• State Resolved
  • Locked Locked
  • Replies 17 replies
  • Subscribers 64 subscribers
  • Views 4192 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

fragility of LM5041

Daniele Nardi
Prodigy 140 points
Other Parts Discussed in Thread: LM5041, LM5041B, LM5041A

I've selected LM5041 in 2005 for a wide DC/DC power converter: 16-150V input, multiple isolated output for about 150W.
The scheme is very similar to evaluation board with some little difference: a diode instead of low side mos for the step-down, and a linear regulator for VCC and VIN pin without an aux return by the switching stage.

In this 10 years more that 90% of issue on this ECU (that include microprocessors, PLDs, and many digital/analog input/output) was from the LM5041 failure. In most cases I was found the IC in short circuit in most of pins whit all other components in regular working conditions.

Some last issue are much more enigmatic in some pieces of LM5041 (that I've buy on 2010):
 - at standard temperature (20°C) work all fine
 - when the power convertert start up at ambient temp under 2-4°C the push-pull stage seem ok, but step-down stage don't start.
In this condition internal Vref is OK (4,999V), UVLO at about 4,9V, CS 0V, and (where I think there is the problem) SS don't go up 0,8V, and COMP stay around 1,4-1,5V whitout any external forcing.
After replacement of LM5041 with a new one the converter work fine in all temp range.

Has LM5041 some failing that kill the component during regular working?
There are revisions of the IC since 2005 to today?
Can I expect an increased robustness in the LM5041 that I buy today against of the pieces I've mounted from 2005?
Can I expect increased immunity on newest LM5041A and LM5041B?

Thank you.

  • 0
    TI__Mastermind 31580 points
    Hi Daniele,

    Can you tell me if this is a low temperature only problem and what % of the LM5041 are affected?

    Thanks

    Peter
  • 0 in reply to Peter Meaney
    Prodigy 140 points

    Hi Peter,
        thank you for your answer.
    We have a 30% of warranty return in total actual production (about 70pcs in railways market), and 90% of that are for LM5041 issues. In 80% of this the LM5041 was in short-circuit, in 20% was with temperature problem.

    We are going to evaluate a new contract for many units, and we must to know if we can expect the same return for this component in new productions.

    Thank

  • 0 in reply to Daniele Nardi
    TI__Mastermind 31580 points
    Hi Daniele,

    If the LM5041 are short circuit on most of the pins, I suspect the device has seen an electrical overstress between the HV pin and ground.

    Is there a system issue that can cause the high voltage supply on Vin to the LM5041 to exceed a 100V?

    In general I recommend using the latest versions of devices if they are compatible with the application.

    Regards

    Peter
  • 0 in reply to Peter Meaney
    Prodigy 140 points

    Hi Peter, thank you for your support.

    Vin and Vcc pins are under the same regulated voltage.

    If I propose to my customer to invest in an upgrade of all the units with a new LM5041 ("A" or "B") I must be sure that it's an improvement.. right?

  • 0 in reply to Daniele Nardi
    TI__Mastermind 31580 points

    Hi Daniele,

    I have checked interanlly and there is no record of any issue with the LM5041.

    The Vcc comes from Q9, is that right? Is there anything else damaged when the LM5041 fails short, power stage components, gate driver?

    Regards

    Peter

     

     

     

  • 0 in reply to Peter Meaney
    Prodigy 140 points
    Yes, Q9 is the voltage regulator for the startup. After the first stage (step-down) another voltage regulator (same scheme as Q9) provide for the Vcc from 16V of step-down.

    When I found LM5041 in short cirquit was enough to replace it to have the power stage in working condition
  • 0 in reply to Daniele Nardi
    TI__Mastermind 31580 points
    The max voltage rating for Vcc is 16V, is it possible that transient load on the system can casue the Vcc rail to exceed 16V. Have you looked at the VCC rail across the IC under abnormal operaing conditions to ensure the linear regulator and zener diode maintain the DC at less than 16V? Are there any voltage spikes on the rail due to switching currents from the gate drive outputs.
  • 0 in reply to Peter Meaney
    Prodigy 140 points
    Vcc is regulated at 12V. I have never found voltage over 13V in all functional condition.That solution is to prevent spike on Vcc for possible switching controller issues.
    What do you mean with "abnormal"?
    Gates seem free from current reaction on push and pull, never replaced the IR2125.

    My questions remain the same:
    can I expect an increased robustness in the LM5041 that I buy today against of the pieces I've mounted from 2005?
    can I expect increased immunity on newest LM5041A and LM5041B?

    Not only the short circuit trouble my customer, but the temperature issue too.

    Thank for your support Peter
  • 0 in reply to Daniele Nardi
    TI__Mastermind 31580 points
    Hi Daniele,

    Since there is no known issue with the LM5041 that should results in the short circuit problem I dont think moving to the A or B version will make any difference. I suggest that you return one or two samples of the short circuited devices for failure analysis and do the same for the temperature issue.

    Regards

    Peter
  • 0 in reply to Peter Meaney
    Prodigy 140 points

    Ok Peter, I'll re-open this thread when I've another LM5041 to give you.

    Thank you very much for your support.

  • 0 in reply to Daniele Nardi
    Prodigy 140 points
    Hi Peter,
    I've another devices that don't startup at low temp: do you want some particular measure before I replace this LM5041?
  • 0 in reply to Daniele Nardi
    TI__Mastermind 31580 points
    Hi Daniele,

    Let me look at some earlier notes to see what signals to look at, I would also like to discuss this with another engineer who is travelling at the moment so it may be this afternoon before I get back to you.

    Regards

    Peter
  • 0 in reply to Peter Meaney
    TI__Mastermind 31580 points

    Hi Daniele

    Can you capture the waveforms, voltage levels, on these pins during start-up COMP, SS, VCC, VREF, UVLO.

    If the IC refuses to start up, try shorting COMP to GND briefly and release it to see if the IC will then start.

    Regards

    Peter

  • 0 in reply to Peter Meaney
    Prodigy 140 points

    Hi Peter,
      these are the record of the signals. I've a 2 ch oscilloscope, than I've syncronized all the screens whit the UVLO signal (a power-control release UVLO after complete charge of input capacitors, than that is the natural sync).

    At 20°C (in LM5041 zone) switching regulator work fine:

    Vcc@20°:

    after the charge of capacitors the Q9 supply with little less of 12V, than (after switching start) auxiliary's regulator supply with little more of 12V

    COMP@20°C:

    at UVLO release, it go up in a ramp and "stop" when feedback intercept the target output

    REF@20°C:

    at UVLO release, it go stable at 5V

    SS@20°C:

    before UVLO release: ramp and still to 0,8V

    after UVLO release: free going up

    At -1°C (in LM5041 zone) switching regulator don't work (don't start-up, or if it's going it stop):

    Vcc@-1°:

    after the charge of capacitors the Q9 supply with little less of 12V

    COMP@-1°C:

    at UVLO release, it go to about 1,25V and still

    REF@-1°C:

    at UVLO release, it go stable at 5V

    SS@-1°C:

    before UVLO release: ramp and still to 0,8V

    after UVLO release: it's locked to less than 1V

    At 1°C (in LM5041 zone) switching regulator don't work, but after many failure it start successfull. Interesting signals are:

    COMP@1°C:

    at UVLO release, it go to about 1,25V and still

    after few moments (don't look at the time in the bottom, because I restart the unit after every screenshot) LM5041 try to start until it start

    REF@1°C:

    at UVLO release, it go stable at 5V

    SS@1°C:

    before UVLO release: ramp and still to 0,8V

    after UVLO release: it's locked to less than 1V

    after few moments LM5041 try to start like COMP signal, until it start

    Now I must replace this LM5041 because my customer need for this unit, can you tell me the shipment address for send you the LM5041?
    Have you a contract for the shipment from Italy?

    Thank for all.

    PS: when LM5041 don't work the COMP reset has no effect.


    PPS: Later I'll send you record of same signals with the new LM5041 (made by a recent sample order)

  • 0 in reply to Daniele Nardi
    TI__Mastermind 31580 points

    Hi Daniele,

    Thanks for all the information. Did you try to short the COMP pin to ground and then release it at low temeprature to see if this would get the IC to start?

    You can ship the IC to:

    Texas Instruments

    5 Eastgate Avenue

    Little Island

    Cork

    Regards

     

    Peter

  • 0 in reply to Peter Meaney
    Prodigy 140 points
    Hi Peter,
    yes, like in "PS" the pull down of COMP has no effect. It's the same with the SS, and the same with a pull-up of 5k resistor on COMP or SS.
    I've some other information that I've before the restored unit was shipped.

    In my mind is that the issue have a relationship with CS input, because when I've shorted this input with a stable wire (not with a manual short circuit) the issue decrease significantly.
    It is as if CS was much more susceptible in the used LM5041 compared to the new one.

    As in the past, after replacement of the LM5041 with a new one the unit can work fine under -10°C.

    I'm looking for a good shipment of 2 LM5041 to you.

    Thank for all
  • 0 in reply to Daniele Nardi
    TI__Mastermind 31580 points

    Hi Daniele,

     

    Feel free to email me at petermeaneyti.com

     

    Regards

     

    Peter