on multiple devices? 

    Three devices over two boards.

    Saminah C said:
    Also, what is the voltage measured on the RLIM?

    1V253

    Robert

  • I think the Rlim side works fine.

    Rlim=25kOhm: 1.25V/25k=50uA, by which the current limit is 50uA*72=3.6mA.

    Rlim=36kOhm 1.25V/36k=35uA, by which the current limit is 35uA*72=2.5mA.

     

    Can you remove the LED to see if the IPx current could decrease?

  • My scope probes were elsewhere and I had some inexpensive replacements coming in so I waited for those before removing an LED.

    Checking the 24V shows less than 20mV of noise.

    Re-checking a second input on the same device shows no difference.

    Finally removing the LED also makes no difference.

    Robert

  • Hi Robert,

    The best way to isolate this issue is to try using only the 882 on a fresh board with 24V input and check what current you see? (without connecting the LED driver or any other external component not required for 882).

    The 20mA current makes me think the IC is damaged but you mentioned you have tested with other 882 devices and saw the same issue?

    Thanks

    Saminah

  • Saminah C said:
    you mentioned you have tested with other 882 devices and saw the same issue?

    Yes, I have.

    Saminah C said:
    The best way to isolate this issue is to try using only the 882 on a fresh board with 24V input and check what current you see? (without connecting the LED driver or any other external component not required for 882).

    A clarification question: The block diagram for the IC shows that the input current on the IP pins flows out to ground through the LED pins. That would make the LEDs (or at least a resistor) mandatory. Is that the case?

    I just received an eval board and replacement ICs for further comparison as well. It looks like I need to figure out how to take this chip off. Maybe a chisel :) Seriously I'm planning on cutting the pins and then seeing what I can attempt to free the IC from the thermal pad.

    I'm leaning towards some sort of bad batch of ICs or some sort of thermal failure during reflow.

    Robert

  • Robert,

    See the following implementation without LED  -The most simple input configuration without input resistors and indicator LEDs is often used in low-cost micro-controller applications, where status LEDs are connected to the general purpose outputs of the controller.

    I am also leaning towards damaged IC.

    Regards

    Saminah

  • Found the root of the problem. I will need to figure out whether the ICs have survived and come up with re-work for further testing

    If you look at the initial post (and highlighted here) you can see that the input and LED pins are reversed, the pin numbers are backwards. I now get about 2V9 across the 1k2 or about 2.4 mA

    Everything fits.

    Robert

    I would still like an answer to my circuit question above. As long as I can either use an open in place of the removed LED or a short then the rework from my testing will be easier.

  • Robert,

    Good to know the root cause got identified.

    For your question on LED, I think either way should work. If REx is short to ground, input current  will flow into both path 1 and 2. If REx is open, the current only goes into path 2. However, grounding might be a better choice since some items are specified this way.  

     

    Thanks

    Saminah